Document Sample
0_ APPROACHES Powered By Docstoc


              PsycnolNn Lyrlc AppRoecnns                     292
              Bnna,vronlhr RnA,troN ro
                rnB Ecorocy on Dencnns                       295
              OrHpn Psycnor-ocrcnl AppnoAcHEs                304


              Parapraxesand Wit                              292
              The Contribution of ExperimentalPsychologyto
                the Formulation of thc Problern of Safety-
          ,     A Brief for Basic Research                   296
              Complenentary Theoriesof SafetyPsychology      304
              PsychologicalClimate and Accidents
                in an Automotive Plant                       309
              An InterdisciplinaryApproach
                to Accident Patternsin Children              313
              Psychologicaland Psychophysiological Factors
                in Motor Vehiclc Accidents                   327
              The Personalityof Drivers and Pedestrians      335
              Risk and Hazard                                337
              The Risk Taken in Crossinga Road               346
              The Risk Taken in Driving
                Under the Influenceof Alcohol                351
 PSYCHOLOGICAL APPROACI.IES                                                                                                                                         289

TUE stuuy oF AccIDENrs                   from the perspectives psychologyhardly requiresjustifi-
cation.To the extentthat ircciclents                    involvchurlan behavior,they must incvitablybe
studiedwith the science                 that dealswith the behaviorof the indivic{ual.                        And. in fact.
psychological             theory has beenappliedto accidentphenomena a wide variety of             by
peoplc,rangingfrom the layman who believes                                that accidents cirused "the nut
                                                                                               are              by
that   holds the wheel" to the cxperimcntalist                           studyingthe legibilityof road signsor
 speedometer           dials.
     Potentially there is hardly a field of psychologythat does not have implications
for accident         research. a field of study,however,psychology
                                      As                                                          has become broad  so
and conrplexthat one cannot generalize                                                             to
                                                                  aborrt its applic.rbility accidents-eithcr
in terms of the utility of its conceptualizations the value of its researchor                                        findings.
Therefore,                        of
                  instead attemptineto subdivide field and cxaminethe relevance
                                                                        the                                                and
the findingsof each of its subdivisions, us examincsonreof the aspects the
                                                                 let                                                    of
individual that relateboth to accidcnts                        and to psychology           and thcn rcvicw examples
from the wide rangeof psychologically                           orientedaccidentresearch.
    Perhaps         the broadest         area of human behaviorthat is relevantto accidents-and
certainly the area that has becn sub.iected the greatestarnount of psyc:hological
research-is thc capacityof thc human organismto rcspondappropriately scnsory                                      to
stimuli.Color perception,                 depth perception,            perceptual     constirncy,    spatialdiscrimina-
tion, reactiontime, kinesthetics,                    and similar essentially             psychophysical         phenomena
all may haverelevance the avoidance hazards,
                                      to                       of               whetherin a prirnitiveor a highly
technological            environrnent.+
    Lcarningtheoryhasnot yet becnappliedsystematically the learning accident-             to                    of
avoidance,but some psychologists                          have hypothesized             that thc disproportionateiy
high frequency highwayaccidents
                             of                             involving adolescents           may be due not only to
charactcristics            peculiarto adolescence also to thc frequency "inappropriate
                                                               but                                   of
responses"         that is charactcristic the carly stages the learningof any complex
                                                    of                            in
skill (seeMcFarland and Moore in Chap. I and referencc2).
    Studiesof the individual'scognitiveprocesses                              have implications         not only for the
teaching safepractices also for the selective
               of                        but                                    screening individuals occu-
                                                                                              of                   for
pations that directly involvc thc safetyof others as well as themselves. undcr-                                  An
standing his cognitive
              of                        processes yicld knowledge how an individualassesses
                                                      may                              of
thc hazardin a givcn situationand would appearto be especially                                      importantnot only
i n d r i v c r s ,p i l t t s , a n d m a c h i n eo p c r a t o r s u t a l s o i n g o v e r n m e n t a l n d i n d u s t r i a l
                                                                     b                                     a
exccutivcs,       manufacturcrs, othcr planners
                                           and                           whosedecisions          in[iucnce safetyof
entire populations,The pertinentcharacteristics such key individuals,however, of
have yet to be exploredscientifically.                      Thc importanceof small numbcrsof crucially
placedindividuals accident       in            causation                              is
                                                               and prevention easilyseen                day by day. The
small numbersof individualsresponsible the dcsign,course,and spccdof the
,S.S.Iiranic, for decisions              with respectto the crashworthiness vehiclcs,          of               and for the
placement        and strengthof dams illustratcthis wcll, as do many other aspects thc                                  of
modern cnvironmcnt.+                   This is perhapsthe most important uncxplored                         area of parti-
cular interestto psychologists                  and othcr behavioralscientists               concerned       with accident
causation and prevention.
* The psychologically              oriented accident    research     literature  may be entered                  the
                                                                                                      through review$
p r c p a r e r l h y T h t r r n d i k e , t M c F a r l a n t l c t a / . . 1C o l d s t c i n , r u n 4 C r e s s w c l la n d F r o g g a t t ' t aa n d t h r o u g h
nrany of the papcrs published in lhe                                            Satety Rcscarch Review.
290                                                  PSYCHOLOCICAL APPROACHES

                                                                        the capacitiesof troth
signiflcantIor an understandingof ccrirrirraccidentsbecause
                                                                      for perception,   and for
the child anrl thc aging indiviJual for c.nccpt formation.
                                                                      adult in  ways that may
motor responscappear to differ sharply frotn those of the
                                                                                    reason the
make them      differentiallyvulnerableio c'vi'onmental insult. For this
                                                                          through the study of
application to ilre young and thc rrgccl conclusionsreached
yourrgadults is highly questionahle'
'                                                                        of     social group on
     dork in social"psychtlogy-not only on the generaleffect the
                                                 the sir-cialvalucsaud processcs     that affect
the behnvior of thc inclividriottut aiso on
                                          is furtherundcrstanding                    o[Iiustra-
                                                                        o|tlrc effects
 risk-taking-is aIsorelevant. too,
 tion, aggression,  and various   environllentaldistractions       upon stimulus4i5slimina-
 tionoattcntion, perception,and motor rt''sPonsc"
      ilchaviorai iteviniinn-whether proclucedby transient or prolonged
  disturbancc,chcmicalagents,or organicdcficit-may            allect the safetynot only of the
  dcviant but of others.               ut tt't* validity of accident proncnessas a global
                                                                        acute or chronic-"may
  concept,there are many ;dications that.emotionalstates--
                                     but the scierrtifically ilcloquate  suppor:ting evidcnccis
  alter thc incidcnceof aclidents,
  not exterrsive. Similarly, althoughtherc is much scietttific      informationon the associ-
  ertionbetweenalcohol anclnrottir vehicleacciclcnts,      little is known at a highly spe9ifi1
  level,about the  effectsof alcohol on many o[ thc segments bchavior that may lead
  to theseand    other irccidents. adriition,although thereare various restrictions
                                                                            the.prccisc        if
  thc licensingof individualswith certain ,.nto.y-i*puirments,
  any, of ,u"h*i,npairments accidcntratesremain
                              on                          tobe definedand mcasurcd,and it
                                                    "cqmmon scnsc" measureswill prove
  is'tikely that at least some of the ctrrrcnt
  inappropriate whcn l'heir merits ate ptopcrly studied'
  ' -'fi*.p,t"
               the potential virluc of psychologi*alapproaches,,loy,"u:.,psychological
  studieshirectly relatedto accidcnt,s    huu* not gcncrallylreen highly fruitful' In terms
  of practical uulue,it $eems   probablc that work in visuirl perceptionhas had thc most
          ond       application-to tte3ign.,of
 prolitablc widespread           thc        g-l9l-lll                       lll. :1.'1':::
                                                                       of various items of
 lolors for signs,aircrirft, unintn*r equipr'cnt, and io the design
 military ancl"civilianharclware  (seeMiFarland in Chap, 2 and rcfcrence4). However,
                     cvidcncs'f the potential contrib*tion of approaches     outsideof the
 there ii increasing
 nsvchophysjcalarea. This is     reflectedin scvelal of the papcrs that foltow and, fbr
                                                                               'f groups of
 5;i,;;#. 1;;-.*,rt *o.t rclati.g psychol.gicaland other charactcristics
 driveis to thcir ittsurer's subscqucntclairn expericnce'5
                              our understancling the relationship
                                                 of                   bctween    personality
  charrrctcristicsand iccidents is still serionslylimited.    For exatnplc,the accident
                                                                  ,subjected. adequately
  experienceof patients in psychotherapyhas not yct becn
                            study; hencelhcrc is little eviclencefrotn psychotherapists  as
  oyit*maticand cxtcnsive
                                    emotional stateslead to accidents.     Other work on
  tL the extent to which sp""ifi"
  personality factorshas focuscdon various prcsurnably        stablecharactcrtlirits in an
                                       "safe" diiver, pilot, rnachine.operstor, al-Bttt
  utt*rnptto clevelop profileof the
                           practical ancltheoreticallimitrrtions.Although,     for cxample,
  tt i, upp.oo.}, has'bot-h
  it is somctinresfeasible a military service, industrialorganization' a commQn
                          for                   an                            or
                                             screening   in pcrsonnelselcction    whereit is
  carrier to usc thrs type of psychol<igical
PSYCHOLOCICAL APPROACHES                                                                                                                                            29I

possibleto reject large percentages individually safe applicantswhose character-
isticscannotbe differentiated     from thoseof groupswith elevated   rates,it is extremcly
doubtful that thc Amcricanpublic would countenance useof pcrsonality
                                                            the                    testsa$
part of the driver licensing     procedurc,particularlysincethesetestscurrently show
little reliability in predicting individualaccidcnl.  experience.0
                                                                 Perhapsmore seriously
 limiting is thc fact that this researchusually:( l) ornitsadcquatc scrutinyof variations
in the exposure hirzardof those studicd; (2) has failed to discriminatcbctween
characteristics   that are stablein time and those that are transientor situational:
(3) doesnot attempt to verify its resultsby rcplicationwith other populations;and
(4) doesnot demonstrate                             of
                              that the perccn(ages individualswho woulclhavc to be
restrictedin order to achievea given reductionin acciclents         makes psychological
            sufficiently "cconomical" to
screening                                  bc politicallyand sociallyacceptable.o
     A relatedand particularlydirngerous      misuseof psychological    irpproachcs tois
justify in a logicallycircular fashionadministrative     actionsagainstinclivicluals who,
for whatever    reason, havebeeninvolvedin repeated      accidents. the prcscntstateof
our knowlcdgc,we cannot reliably attribute an individual'saccidenthistory to his
psychological    characteristics, hencethereare no groundsfor basingadministra-
tive action upon psyclrological      thcory or for attemptingto validate psychological
hypotheses     on the basis of administrativclyaccumulatedaccident histories.As
Daniel P. Moynihan has pointed out:

         It is particularly benoted
                         to       thattheadministration traffic
                                                     of                     ignored
   the statistical laws that govern the random distribution of comparatively infrcquent cvents
   alnong a large nunrber ofpersons. (Thcre is only onc fatality for 2000 years ofaverage tlriving
   e x p e r i e n c e .T h e s el a w s d i c t a t e ,o f c o u r s e ,t h a t a r a t h e r l a r g e s h a r c o f t h c c v e n t sw i l l ( ) c c u rt o
   a r a t h c r s m a l l p o r t i o n o f t h e g r o u p . T h i s r t i s e s t h e p o s s i b i l i t y ,e v e n l h e l i k e l i h o o d , t h a t r n a n y
   of thc                       violators" of thc traffic laws irre innoccnt victitns of the Poisson tlistribution
   whose misfortunes have bccn compounilcd bv a statistically illitcrate bureirucracy.T

    Another obstacleto productive psychologicalresearchstemsfrom the fact that,
unlikc many other kinds of human hchavior,risk-takingor the precipitation acci-
dents cannot usually be clicited or effectivelysimulatedunder laboratory conditions.
The kind of risk-taking   behaviorthat can easilybe studiedin the laboratoryconsists
essentially the cirlculation probabilityand the alterationof decision-making
                               of                                                  by
various social proccsscs.  But such laboratory behavior rlay lack the spontaneity,the
motivation,the social climate, the "punishment," and rlany other elements        that
itrfluencerisk-takingand acciclcnt   precipitation a naturalsctting.
                                                  in                 This raises
a point emphasized Clhapter
                     in           3-narncly, the impoltanceof studyingaccidents  and
their relatcdphenomena the real world in which they occur and not merelyunder
thc necessarily artificialconditionsof thc laboratory.
    The readingsselected    for this chaptcr rcflcct in many ways the points we have
made in the foregoingdiscussion. gcncral,their virluelies more in the potentials
they point to than in the excellence their methodology the validity of their con-
                                      of                   or
clusions.Thus, theyare more usefulin pointingout general               for
                                                             directions further a rd
morc rigorous research   than in providing findingsthat can be usedas a firm basisfor
action programs.
292                                                    PSYCHOLOGICAL APPROACHES

PsycHOI,III,TVTIC PPROA     CHES                                          "the
     Frcud,s emphasison the role of $ubconsciousfactors in what he termed
                                                      The method employed for
ir.,'d.sp..iolty"u*ong Iayr*n concernedwith accidents.
                                                                  has, however,
demon,strating impoitance of such factors in accidentcau$ation
progressed   but little since Freud's work'

  * Charlesllrennet, M.D,

    It is not the purposehere to discuss &dequacy.
                                             thc          ofthe scientificcvidcnceunder-
lyint psychoana'tytl"    theory. However-, must.bepointed llt th.utits applications
                                                                        rigorous and sys-
accidcnt causationtrune yit to be substantiallysupported^by.the
ternaticresr:arch   that is regardedas csscntiatin coilnteral fields' The cviclcucc
which is esscntiallyanecdltal, is usuatly derived       post hoc under conditions which
                                                                rcporting of information
might be expectedselectivclyto favor the production rrncl
c-oilfatitrre *itt, ttr. "on"*ptu,,l framework    cmployed.Casc serieshave been statisti-
                         Contiols havebccn virtuaily nonexistent'   The-data  havescldom
cattytrigtrty  selectccl.
 been val.idatecl  everrwhen indepenclcnt      sourceshave been available' Psychoanrtlytic
                                                                                    This is
 interpretationshave beenfavorc.l to the exclusionof alternativeexplanations'
                                  likelihoocl  that someaccidents are initiatedby psycho-
 not to deny the consiclcrable
 dynamic   factors or to suggest   that the theoriesthcmsclvcs incorrcct' Rather' it is
 tJ p.int out thc **oknerlJ, ofthe rcscarch              supporting presumplionthat
                                                  evidence           the
 such factors contrihute to accidents.    Thcse characterisl"icsare well illustratcd by the
 sclection  that follows.

 LETus Now CoNSIDER classofparapraxes
                         the                        was struck by lightning during an electrical
 which are ordinarily referredto as accidental      storm. we should ordinarily be quite confi-
 mishaps,whcther the ntishirp occurs to one-        dent that the mishap was truly accidental
 self or to anotlrer as the result of onc's own     and could not possibly have bccn utlcon-
 '*carelessness." must make it clcar at the
                  Wc                                sciously intendod by the victin-r' After all'
 outset that the only accidentswith which we        who can tell where lightning will strike'l
 are hcrc concernedare thosc which the sub-         Howevcr, if we letrrn thrt the rictirn was
 ject causedby his own actions, although he         sitting under a tall, solitary tree next to a
 iad. of course,no cansclotr.lintention to do so'   heavy, steel chain that dangled frorn one of
  A mishapwhichisheyondthesubject's      control    the branches to within a few fbct of the
  is ofno interest usin our present
                 to                   discussion'   ground, then we nright as well begin to
     It is often easy to decide whether the         wonder whether the victim was or was not
  subject was tesponsible for the firishap          aware, bEforc the accident, of the relatively
  under consideration, but it is by no means         great danger that a petson in such a situation
  always such a simPle matter to do so,             will in fact be struck by lightning. lf wc then
  For example, if we are told that someone           discover that this was well known to the

       fRenrinted. with perrnission,from An Elementarv Taxtbook in
       li;d-i;;: N-; yoik: Doubletlay,Anchor, t957.Copyright                            )I
       I                                    Press.
 PARAPRAXES           AND WIT                                                                             293

 victim and if, having recovered from his      will lead to an understanding of his uncon-
 mishap,hc honestlydisclaimsany conscious      scious motives for causingthe mishap that
 intent to endangerhis life, we must conclude             at
                                               secrned first glanceto be quitc accidentirl.
 that this particular victim of lightning was  It happens not infrequently that, in the
 deliberutcly,  tbough unconsciously,          courseof the analysisof such a mishap,the
                                         trying to
                                               subject recallsthat hc knew for a moment
 gct it to strike him. In the sanrewily. an iruto-
 mobile accident may be due to a purely        that the                       was going to happen,
 mechanical failr.rrearrd have nothing what-   just before he pcrformed thc action that pro-
 ever to do with the driver's unconscious      duccd it. Obviously, lrc could know such a
 intent, or it may, on the other hand, have    thing bcfore the fact only if he intendedthat
 been either dircctly causedor made possible   it should happen.This partial awareness                        of
 by unconsciouslyintentional acts of com-      intent is usuallyreprcsscd,            that is, forgotten,
 mission or omission by the driver.            during or just after thc mishap and is oniy
    The reader may ask whether we propose      restoredto consciousnremory if the mishap
the view that every mishap that t:ould have    is analyzcd,Thus, without analysisthc sub-
 beencausedor lacilitatedby an unconscious    ject himself usuallyis quite convincecl the                of
 intent on the part of the subjectwas in fact  purely accidcntalnature of thc rnishapthat
 so causcd. ls therc to be no room left flor   in fact he himself intentionally caused.
 human imperfection?Arc wc to assume,for          Naturally it is in the courseof psychoana'
 instance, that no one would evcr have an      lytic therapythat the opportunity arises                   most
 automobile accidcnt unlesshc unconsciously    often for studying such rnishaps                  dircctly, as
 wanted to?                                    opposedto merclyspcculating                 about them in
    The answerto this qucstionis, in principle a more or lessconvincingway on thc basisof
 an unequivocal                                external. circumstantial evidence.Most of
                    one. Insofar as a foreseeable
 mishapis caused a "human imperfection"
                      by                       o u r e x a m p l c sw i l l c o n s c q u c r r t l b c d r a w n
in thc performancc of some action or other,   from this source!though such rnishapsare
 we assume    that it was unconsciously       by no means more frequent in the lives of
 by the performero[ thirt action.It is true, of                     patientsthan they are in the
 course, that fatigue, boredom induced by     lives of other persons,
monotony, and other, similar factors ntay         On one occasiona patient, while driving
increase   the frequencyof such mishapsto a   to work, was making a left turn at a fairly
greater or lessextent, but we are here in the busy intersection.                      of
                                                                          Because the number of
same position as that which we took with      pedcstrians who were crossing, he had
rcspect to slips of the pen or tongue. The    slowedto a speedof about five milesan hour
necessary   condition for a rnishapof this sort,
                                              when he suddenlystruck arrelderlyrlan with
which is often a sufficient condition as well,his left, front fender trnd knocked him to the
is an unconsciousintent to produce it. Fa-    ground. As lar as the patient was aware
tigue,boredorn,etc.,are merelyaccessory       when he Iirst told thc story of the mishap,he
adjuvant factors.                             had not seenthe man at all, Later. however.
   If thc readcr now asks how we can be so    he was able to recall that he was not sur-
slrre that mishapswithin the control of the   prised when he felt his car hit something.In
subject were in frrct unconsciouslyproduced   other words, he was dimly aware of his
by him, our answerrnustbe that this conclu-   unconsciousintent to strikc the rrran with
sion is a generalization   which has been mrrde
                                              his fenderat the moment of thc
on the basis of those casesof such mishaos    On the basis his associations the various
                                                              of                           to
 hich havebeenaccessible directstudy. circuntstances what had happenedit was
                            to                                    of
Hcre again, as in the case of other para- possihlc to discover that the chief, uncon-
      , direct study meansthc applicationof   scious motive for the mishap was the pa-
hc psychoanalytic  technique.lfthe subject's tient's wish to destroyhis father. In fact, his
     ri.ltioncan be obtained.his associations father had bccn dead for a number of vears.
                                                            FSYCHOLOCICAL             APPROACHES

but the wish was one which had been most               sincc in this instance it is not easy to distin'
.rctive   rlilring the patient'soedipalphase,   had guish with ccrtaitrty bctween thc two'
                        repressed that timc and
                                   at                     Unconscious activity of the superego fre-
bccn energetictrlly                                                                                causing
had livcd on in his id thenccforth'        We can quently plays an important part in
                                                                                 sort. Many nrishaps are
unclerstantl     that this wish was displacedin pa.ap.r.*.t of this
                                                                            intended to rtsult in loss or
the wa1,that is characteristicfor the primary unconsc,iously
                                                                        In the motivation of such cases
 oro.*r, onto an unknown, cldcrly rnan who self-injury.
                                                          large role    is playcd by an unconsciousneed
was in thc p+th of the paticnt's car and who a
                                                             punishment, lor sacrifice,or ftrr rnaking
 theretirre became the victim of what was for
                                                                       for some previous act or wish' All
 apparentlyan accidcnt'It is understandable restitution
                                                            these rnotives belong to the supcrcgo' as
 aiso tttat dcspitc the fact that thc victim            of
                                                             reader will rernember.
 sustaincd no injrrries and that the patient the
              was fully insured, he nevertheless           As an exarnpleof such motivation we may
                                                              the fotkrwing case. The patient t:rf our
  felt both frightcncd and guilty to a degrce cite
                                                                                            the right liont
  that was considerably       out of proportron to first examPle one day drove
                                                                 ol  his car againstthe corner ofa curb-
  the actually trivial nature of thc accident' wheel
                                                                                           park with such
   Krrowingthe unconscious        motiveswhich led stonc while attcmpting to
                                                                                                of the tire
  1o his knocking the      man down, we can real- force as to tcar the sidcwall
                                                                    repair. It is uncotnmon for an ex-
  ize that it was these rrrolive$which were the beyond
                                                        peiienccd driver to have such an accident
   more important sources the pirtient'ssub-
                                                                                           more surprtstng
   sequenlguilt fears'     In other words, his reac' and this one was irll the
                                                                    it occurred at the curb in front of the
   tion to tlre accident was only apptrently a bccausc
                                                                      61vn fi1ruse'where he had parked
   disproportiorrate ' lt was quite in propor- oatient's
                                                                  times beforc without incidcnt' How-
   tion to his repressedwish to destroy his many
                                                         ,u"a. hit tssociation furnished tlre explana-
                                                                At the tirne of the mishap he wtts re-
        Another example, which is so trivial that tion.
                                                                     from a visit to his grandfather's
    it hardly dcscrvcs to bc called a rnishap, is turfling
    one which we mentioned in Chapter            I' In hrruse on thc rrroming trftcr the latter had
                                                                followirrg irn illness of scveral months'
    that case a young man, driving to his fian' died
                                                                               the paticnt ielt guilty as a
    c€e's home on the morning of his wedding' Urconsciously                                                of
     str)ppedat a green traffic light   and was not result of his granclfirthcr'sdcath bccause
                                                                      hostile   wishestoward the old man,
     uwoie of his mistake until after it had his own
                                                          wishesthat were to a considerubledegreethe
     changedto red, In this casethe driver's asso'
                                                                             of similar, uncotrsciouswishes
     ciationsled to thc discovcryof uncotrsctous counterparts
                                                                     his own fnthcr. 11* 56ashcd the tire
     leelings of reluctance to go ahead with his toward
                                                  guilt on his own car to satisfy the unconsciousdc-
     *urriug* which were chiefly due to the
                                                                                             hc be prrnishcd
     and fear cotrtrectcd   with ccrtain unconscious mancl of his supercgo that
                                                          for having, in his unconscious      fantasy, willcd
     sexutrlfantasiesof tr sadisticand ince$tuous'
                                                           his grandfather'scletrth.
     that is, oediPalnaturc.
                                                                              such a rrtishap combines both
         In the first of the two examplcswhich we            Sornetirnes
      havej ust given the mishap was d    ue to inade' the crirne and the punisllment' We may
                                                           pcct, lbr instancc'     thnt in thc cxarnple just
      cluat; or incompleterepressiort a hostile'
                                                           givcu, some reprcsscdlirntasy of smashing
      ii impulse. The id impulse in question es-
                                                                                                 or symbolic
      capccl in part fronr repression'   as it is ofterr ilis father achieved a displaced
                                      writings' ln the gratificationin the         patient'saction of smash-
      exprcssed psychoanalytic
                                                           ing his car againstthe curb. In this particular
      sec.rnd e^ample the parapraxis was the                                                   paticnt's asso-
       result of either a deferrse  againstcertain id -*ir.1rlc, as it happcncd, thc
                                                           ciations did      not point in that direction, so
       impulses or of a superegoprohibitiorr-direc-
                                                           that we are left with no more than a suspi-
       ted against thcm, or evcn, pcrhaps, of both'
FARAPRAXES        AND   WIT                                                              295

cion or conjecture. However, in other pases,    Smashing up his car was an unconscious
thcre is no doubt of the fact that crime and    expression this anger,which she was un-
punishment are both containcd in a single       able to display openly and tlirectly against
action.                                         him. For another thing, she fett very guilty
   For instance,a patient, while driving her    as a rcsult of what shc unconsciously   wanted
husband's car, stopped so suddenly in traflic   to do to her husband in her rrrgeat him and
that the cat behind her crumpled one of the     damaging his car was an excellentway to get
rear fcndcrs of thc car she was in. The ana-    hirn to punish her. As soon as thc accident
lysis of this mishap revealeda conrplicatcd     happened,she knew shc was "in for it." For
set of unconscious motivcs. Apparently          a third thing, the patient had strong sexual
three different, though related ones were       desiresrvhich hcr husband was unable to
present, For one thing, the patient wds ufl.    satisfy and rvhich she hersclf had strongly
consciously very angry at her husband be-       represscd.Thesc unconscious,scxnal wishes
cause of the way he mistreated her. As she      were symbolicatly   gratifiedby having a man
put it, he was always shoving her around.       "bang
                                                       into [her] tail," as she put it.

     Psychoanalytic  explanationsof accidentcausationare of sulficientinterest,influ-
ence,itnd plausibility to justify thcir scientificevaluation.Howcver, the cost and
duration of the pr:ocess psychoanalysis, special
                         of                  the         characteristics patie
                                                                          of      nts, arrd
the clinicalolicntationof psychoanal_ysts createda widespread
                                            have                          imprcssion  that
rescarch more diffic:ult psychoanalysis
                            in                 than it is in rnany othcr liekls. Despite
thcse difficulties,however,there are rnany possibilitics thc scicntilicstudy crfacci-
dents fronr a psychoanalytic    point of view. For example,& range of irppropriately
chosen  individuals might be diagnosed   psychoanalytically ileans of thc Rorschach,
T'A'T., or other projectivetest and then followed to determinewhethcr their sub-
sequentaccidcntcxpct'ience     bore a relationship their psychodynantic
                                                   to                           character*
istics.HowcvLrr,  cvena high degreeof association  revealed such a stuciywould not
cll'ilsclIconstituteadequate  proof. It would be necessary, addition,to drmonstrate
that the observed   associations bctwecnthe presence absence'of
                                                         ar               certain psycho-
dynamic patterns and subsequent        acciclcntcxpcricnccrrecessarily     followed from
psychoanalytic   formulations,that other cxplanations     did nol. fit the facts,and that
thc.Iindings wereconfirmed whcn the samemethodswereapplieclto other groups.

llrllavloa rN RELATIoNrtr
THE ECOLOGY OI' DAIICTT<S                       :

    In sharp contra$tto the approaches Frcud and his discipleshas been the work
of experimcntalpsychologists   using scientificmcthods in the study of a widc valicty
of preciselydefinableaccidentvariirbles.There have been excellentinvestigations      of
perceptionand reac-tion titnc as they are affectedby suchfactors as drugs irnd firtigue
(seeLoomis ancl West in Chap. 3 and refclcnce2). Most psychologically        oricntcd
accident rcseilrch,hou'ever, is narrowly focused: it views accidentsin substirntial
isolationfrom the characteristics the environments which they occurand it fails
                                   of                 in
to provide an adequateconceptualfrumework within which the bload gamut of
accidenttypes and circumstances     ma-ybe rationally approachcd.
THE CONTRIBUTION OF EXPERIMENTAL                                                                                                               T
pivtHolocY     To rHE FoRMULATIoN oF THE                                                                                                        I
                                 BASIC RESEARCH
           OF                                                                                                                                   I
   -Jumes J. Gih'son,
     Theselectionthatfollowsisanexceptiontothisgeneralruleandis            p*ythT
                               to                 Cibson, experimenFf
                  contrib;ii;ns accidc'f[itcraturc.        an
                                             inretarion "the
mostsisnificant                                               ecologvda"g:t::' I
                                  c'f                  ro
ffiil, ffi;;;,hdffii;;alysis refreshing  departure both from purely Statistical I
His formulat,nnr*p,*,*nts a                                       of onetffie and I
                                                    with var.iables
proaches from tn. nur.o* studythat dlals only
                                                                 tTl               I
                       contcxt' He'e we haveconvincing   eviclence
 ncglects ecologic
           the                                                        l,t:-^-::Tl I
                                 cango far bcyondthe standard reaction-time
 f"ir-pi""if".torr"in accidents                   popularopinionnotwithstanding'    I
 ment,whicht',urprou*A-"markably-urrprotluctive'                                    l

Ar-r rxpgnlMeNTAt PsYcHoLoGIST is con-                                   is fearsome, or properly to be avoided' in the
cerne<lwith the bchavior of men and oLher                                environtnent of the individuals in cluestion'
;llinrrrls relative to thcir environment. He                                 Man is a terrestrial animal; he breathesair
h^s no particular tnowtetlge atout thc spe-                              an<1resists gravity and gets about on the
.*f f*ff**      of acciclentp-reventi.n antl_no                          surface of the earth. He carrnot live in
pr*.irc-i rccornrnendations to mske. But                                 mcdium      of water or mud, and he is not
since he is intcrcsted in motivated behavior                              naturally cquippcd to swim, burrow, or fly'
unJ p...*ptiun irncllearning. hc docs have a,                             He rrdapts his behavior to thc pcrmancnt
 way .f t'orrnularing th. ;;;;;;i    problcrn of                          features of thc earth's surface-the local
 adaptive behavior, or ad]ustment, .rl yh_"I                              terrain. I{e rcsponds, also, to the flux of
 ever you like to call it. This can be helpful in                         energieswhich sutround him-gravitational
 conceivirrgclearly what wc should mcan by                                antl mechanical, rirdiant, thermal' and che-
 sa/ety and 6l6nger,Thepracticirl busincss of                             mical, some timited fields and ranges of
 in".*u*ins the formcr and decreasing the                                 energy provide stimuli for. his senseofgans;
 iutt., A**p*nasupo' having a cleai and                                    others induce physiological adjustments;
 .*pii.i, tireory to guide us.                                             still othcrs produce injury'-These last etrergy
     The experimental psychologist makes ex-                               exchangesotc propetly called dangeroils'
  periments, .f course, mostly iri a )irb.ratory,                          tlungcr is herc dcfined.as an external source
  But what hc aims for is a behavior theory                                of potential injury. tnjury rcsr'rltg, course,

  that is valid f.r the whole cnvironment, not                             o"i' if thete occttts what I wilt call an en'
  nrerely for the laborrrtory. He isolirtes. and                           rcunter'
  controls abstract features of the errviron-                                The geographical environment of tnan'
                                                                                                                              v i r own cf-
 m e n t , a n d h i s e x p e r i m e n t s a r e m e a n i n g f u l i n c l u d i n g t h e l o c a l e n e r g y e nhis o n m e n t ,
                                                                                  considerably modificd by
 only to the extent that the featuresisolated been
                                                                                  rluring the pa$t few. thousand years'
 for study arc gc'uine or iypical. Hcncc he forts
                                                                                   has bcen earth-rnoving and construc-
 ought to be a student of the environment a$ There
                                                                                  and provisitln fttr artificial light' and
 well as ol behavior. He shorrld be something tion,
                                                                                          thc paths' bridgcs' fcnccs' shelters'
 performs experimentson rtr" aeuttopin-nt of and large'
                                                                                                                                     a other
 f e a r a n d a v o i d a n c c i n c h i l d r e n a n d a n i n r a l s , t o o l s ' m a c h i n e s ' c l o t h i n g ' r a d iallt o r s ' f o
                                                                         water faucets, conditioners'and
  but he orrghtalso to bc.onr*.n*d with what

THE CONTRIBUTION OF EXPERIMENTAL PSYCHOLOGY                                                   297

goods of the artificial environment have alle-      more impprtant. (ln either case impact with
viated old sourccso{'danger, but they have          the ground is the causeof injury. Impact can
also introduced new ones, Human children            be measured rate ofchange ofvelocity, or
are now in lessdanger frorn rock-falls but in       amount of deccleration.) falting-ofT
                                                                               A             place,
rrrore danger frorn electrical outlets.             such as a cliff or steepslope of ground or a
   The natural and the artificial environ-          stairway, is a particular danger. Also, a child
ments comprise what might be called the             may collide with an obstacleduring loco-
objectivccnvironment.But this is not every-         motion,
thing. Actually it is only a basic framework           2. Passive impact. One may be struck by a
for the animate environment and thc socral          falling object:a rock, an avalanche, plaster
environment-that is, the world of other             ceiling, or, for that matter, a meteorite.Or
anirlirl$ and other Fersons.These arc also          one may be struck by a moving object: a
sourcesof danger, Animals can bite, claw,           moving automobile, an explodcd chunk of
crush, or sting. And other personscan be            somcthing,or a missile.Thesc constitutc a
cven mere deadly,espccially      when equippcd      secondtype of danger, but thcy arc irrtpcr-
with weapons,or when they misuse tools,             manent rather than permanent featuresof
machines,or vehicles.                               the environment.   They arc caused melting
   From this point of view, dangersare en-          snow, say, or by a carclcssor angry man,
vironmentalfacts.Theoretically is possible
                                    it              They do not stay put, like an uncovered    well
to locateand specifyall sources     ofdanger in     or a tree in the nriddleof a path.
the geographical,   the artificial,the anirrtatc,      3. Interference  with brcathing,It is poss-
and the socialenvironments. excludefrom
                                  I                 ible for a terrcstrialanirlal, young or old, to
consideration  discase. at leastinfection,as
                        or                          encounter the wrong meclium and to be
a sourceofdanger bccause is a special
                              it            type    sulJbcated a solid or drowned by a liquid.
of dangerwith it$ own peculiaritics. exclude
                                       I            Sand banksand bodiesofwatcr are local and
also starvation becauseit is another well-          specificdangers,
definedphenomcnon,      fuirly well understood.        4. Tool and machine forces. We primates
We are prirnarilyconcerned      with safcty,and,    are cquipped for nranipulrrtionas well as
in ordinary speech,  this is ciistinguishedfrom     locomotion. We have constructcd tools.
health and nutrition. Allowing for these            varying from stone chiselsto stamping mtr-
omissions,let us corrsiderdirngersin more           chines, which can shcirr tissue as well as
detail.                                             wood and metal. For the urtskilleduser or
                                                    the child, thc handling of a tool involves
A ClessrtrcAttot*top DeHcEns                        dangcr as well as achievement,     with a thin
  Injuries to a living organismcan be pro-          line betweenthc two,
duce<i only by some energy interchange-                5, Machine failures. As a result of the
Conseqr,rcntly, most cffectiveway of classi-
               a                                    complcx and elaboratemechanical       environ-
fying sourcesof injury is according to the          ment, and all the ways in which devices    can
forms of physicalencrgyinvolved.The ana-            break, cxplode, or otherwise fail, a vast
lysiscan thus be exhaustivc and conceptually        array of dangerssurroundsus all, Machincs
     . Physicalenergy is either rnechanical,        are designednot to fail, maintenanceis in-
                                                                                       "safcty de-
 hermal, radiant, chemical,or electrical.           tended to prevent failure, and
  tvlechary4a! .E!l!!E!. Physical cvents such       vices" are installedin order to  signalfailure.
    impact, breakage, shear, and flow (or           Nevertheless,  flywhcels and mower blades   do
       ge offlow) all involve the expenditure       brcak, tires do blow out, and airplanesdo
   mechanicalenergy.The cncountcrsof a              corneapart in flight. But the
   ld with his cnvironrlent include seven           in machine breakdown is thcoreticallycap-
 ypesol suchevents,                                 able of bcing eliminated.The great sourceof
   l. Active impact. An active terrestrial ani-     danger from machines stems from the
     , infant, child. or adult niay either fall     relation bctween them and their opera-
         " or fall "off." Thc latter case is the    tors.
298                                                        PSYCHOLOCICAI         LAPPROAC
    6. Animal forces' Other speciesthan our          rnust learn new modes ofl avoiding them' I
own are cquippedto injurt irl varioussimple Oftcn enough they must be identified bf I
mechanicalways, Thcy can hite, claw, kick,           signs or syrnbols.                                 I
butt, stirrg.   ttnd so on. Encounters    with ani-     tJ"fyl:4E:::El.'A       lethalflow of electric
mals, cspec:ially     with thr,rsc we loosely call curreirt-is rare in tirc tratural environtnent'I
"wild," constitutc dangers.                          To be struck by a bolt ol lightning is thc I
                                                                                  "act of Cod," or an
    7, Weapou-produced         forces. Our fellow- classical    cxampleof irn                           I
men have irrvented1ools for doing injury-            unpredictableaccidcnt.But nowtdays cur- |
that is, weapons. An armetl rnan fairly radi- rent outlets, live wires, i'tntlso-callccl                 I
ates danger. A knife or an axe is lethat hazards"are prcvalent.'l'hcyare not always I
enouglr,but the weaponsof high civilization easyto identify as srrclr,but they differ from I
will cut, ()r pcnetrate,or explodc at a dis- Iightning in an important way: the locus of I
tance, Year by year, the distLrrrce     increases. danger is fixed, wl'tcrcas the locLrtion of I
                                                                                                     of I
     Thermal Ener&!'; We are animals with a lightning is variable. Thc occurrencc
closely regulatcdtenrpcrature,trnd       we cirnnot shock at a cument sourcc is certtin. The J
withstantt extreme or prolonged heat gain or         occurrsnce oi Iightning at a givelr place hirs
 heat [oss. Extremetemperatures the sur- only a
 rounding medium         will darnagc us, and so
 will very hot or cold substances'         Cltildren     To sum up, environrrrentaldangerscan be
 have alwrrys     had to lcarn  just how closethey analyzed and classified. They are physical
                                                                                   tnore claborate ana-
 can coile to ir firc. Bctween warming and facts or events,A mttch
                                                                              have given, but an eco-
 burning, the gri',dicnt is fairly steep' Nowa- lysis is neetledthan I
                                                                                      An esscntial part
 dnys children arc having ttl learn the ana- Iogy of daugcr is possible.
 Iogous rule for      clry icc. The entities in the of the ritruggle fttr existenc,;ovsl thE agcs
 world that cau5gburn or frostbite arc dan' was the avoidancc             or the ovcrcomingofsuch
 gerous. Visible llanresirrc casy to     detect, but dangers,   Falls,collisions,   immersions,  blows,
 some of      thcse entities are not. Explosions, cuts,bites,burns,and poisonswcre what our
 for exarlplc, which often cotrrbine       heat and anccstorseither survived or, fiorc clevcrly,
 impact, are notoriousty         unpredictablc.       averted, Dangers werc agents of natural
                Enet|,y:The radiarrt energy of the selection. They wcre also objccts of discri
 $un, as filtered by the atmosphcre,is gcne- minative perceptiotrarrd selectivelearning,
  rally not dangerous.Apart from overexpo' They must have been closely attendcd to.
  surc to ultmviolct and to radirrntheat-/.e.' Thcy governed the evolutiorr of the capacity
  sunburn and !;unstrokc-the clanger radi- for looking and listening, I'hey motivated
  ant eriergy comes ftom nranmadestturcesof the trainirrg of children. They gnve rise to
  gamma rirys, Atomic errcrgy in war, how- rules and laws.                          led to cooperative
  everl we had better       leavc out of considera* effort. In the modcrn world, so firr as
  tion. Even tnore than starvationand disease, possible, we try to clirninrrtethem from the
  this fearsomethreat is unique.                      environment, Whether cornplete         climination
      Chemical Energy:       We call injurious che- is possible (or even desirnble)is another
  micals poisons' They can be absorbed by qucstion. In any ca.se is academic,for ncw
  breatlring, by skin contact,      by eLrting, by dirngerssecm to arise witlr evety nev inven-
  drinking. Thcy existed in the nirturul envi- tion ofl our techtrological            culture, and tlrcir
                                                                                           par:t of the
  ronment of tlur anccstors' and they still alleviation is an cssential
  exist.   Prinritive mcn developed,or lcarned technology.
   and taught, ways of distinguishing       them by
                                                       Explonrronv BFHAvIoR
   sight, srnell,or taste,nnd then avoidedcon-
                                                       AND MARCINS SAFETY
   tact with thcm or rejecteilthem as food' But
   ncw poisons cotltinually appear           irr our      Now for the second assumption in my
                                                                                      mirrgin ol' safety'
   manufacturcd envirorrtrrcttt,and children line of rcasoning-the

  Children, like nearly all anirnals. are            and behavior. Psychotogicalresearchbased
  strongly inclined toward locomotion and,           on this assumptionis scanty.hrrt I am here
  like all prinrates, are chronically bcnt on        suggesting       that it should he encouraged.
  manipulation. They tcnd to explore the                Thc simplcst situation of this sort is the
  environment and to investigatethe possibili-       cdge of a clilT', a brink. Any terrcstrial
  ties of ohjects. They get about and they pry       anirlul rvhose center of gravity passesthe
  into things, During such exploration they          brink by a hairsbreadthwill full. LJp to the
  learn about places (place-learninghas been         brink, no matter how closc.it will not fall.
  much studied by cxpcrimental psycho-               So long as a rlargin is rcscrved,                thc animal
  logists), thcy lcarn about objects (object-        is safe.Obviously,the visual discrimination
  identi{ication has been investigated), and         of this margin is important to mor.rntain
  they learn about events(a kind ofl learning        goats and to sttucturul steel workers. An
  which is only beginningto be studied).Sorne        experinrctrtal      study of hrink-avoidance+                 has
  placcs,somc objccts,and someeventsin the           reccntly been finished at Cornell, using as
  cnvironment arc datrgcrous.    Hencechildren       s u b j c c t sm a n y k i n d s o f a n i m a l s ,i n c l u d i n g
  necessarily   and unavoidilbly corrrec)osc to      babies.The device used for all species                        was
  certain kinds of dange       r, Lor;omotion,       callcd a "visual cliff." Another, more com-
  rnanipulation,play, invcstigation,  trial-and-     plcx, cxanrple can be given with regard to
  error, insight, and problcnr-solving   involve     powered locomotion. In autonrobiledriving
  the skirting ol dangers.Evcn so simple an          there is at evcry instant a minimum braking
  act ns running involves(a) not falling down,       eoae which is fixed by sevcral deterrninants.
  (6) not falling into holes and ditches, (t.) not   Most notably, it increascswith speed,At
  colliding with obstacles or with other             every instant thcre is also a /e/r/ o,f .free
  individuals. It requires choice of footing,        lrolel which is lrxed by road, obstacles,                     and
  selection of path, steering, and stopping at       curve (if present).If the minimum braking
   the goal instead of bumping into it. The          zone is ever allowed to exceed the field of
   dangersof irnpact have to be perceived,and        travel by thc slightestarnount, thc driver
  the locomotor action has to be adjusted            will crash. Thc ratio of thc two fields, or
  continuouslyta the variousgapsor distances         spaces,rnu$t never reach unity. This is the
   betweer the body and the solid world              first law for control of a wheeled vehicle.
   around it. This is what I call practical          Sornewhatsimilar laws can hc formulated
   spacc-r)erception,                                for the landing of an ncrial vehicle or, for
     Thc wold "margin" means (according to           that ilatter, the landing of a hird on a perch,
  the dictionary) not only an cdge, brink, or        or a bee on a flower. The essence the                    of
  limit but also an extra space, a gap, or a         matter is that the operntormust so act :rsto
  difference.In this latter senseofthe terrn we      maintain a certain physicalmargin between
           of Lht norgin o.f safety which cxists
I speak                                              two changing stimuli, or two stirnulus
               pefsrln and the edgc of danger.
I between a                                          variables.The birds and the bccs can do it.
I Psychologists    n,ho study skill now realize      and it is not surprising that human beings
          gap, a discrcpancy
I that a                     which changes ith
                                             u       love to practice doing it, starting with tri-
Itime,   is a governingstimulusfor alt kinds of      cyclesand ending with aircraft. Obviously,
  purposive behavior. The responseof the             the two stirnuli must be detrcted or. better.
Incxt momcnt is controllcd by thc gap of the         the gap betweenthem must be pcrceived,
  p.csentnlonrent.One type of gap, the error            It should be notcd that the nrargin-of-
lor mismatch, has to be minimized, and               safetyconceptinvolveseither or both of two
lcontrol devices, servomechanisms, be
                    or                    can        kinds of gaps-tbe closenes$ danger in        of
ldesignedto do this. But anothertype ofgap,          spirct and thc irnnrincrrce flangs1in time.
fthe margin of safety, has to be maintained,         Often enor.rgh       there exists stimulus inforrta-
lThis is a wholly differentsort of discrepancy,
       it is equally a stimulus lor perception       t Sce"The VisualCliff," Chrp, 3, Eds.
300                                                     PSYCHOLOGICAL          APPROACHES

 tion in light, sound, or odor for imminence        events should not prevent us from studying
 as well as for closeness.  The signsof dangcr      the perrnancnt, fixcd, and perfectly predict-
 may be subtle and easily overlooked but           able dangers that surround us. The ecology
 physically present nevertheless.When they          of danger and the psychology of safety
 cannot bc detected by thc unaidcd senses,         rrrirrginsdo not haveto be based on statis-
 thcy must be signaledby instrumcntswhose          tical reasoning.
 output rJ noticeable. eilher ca$e,
                        In            however,        It shouldalso be admittedthat the margin-
 the perceiving of the "safcty reserve" in         of-safety formula is not sclf-evident when
 various situations is somcthing which all         applied to certain types of danger in the
 childrcn want and nccd to do. And the             classificationproposed a few pages back.
 appropriateadjustntcntof behaviorto such          The pcrccption of hostile intent in an
 margins is a skill worth practicing, Parents      animal or in another person is not like the
 are comrnonly so reluctant to have their          perception of a cliff edge. The signs of
children "exposed" to dangers or, as we            imminent attack are not always evident.
 say, to "risk accidcnts," that they often         T he perception of thc line between food
 discourage the very cxploratory actions           value and poison value is not a sharp one,
 which would lead the child to form habits of      and the clues are subtle, Nevertheless, I
 pruclence and to avoid unsafe practiccs.          believe that the formr.rlirof maintaining the
 Such an overprotected     child does not grow     safety gap is the only one that promises a
into an adult who is safefrom the dangersof        rational psychologyof caution, or a theory
the environrnent;on the contrary, he is at         ol'avoidance bchavior. A child growing up
their mercy, I suspectthat -'acciderrtprone"       in society must learn the social margirrs of
individuals are of this sort rathcr than           interaction-that is, "how far one can go"
personalities   with an unconsciorrs  dcsire to    with another person before aggressionwill
injure themselves.                                 appear. I-ikewise he must learn the pcr-
    An individual who habitually maximizes         missibletolerances   for edible substances  in
the rnargins of salety irr dcaling with his       his world-and latcr for drink.
environrrrentis said to be timid, He is handi-        Some of the dangers classified earlier
capped by his l'earfulness,Safety demands         produce all-or,none injury, where "a miss
the mttintaining of a. safety rescrve, not the     is as good as a milc." In othcrs, such as
maximizing of it. One often needsto come          sunburn or other radiation and cherrrica
closeto danger,in spaceor time, by reason         damage to the body, the amount of injury is
of some other rrrotive. long as the margin
                          So                      proportional to tlrc amount ofexposure and
is perceivcd and the behavior is controlled,      the concentration of cncrgy, ln thc latter
one is safc.                                      case$ the rnargin of safety is not easily
    Evcrything that has been said so far about    determined. Neverthelcss, threshold has to
margirrs of safety presupposesthat dangers        be established, and a tolerance estimated.
are perceptible, that is, are potentially sig-    even if thesc be somewhat arbitrary. The
nalcd by stimulus information. Many               study of"radiation hazard" provesas much,
dangersare. But sontc afe not. or at leilst       as does also the research the supposedly
are not givendirectly.Neither the imminence       cumulative effectsof "additivcs" in food,
nor the closcness a shark bite or a bolt of
                    of                                The relation of pain to injury and the role
lightning is easy to detect,In the rrranmade      of pain in thc learning of avoidancer.cactions
environment, explosions are difficult to          involve a number of unsettled qr,restions    in
prcdict or locate in advance. We call them        physiology and psychology. They will
accidents.  Insof,ar they arc unpredictable
                     as                           deferred to a later section. We havc so far
events (in the present state of knowledge         bccn concerned with objective dangers and
about sharks, weathcr, and chemistry),the         objective margins of safety, not with pa
terrn is justified. But the occurrcnce these
                                       of         or fcar.
THE CONTRIBUTION OF EXPERIMENTAL PSYCHOLOGY                                       301

THn CnuserloN oF So-cAltrn Acctnr!'lrs  organs) is to train the individual in his
    ln acsotdancewith thc foregoing assump-
                                        ability to discriminate,and to cducate his
tions, accidents    occur for one of two general
                                        attcntion. Children have to learn to see
rcasons. The first is misperception of adanger in the natural environment and to
dangcr in the environment,or of its proxim-
                                        intffpret the conventional signs of danger in
ity, or of its imminence. The second is the civilized cnvironrnent.They have to be
inappropriate reaction to a perceived dan"
                                        attuned, as it were, to the inputs of their
ger. In short, an individual may sufferharm
                                        sense organs, This should bc a very large
from a failure to perceivc or a failure to act,
                                        cornponent of safety education.
a failure of sensoryinput or of motor skill.
                                           Failure in Reactingj The individual may
We may take them up separately,         react inappropriately or incffectively in the
    fgilure in PerrcU4E; As we know, per-
                                        face ofl perceived danger. Thc reasons are
ception nray fail for a variety of reasons,
                                        still impcrfectly understood by psycho-
both external and internal. With respectto
                                        logists.In the caseof a child, motor develop,
extcrnal causes, in the first place, thement may be immature, There may occur
physical stimulus information may be    conflicting tendencies to react-that is,
absent (as in complete darkness) or im- habit interfcrcnce. There may sometimcs be
poverished (as undcr poor illumination),inhibition of reaction by fear. But the main
One cannot seeobstircles night. Sccondly,
                             at         reason is what we call insufficient motor
the availablephysical stimulus information
                                        skill. The essence skill seems lie not so
                                                           of             to
may be unreliable or undcpendable, it       as
                                        much in the connecting of single reactions
often is for the predictionof lightning or an
                                       to singlestimuli as in the carrrrol the flow
€xplosion. The remedy for both is to provide
                                       of action, and the co-orderingof output to
        infornration by artificial means, by
                                       input. Skill is inseparablefrom pcrception
illuminating the environrnent, say, or hy
                                       with respect the patterningand thc timing
   oviding substitute information in the
                                       of muscularaction. It is related to the gaps
 brm of signs or signals,or by instruments.
                                       and discrepancies   referredto earlier,lt is not
   ith instrumcnts, almost any object or
                                       a matter of such popular misconceptions       as
     nt is detectable.The remedy for "hidden"   reflexes" or "quick reaction time,"
             is to make them evidcnt. The best
                                       The rcmedy for ineffective reaction to
      ruments are those that give quantitative
                                       dangers,then, is to train the skills of the
      rmation about margins of safety-thosc
                                       child, or allow him to learn. He must co-
     which the impending character of the
                                       ordinate his behavior to the margins of
          is made evident.             safety.
    With respectto internal causesr       The Concept ol'"Attident":
                                      perception                           If harmful
     y fail becausethe physical stimulus in-
                                       €ncounters between an individual and his
     rnation iri not rcgistered the individual,
                               by      environmentdo occur in thc way described
  Ithough it is physicallypresentar his sense
                                       above, to that extent they are not unavoid-
                                       able accidents.
        ns. This ilay occur, first, by reason of         Remedies   are available,and
                                       preventionis possible.
            in the senseorgans or, second, by                    The term "accident,"
                                       it seemsto mc, refersto a makeshift concept
            of immaturity of the sensoryequip-
                                       with a hodgepodgeof legal, medical, and
     nt or, third, by reason of temporary
       pacity from drugs or illnessor, fourth,
                                       statistical overtones.Two of its meanings
                                       are incompatible.Defined as a harmful cn'
      reason of untrained discrimination or,
       , by reason of so-called        counter with the environment, a danger not
                                       averted, an accident is a psychological
   hich can vary from sleepto distraction-
                                       phenomenon, subject to precliction and
         is, misdirectcd attention. The failure
     y occur at any stage of the perceptual
                                       control, But defined as an unpredictable
          The remedy (for normal sense event, it is by delinition uncontrollable. The
joz                                                       PSYCHOLOG TCAL APPROACHES

two meanings are hopelessly entangled in            tical studies which seem to show that some
                                                    individualsare especitrllyprone to accidents,
thc comtnon usage.There is no hope of
defining it for research purposes. Hence I          they assume that these individuals wish
$uggest that the word be discarded in               themselves hartn, or seek injury' Qr cvcn
scientific discussion.                              /ike pain, I cannot believe that they literally
    The problem of accident prevcntion              do. Such individr.ralsmay scek punishment,
should be renamed-perhaps by calling it             but that is a rnuch tnore complex matter'
the problcm of safety. lt is thcn' on the one
                                                     Expnntupnrnt RnsEnncn         Rnlrvaur
hand, a mattcr ofthe ecologyofdangcrs and
                                                     TO P|jRCIIVING AND AVOIDINC D.q,T*]CERS
the natural or artificial signs of danger and,
on the other hand. a matter of the psycho'              Apart from anccdotes, experimcntal evi-
                                                                                      is scant' What
IoBy of the perceptions and reirctions dcnce about fear in chilclren
 aroused by these signs.Whcn thus         reformu- thcrc is stems rnostly from the formula that
                                                     "the burnt child drcads the firc"' (It is
 latcd, thc problenr appears in a quite                                                            "the
 dill'erent light.                                   interestingto note that thc implication
     'l                                                                     rrot drcad the   fite" has
        he presenccof a danger does not have to unhurnt child <loes
                                       "risk'" An    never bccn studied.) Experimenters have
 be conceived as exposure to
                                                                          with the irnportance of
 approach to the edge of dangcr is not been concerned
                                   "chance'" The lerrrning,      Instinctshave long had a bad name
 nccessarily    the taking of a
                                                                                   Forty years agrt'
  purely statistical concept of danger leads to in human psychology'
  mystification in a// realms, but          mystcry J. B. Watson believcd hc had shown that
                                                                           loss of support were the
  applies only to thc realm of the unknown' loud sounds and
  The great majority of simple dirtrgcrsare not only        stimuli thnt arouscd fear in the young
                                                                             however, it has hecomc
  hidden. Exploration, whethet geographical infant. Meanwhile,
                                                                                                 of in-
  or mechanical    or social,docs ncrthrrveto be clcar tltat the emotional rcsponses
  thought of as the accepting of a calculated         fants continue to differentiate for a long
                                                                        and that this differentiation
  risk weighed against the calcultrted ad- period of time,
                                                                               tnaturrttion and learn-
  vantage. Game theory is not thc only or the is a product of both
  bestway to analyzemotivated        lrunran action' ing, inseparably mixed. Presurnably the
                                                                               fear also differentiate
  It appliesto gamcs of chance. Life is not a stimuli which clicit
  gamble,just as all gamesare not girtrrcs        of Whether      children have or auluirt a fear of
                                                                                   bc decided sinrply-
  "hutl**. Consider the difference between the snakes,therefore, canuot
                                                         The conditioning        and subsequent cx'
  sport of mountain climbing and the sport of
                                                                              response to a fleutral
   iussian roulette. Eithet one can kill you' tinction of a fear
   But in the     former the margin of safety stimulus has often been studicd,
   reserved is always under control if you are with          Watson and Rayner's expcritnent with
                                                                         white rat, But no onc until
   an expert, whereas in the latter control is Albert and the
                                                                                                  of fear
   impossible.   No otle catr become an expert at recentlyhas studiedthe development
   it, and no dne calr         It is a stupid  girme' in responseto a genuirreccological danger.
                                                                                    inverrtcd a device
        The Role of Motivdtion.' lt has been Gihson and Walk have
   ass@irnPlc                                fcar of   which   isolates all the visual stimulatiorr for
   harm is a motive for alt individuals, from          seeingthe cdgc of a cliff without the physica
                                                                             They have fiied it with
   infant to adult, just as the exploratorydrive existence of one.
   is a motive. (To what extent and in          what human infants at thc crawling stage ol
                                                                               babics consistently re'
   way fear is learned is a separate question') development, The
    Any psychological theory of trvoidilnce be- fused to crawl          out ovcr a glass surfacewhcr
    havior and, irbtlve that, of cautit)us of          the visible surface was far below, althougl
                                                                                              the visiblt
    prudcnt belravior   must make such an a$' they would readily do so when
    sulnption. But somc students of accidents surface            was close bclow the glass' Evident\
                                                                              support as wcll as sub
    appear to deny it. lmpressed by the statis- they need optical
THE CONTRIBUTION OF EXPERIMENTAL PSYT]HOLOGY                                              303

stantial gravitational support. The same          is, of coutte, a very general capacity among
resultwas obtaincdwith turtles,chicks,rats,       animals. It is sometimesachieved without
kittens,goats,lambs,pigs, and puppies.For        vision by the picking up of auditory echoes,
some of these animals, it has been shown         as we now know frorn studicsofthe behavior
that the avoidance  doesnot dependon prior       of blind men and hirts,
visual cxperience.Newly hatched chicks,              Apart from theseirrstances, degreeto
one-day-oldkids, and rats rearedin the dark      which childrencrfldiflerent  agesdo or do not
until capahlc of locomotion could distin-        identify the 9cr61111..  dangcrs of their cn-
guish at oncc betweena short depth and a         vironment is alrlost unknown. Parentstend
great depth. Neither of these species has        to assumc   that they do not. The developmcnt
eyeswhich could registerbinocular parallax       of irrational or misplirced feirrs in some
as we do, Thc invcstigatorshirve demon-          childrcn suBgeststhat they do not. The
strated neatly tlrat the effective stimulus      imaginary dangerswhich dominate thc lives
inf'olrnation is mostly carried by motion        of sorneadults suggest   that the identification
narallirx.                                       is never perlect. But thc qucstion ought to
   A clrop-off of the ground is given by        be onc for research,not speculation.
certain pcrspective  information in thc licld       The cvidcncc about skill is in much the
of vicw, Tlre experimentefshave herc iso-       same$tatcas the evidence     about perception.
latcd or abstrircted from the world a very      Psychologists have accumulatecl a grcat
cornmorl type of danger and havc shown          number of f'actsabout the developmentof
that childrcn irrrd animals react appropri-     motor skills but not about the peculiartype
ately to the visualstimulus.Additional basic    ofl skill that can be developedin rhe facc of
research of this type is needed.                danger.The margin-of,safely      hypothesis has
   Another cnvironmentaldangerconsists     of   never bcen worked out in detail. What is
somethingmoving toward one's hody. This         needed, think, is the inventir-rn irpparatus
                                                           I                       of
is signalcdby a rapid expansion-i.c., nrag-     for simulating a particular danger while
nificirtion-of a textureclfigure in thc field   allowing the subject to acr. His action, in
ofl view. This particulrrrrnotion stimulus is   turn, should cor'.trolthe danger. Such de-
one I call "loorning." l havc shown that the    vices are beginningto appcar. A true simu-
 uddcn enlargement a harrnless
                     of            shadowin     lation ol' automotrile driving, for e1s6p1*,
 he ficld of an adult observercauseshim to      is possibleto construct. lt should bc uscd
   nk and dodge. It should be tried with        primarily for rcsearch to discover how
 nfants and animals.Sensitivitvto obstacles     people do in fact drive automobiles.

   This articledelineates wide rangeof accidentproblemsthat can be studied
rorn this standpoint.Insteadof classifying           on
                                           accidents such nonexplanatory      bases
s placeof occurrence typc of irrjury,cibson proposcs ccologyof clangers
                     or                                an                     which
   lyzes and classifies thc various cnvironmentalforccs that constitutepotential
hreatsto safety.The waysin which man and other animalsperceive     and avoid these
  ngcrsconstitutea fertile field for psychological
   cibson is on solid ground in suggestingthat "thc tcrrn 'accident'be discarded in
 ientific discussions." view of tlre necessary
                       In                       lole of thc nhvsicalarrd chemical
gcntsinvolvcd,it is more reasonahle classify resultant
                                      to       thc          injuricscorresponding-
I for example, electrical,
                as                                          in
                              mechanic-al other iniur-ies, line with our earlier
iscussion the role of such agents.tThis has the additional advanrages
           of                                                               hoth of
 singclassificatiotr the natureof the responsihle
                   on                              agents-as is donc irr thc casr-'of

                                 irre now beginningto be taught "accident" prcventionfrom
  It is of interestthat physicians
     ially thc sarlc ecologic stirndpoint.4
304                                                 PSYCHOLOGICAL APPROACHES

the viral, rickettsial, bacterial,and other biological insults to the 6ody-and also of
                                                                         "accident." Thc
avoiding the folklore and extra-rationalism     associated with thc tcrm
concept                 however, is too strongly entrenchcdto disappear quickly, al-
though wc believeit will eventually,at leastin the scicntificliterature.
    Although not primarily a psychologicalpoint, Gibson's mention of the role of
environmental dangersin natural selectionis of particular intercst. For examplc,
geneticfactors associated                                                       to
                             with the initiation ofaccidents,with susceptibility injury,
and  with the ability to recoverfrom injury are probably being diffcrentially sclccted
evenin prescnt-day    populations,   just as they haveundoubtedlybeen in the past. ln
fact, since accidentsare currently the leading cause of death in many societie
throughout most of childhood and the childbearing years, it is quite possible that
they are one of the principal contemporary means by which the composition of
population genepools is changing,a point almost univcrsallyoverlookcclin rescarch
and evolutionary   speculation geneticists others.Nonetheless,
                                 by              and                      many evidence
of geneticfactors of possiblerelevance      may be cited. Theseinclude the apparent role
of geneticfactors in myopia;r the demonstrationof strain differences experimenta
animalsin preferencc alcohol;'t
                         for           genctically mediated variationsin susceptibilityto
injury, as in hcmophilia, albinism, porphyria, and osteogenesis       imperfecta; and the
probability among more normal individuals of similar genetically      mediatedvariations
in ability to survive injuries of various types once they are sustained.

OrnnR PsycrroroclcAl, AppRoACHES
   Intermediatebetweenthe extremes method and emphasisrepresented lrcud
                                       of                                   by
and Gibson lie a variety of psychologicallyoricnted contributions, ranging from
speculationunsupportedby data to rcports of tests of various psychologicalhypo-
theses. Like much of the rest of the accidentliterature, thesevary grcatly in quality'
Since we can merely sirmple this extensiveliterature, wc havc chosen papers that
demonstrateits variabiliry rather than limiting ourselvesexclusivelyto examplcs
sufficientquality to serveas modelsfor subseqnent   research.f

  -Willard Kerr. Ph.D.

                                                of                    theor
                         briefly the inadequacies the accident-proneness
   This article summarizes
and then proceeds to develop two additional theories: of psychological climate and
adjustmcnt to stress.Basically,Kcrr hypothcsizes that a work climate cha
by greatfrecdom to scr rcasonably                                             Suc
                                  attainablegoalswill resultin fewer accidents.
freedom is said to increasethe worker's senseof responsibility and therefore hi
alertnessin preventing accidents.Although he does not presentthe basic data, t
author cites lower accidentrates in organizationswith the prescribedpsychologica

+ Additional examples are found in Chap. 2 (McFarland),in Chap. 3 (Loomis and West
Cibson),in Chap.5 (Orleans          and in Chap.7, all of whichis concerncd
                           and Ross),                                     with the literatu
on accidentDroncncss.
 COMPLEMENTARY THEORIES OF SAFETY PSYCHOLO(iY                                                                     305

 climate. The "adjustment stress" theory propose$ that distracting stress upon the
 worker will increase accidcnts.
     Thesethcoriesare not oll'eredto disprovethe accident-proneness  thcory but to
 explainthe largestproportion of industriataccidents. Thc author estimatcs that the
 individual goals-opportunity-alertncss
                                      hypothesis will explain 30 to 40 percentand
 that the adjustment-stresshypothcsiswill cxplirin an additional 45 to 60 percent,
 leavingonly I to l5 percentfor the accident-proneness hypothesis. Theseestimates
 are not substantiallysupportcd by thc evidencepresented.

 Pnonlsly rHE Mosr universally ignored area        "accident
                                                                 prone" in a general sensebecausc
  ofsafety psychology that pertainingto the
                        is                         there are potcntial tasks that no human
 psychological climate of the workplace. A         being can perform without accidcnt (e.g.,
 devotion to safety gadgets on the one hand        climb the outside walls of the Empire State
 and conccrn for the alleged pronenessfac-         Building to the top with one's birrc hanrls).
 tors within the accident repeirte on the
                                      r               ProfessorsMintz and Blum and Maritz
 other hand has led to the irlmost total           have shown that the accident proneness
 neglect ofl the situational factors which help    theory has becncxplainingcntirely too much
 shape work personality and help manu-             of the industrial accidcntrate. The re$earch
 facture accident-free or accident-liable em-      of Cobb, Johnson, Whitlock and Crannell,
 ployees.                                          Forbes, Farrncr and Chambcrs,ancl Harris
    Many investigatorshave shown that bc-         point toward the sarle conclusion. Mintz
coming a safe worker is a typical learning        and Blum showed that the frequency of
function, Thc decline in accidentsfrom date       '*repeater"
                                                                   accidents approxirnatesa pure
of cnrployment in the typical job is a repre-     c h a n c e( P o i s s o n ) i s t r i b u t i o n .M a r i t z t h e n
sentative.learning curve. But like other          suggested      that thc final crucial test of vari-
learning curves, the decline in error per-        ancein the industrialaccidentrate accounted
formance can be obstructedby a multitude          for by proncrrcss thc correlation bctween
of other factors. lt now appears that a chief     onc's accidentscxperiencedover two diffcr-
obstruction to the rapid decline in error         ent periods of accidcnt exposure-such as
performancc is defective psychological cli-       the last two years and the next two years,
mate. This conclusion, to be supported in         Chiselli and Brown have collated lB such
this paper, standsin sharp contrast to past      coefficients frorn the literature, and the
emphasis upon the accident proneness             presentauthor hirs computed their median;
theory.                                           it is .38. This typical valuesuggests                  that only
                                                 about l5 per cent ofthe variancein individ-
THn AccturFJr PRoNENEss       THEony             ual accidents accourrted by variancein
                                                                      is                   for
   Beforepresenting crucialevidence
                      the                    on individual accidentproneness;                       furthermore.
this theory, the term "accidentproneness" this even may be spuriously high bccause
should be defined,Attitlent proneness a such coefficicntsare contaminated by the
     titutional (i,e., permanent) tendencv correlation of the workcr's position lrazard
   hin the organism to cngage in unsafe with the consistcncyofl his accidcntsover
behavior within some stated field of voca- sptit timc periods.I n fact, it is almost certain
      I activity. A temporary tendency to        that much, if not most, of this l5 per cent of
      accidents not proneness; is liability. potential variancedue to accidentproneness
                is               it
       pronenesliis not general; that is, its actually is due to environmental factor$
  ferrent to an activity field must be linrited (tenrperaturc diff'erences,              fumcs, congestion-
o be meaningful, for, obviously, everyone     is space-threat dill'erences, etc.) left uncon-

      f_ Reprintetl,with permission,
                                   from TheJournalof SocialPsychologv,
                                                                               PSYCHOLOGICAL                      APPROACHES
                                                      system must be geared to support high
trolled and hence correlated with each other
                                                      quality work bchavior'
in the lI coeihcientscited'
                     the unreasonablc assunrp-           ln businesspractice some tralnlng rnter-
     Even allowirrg                                                        "tclling wh$t to do and
                                                      feres by too much
tion that these l8 cocfficicntSv,erc not influ-
                                                   in what not to do" and too little encouragc'
cncud by fatiguc and sttess diffetences                                                             owrr
                                                   of ment to the new worksr to do his
 differenfjob locations, the 15 per cent
                           "accounted for" by thinking and '-standon his own feet"' Union
 variance ln accidents
                                                      leadership likewisc often is guilty of too
 pronenessstill leaves 85 per ccnt of the
                                                       much propaganclizing and not cnough
 variance in accidcnt rates uflaccounted for'          "asking;' in rtletions with new workers'
                                           study   of
     It is intercsting that an eatlier                                                 trew worker is
                                        arrived in- Such iiritial clinrate l'or the
 auioltrobile drivers by Forhes                                       to alertness  thirn to a relative-
                                                 that Iesscouducivc
 .1*p*ncle.tty at the similar conclusion                                rcsigned' passive   conformity
 the accident    repeater contributes flot more ly unmotivirtcd.                                    total
                                      the accident to thc apparcntly        alreacly-structured
 iha,r thrce or four per ccnt to
                                                 who situation.
 ,rr.rnf**' Both Jolrnsonand Thorndikc
 'iater                                                   Acci<lents.of course' show that the total
         surucye<lthe entirc reseatch litcratnre                                              rlnd from
                                        firund that situatiotr is nof lirmly structured
  on automibile safety tikewisc                                                                more self
                                  hzrsic   ilptitudes  thcm thc workcr gmdually accepts
  ,*u*fl"onttitutional flactorsas                                         order to survive' But an
                                      with accidcnt resuonsibility in
  vicldcd negligiblcrelationships
                       also"is the fact that Ilunt'     accident is an expensive teaching dcvicc'
  ie.ords. Rilcuant,                                                                      ir climatc in
  Wittson, and Burtotr cornputed         the psyctrt- Futthermote, if it occurs in
                                           inductton which the employee      is expectcdto supply his
   atric discharge rats Bt Naval
                                                               but not his opinions or ideas, the
   stations  and subsequentlyduring World War energy                                         foreign in-
                                     nine per cent a.ciJent is misundcrstood as a
   ll to vary betweerr four and                                                  belorg in the schemc
   (such dischargee$   were' of course' those in- trucler which does not
                            "prone" to behavior of events. In such circumstances, rarely        it
   iiuiOunft             as
                                    tlrcir country)' occurs to lran'.1geil\ent,     union' or worker
   unsafc to themselvesandfor                                                                     and in-
                                       theories may that an rrcciclentis made ncce$sary
       Two situationtl or climirtic                                                                    his
                            non-chance      variance' evitable in order to teach the emlrloyee
    a*fluin the rerrraining                                                   irnd essential personal
                                                        own indivi<luality
                                                                    union             Even the teaching efficiency of the acci-
       Plainly, botll managcment and
                                                                                   dent itself is i*terfcrcd with, however, if
  training activities, pori.i-*l ^r.r letrderships
                                                                                                        t s o f h e t the p s y c h o l o i c a
  i r r e r e s p o n s i b l c f t r r s o m e i n t c r t t r c n c e w i t h t n o s t a s p c c denytthat o t a lindividual'sgown l
                                                    perftrrmzrnce.                 rnate i. eI}'ect
  the norr'al declirre of error
                                                                       that mental             cttntent is important'
       In stating this theory now' we holcl
  g r e a t f r t e t l o m t r * t ' * u ' " ' n a h l y a t t a i n u h l a l f t h e c l i m a t e e n c o u r a g e s t h e i n d goals with
                                                                    qu.ality set up long-term and short-term
  goals is att:ompanied tl'pi':atty nieh              ni
                                                                                                       b a b i l i t y seems i n m e n t
   u , o r k P e r f o r m a n c e . T h i s t h c o r y r e g a r d s a n r - e a s o n a b l e p r o situationo f a t t alessfixed,
                                                                 be- Gestultol'the work
   accitlcnt urcrely as o to*-qu*fity iork
                     "scrappnge" that happcns to a and the worker
                                                   --ft."iting                                    fcels himself to be a signi'
                                                                         ficant pnrticilrant' Significant participation
   person instead of to ^ ir,i"g.                               -the                                                          prohlcm-
                                                                   of makes fot habits of alertncss,
   level of quality involves .uiri"ng tlr* lciel                                                                         'l'he
                                                                   be raising,        a.<I problem-solving'
   alcrtness; such high ur*rtn**i crrnnot.                                                                                          the
                                                           psycho- logical work environment must reward
   sustainedexccpt within u-'**attfing                                                                     tbr bcing alcrt, for
                                                                  the worker erilotionally
   logical climate.Thc nrorl iich, thercforc,
                                                                                                                       a p t o not to ge
   e c o n o m i c ) r e w i r r d o p p o f t u n i t i e s ' t h e h i g h e r g e s t i o n s ' f o r p a s s i n g ort ilrow a c o - w o t k
                                                                    the horv bcst to d. sornetlting
   the levcl of alertness-aiJ-the'higher-                                                                                      of the ordi'
   level of work quality. oh;;riy,                   tire"iewaro. hurt, and for achievement out
 COMPLEMENTARY THEORIIS OF SAFETY PSYCHOLOGY                                                   307
   nary. The worker must feel free to exercise     relatively "dead end" jobs should operate to
   influence over his environtncnt.                raisc thc averagelevel of alertness,not just
      Considerable  evidence supportsthis theo-    to hazards but to everything.
   ry, Factoty depirrtrnents rvith rnore move-
   ment of personnel   among departments,   that THs Aurusn'{ENT-SrREssTHnoRy
  is, intra-cornpany transler mobility, have         Thc individual goals-opportunity-alertness
  fewer accidents;   the sarncjs true of depart-  theory of safety seeil$,to cover much of the
  mcnts with greater pronrotion probability       variance trot covered by the pronencss
  for thc typical employee(r is -.40). Dcpart-    theory, but sornevariancestilI rernainsand
  ments with tlrc bcst suggestion   rccords (re-  it irppears neccri$aryto verbalize a third
  warde<l)  lend to havc fewer accidents. Addi-  theory. Probably almost all ttf the remainins
  tional evidence of the influence of the        variance  can bc cxplainedby a rhird theoryJ
  stimulating individual clinratc on safety is   an adjustment stress \1rrurr. Ir holds that un-
  founr.l in the tentlcncy toward fewer acci-     usual, negntive, dislrtttfing stress upon the
  dents in individual-typc than in crew-type     orguni.tttt fur'reaJes its liahility to artirlent or
 jobs at the Intcrnational Harvesler Works.      to othcr lox, quality behttvior. This too is a
  In individual-type  work, the employeerarely   climatic theory, bccause environment ig
 is uncertainilbout his rcsponsibility con-
                                        for      intcrnal as well trs external,and this theorv
  sequcnccs; better knows his immediate
               he                                refers to distractive negative strcssesim-
  work goals, Another interestinghit of evi-     posed sp11nthe individual organism either
 dence is that in two dilTercnt studics the      by internaI environment (suclr as disease
 factory departments with incentive pay          organisms,alcohol, or toxic iterrrs)or bv
 systerns,  although problcrn departmentsin     cxternal envirthnrent (such as tempcrature
 regards to monLrtony.Iowcr job prestige,       excc$ses,   poor illumination, excessive        noise
 and lowcr promotion probability, still have    lcvel, cxccssivcphysical work strain), Its
 no mofe accidentsthan departmcntswithout       strcsscs  are different fronr thosc expericnced
 inccntive pay systcnrs, This seernsin such     by thc accidcnt profle; thcir stressesresult
 defiance of expectations as to suggest that    from a utn.rtitutiona! inadequacy. Ordinary
incentive pay systcms restrict accidenls by     adjustment stressis nol the result of con-
encouraginggrciiterindividual initiative and    stitutional inadequacy bul of temporary
alertness.                                      conclitions.
    Accidcnts are more frequen[ in jobs of          What oftcn appears at first to be consti-
lower-rated prcstige; one interpretation of    tutional irccidentpronenessmay be shown
this finding is that clirnaticallythe job must very clearly upon more carcful examination
seemworthy enough trr the worker to sustain    to be the operation of tclrlorar.y stress
his euphoria level. This intcrprctation is     factors, The nrost sobcring example of this
supported by the finding of Hersey that out    is found in the curve of accident ratcs of
of 400 accidcntswhich were studied clinical.   $ucccsslve groups of industrial workers.
Iy, morc than half took place when the         This curve shows high rates in the first I0
worker was in a worried, a;prehensive,or       years of thc worklile and a secondarvin-
some othcr low emotional state,               creasein ratcs betweenthe agcsof.10 and 55.
    This individual goals-freedom-alefrness   These age periods al$o are the grcat stress
  heory suggcsts the c)inratic need for pro-  periods in the typical worklife; this is sue-
   ng emotionalrewarclopportunities      for gested by the fact thar the acciclent rale
           such as special economic incen- curve and the turnover fate curve super-
     prestigc-building   honors, cxtra privi- impose almost perfectly upon each other
     machine and work space decoration when plotted through successive                 age groups
       pilrticipation, and reprcsentationon of thc industrial population. The alleged
pecial committcesand councils, These re- proneness within the young uccident re-
 ardsheldasattainable    goalsby workersin pcatcr is largely dissipatedwhen onc con-
308                                                    PSYCHOLOGICAL               APPROACHES

siders that most of the stres$ is environ-       family and marital affairs,etc., still will carry
mcntal-and associatedwith adjustment to          over into the workplac:epsychologicalcli-
work discipline, attaining self-sufficiency      mate and causc accidents.
away from pirrental ties,courtship, marriage,       In the adjustrnent strcsstheory it must be
assumption of family economic responsi-          admitted that individual dilJ'erences cxist
bilities, and thc struggle for a foothold on a   in ability to withstandwhat ordinarily would
vocational ladder that seemsto lead some-        be strcss-inducing  situations, Yet, such indi-
where worthwhile, Another set of obvious         vidual differencesaccount for lessthan one-
stress explanations comes to mind to ac-         tifth of the variancc in individual accident
count for the -'rniddlc-ageboom" in accident     ratesr thereforc. thc limitations on the acci-
rate.                                            dent proneness theory appear to be much
   Temporary stress factors which already        morc severethan thosc on the adjustmcnt
have been found signilicantly correlated         stresstheory. The lrrct is that both cnrployer
with accident rates include employee age,        judgment and joh applicantjudgment opcr-
workplirce tefiperaturc, illumination, mean      ate to prevent the opcration of any grcat
rated comfort of the shop (r is --.70), degree   amount of accident proneness.While all of
of operational congestion,obvious danger         us are accident prone for one task or
factor thrcatcningly prcscnt, rnanual effort     another. wc don't orclinarily apply for or
involved in job (r is .47), weight of parts      allow ourselvcs to be engaged in such
handled,frequency    ofparts handled,alcohol     tasks-and we probahly couldn't get hired
consumption, arrd influence of discrrscor-       for such tasks if we triec.
ganisms.                                            On thc basis of the evidencesummarized
                                                 and the author's own estimates,the variance
                                                 in accidentratesamong industrialpersonnel
          tuny Ltultatlons
CottpLnlvtet                                     probably distributes in terms of thettretical
awn InrEnenrrATIoNS                              causation according to the following pat-
  It seemswise to emphasize that both of         tcrn;
these new theories of safety complernent           Accident proneness                    ly,to t5.]/"
each other as well as existing proneness           l n d i v i du a l
                                                       goals-opport   unity-alertness   3ol( to 409d
thcory. In thc goals-freedom-alertncss theo-
                                                   Adjustnrcnl $lrcss                   45ii to 60'l
ry it must be recognized that (rr)even under
                                                        Total Variance                     I007;
an optimal opportunity climate, individuals
who lack the characteristics necessaryfor           Constructive thinking about the individual
the work probably will continue to have          goals-opporttrnity climate and about ad-
accidents;(b) excessive physicat sttessescan     justment stressesshould assist industry to
causcaccidcntsin any psychologicalclimate;       escapc the defeatism of the overly-cmpha-
and (<:) psychological stresscs relative to      sized proneness theory rrnd better under-
adjustment to changing life aspirations,         stand and control accidents.

   Although the two theories proposed in this article are suggestiveand are said to be
supported by rescarchfindings, evidencefrom researchspecificallydcsignedto test
thesetheoriesis requiredbeforethey can bc considered    seriously.The sources cited in
most cases not contributemore than circumstantial
            do                                            evidence. For example,   the
validity of equatingintracompany   transfermobility with freedomto set goals,and
employee   agc with adjustmentto strcss, highly questionable,
                                         is                        The assignment   of
estimates the relativeimportanceof accidentproneness,
          for                                                 psychological  climate,
and stressin accountingfor 100pcrcent of industrial accidents complctelyunwat-
ranted on thc basis of the cviclcncepresenl.ed.
PSYCHOLOGICAL CLIMATE AND ACCIDENTS                                                        309

  - Vernon                                                M.'4.
               8.5., lltillard Kerr,Ph.D., William Sherman,

   The study that follows servedas the basisfor much of the discussion the pre-
cccling selection.A panel of cight expericncedjudges rated44 shop departments a   in
tractor factory on variablcspostulatcdas determiningthe psychologicul        working
climate. Theseratings wcre then correlatedwith the dcpartmentzrl   lost-time accident
rates for five years. A strong point of the designwas the analysisof the reliabilityof
the judges.

IN rHE sELtEn  that certain factors of psycho-   the worker and increases         which
logical clinrate and physical environment        are unrclatedto the obviousdangcr.
may be of great importanccin the causation
of accidentsin hcavy industry, the authors Suusers
formulated a seriesofhypothesesconcerning
                                             Lost-time    accident        of
                                                                  records 7,103   (aver-
cortelates of acc,identsand sub.lccted each to
                                           age)pcrsonnel a tractorfactoryfor the
an experimcntal design for tcsting. The
                                           1944-48  five-year   periodweretabulatedfor
hypotheses were:
                                           cachof the44 shopdepartments theplant,
  l Excessivevariation in pressureto "get  and accident     rate per 100 personncl  was
out production" inducesstressand tension computedfor each deprrtment.The enter-
contributing to accidents,                 priseexperienccd     1,943 lost-tirneaccidents
  2, Intcnsity of normal production sched- during this periocl,  ancidepartnrental rates
ules contributes to flccidents.                                     pcr
                                           per hundredpersonnel yearrangcdfrom
  3. IIigh promotion probability encour- 1 . 5t o 1 3 . 2 .
ages alertness which rcduces accidcnts,
   4. A conrlortable shop environnlent re-       PR()cEDURE
dugss ascidents by minimizing physical             Five-point continuum ratings on each of
annoyanccsand distractions.                      the variablesshown in Table I werc collected
   5. Ernphasison individual responsibility      independently  from each rnemberof a
(rather than upon crew work) act$ to reduce      informed" panel of judges. This panel of
accidents.                                       eight judges, all indivitluals with at lcast ten
   6. A reasonahle level of job prestige is      years of divcrsiliedexpericnccin the plant,
requircd for safe behavior.                      includes the staff assistant to the works
   7- Heavy rnanual work induces stress,         mQnagcf,the assistantwprks manager,the
fatigue, and weariness, which increase           general superintcndent,    the plant engineer,
probability of accident.                         the chief of methodsand ratcs,the planning
   ll. Inccntive piry $ystemsencouragealert-     engineer, the safety supcrvisor, and the in-
nesswhich in turn promotcs safe behavior.        dustrial relations manager.Split-half (odd-
  9. Relative adequacy-predictability of         versus-evenjudgcs) rcliahility coefficients
work space is essential for avoidance of         correctedvia the Spcarman-Brown       prophecy
distractiorr-worry and saiety.                   formula for thesc ratings arc indicated in
   10. Presenccof an obvious and threaten-       Table |.
ing danger has a scnri-hypnotic  cffect upon        Mean judge ratings on each department,

     f-Reprinted, with permission, from the JDunal of ApptrcdPlyrhoktgy,l5:2:l0S-l I ll
     L     ) 951 Copyright, ) 95l, by The AmericanPsychologicat
                .                                               Association,
                                                                           Inc,       )
310                                                                       PSYCHOLOGICAL APPROACHES

                                     INTERCoRRELATIoNS'ltlonc Acctnntr      Rrtt rrno EEttl oI Tsr't OrHen VARIABLES lN
Tlsle     l.-PtensoNIAN
                                    44 Drpnnrvrvrs oF e Tn'rrc:ton 't:'q'ctont'(7'103 Prnsorrrurl)*
                                                           2    3     4     5      6       7       I       e       10      1l
             l/a rinble                              1
  t                      to                                                                                             'll
                                                   -.42 .21 .00 .60 .sd,-'30              '24 'ss         '78     'rl
 t'                   to sct                                                                                          -'06
    [':i];:"li:,iJ'ssure                            .21 '"?d .23 -'e4 - ' 2 s -'o8
                                                                          42              's4      67     's0     '6s
                                                                                                                  '43 ' '01
                                                    :00   'ii    -'67
                                                                'g2               '94 -'l I       '00      s6
  3. Promrrtion robabilitv
  4 . c l o m f o r t ^ L l l s l r o pe n v t -
                              e                                                          .44 -.46 -..59 --.1I -,s3
                                               --.60 _.94 _.62            .g2 _.s0
                                                                                                      '46   '13     '37    '33
  5. Degree f     o    crewwork                   :*       'lz -'zs -'so         '91 '00       '58
                                                                                         '99 - 30 -'34      -Os '18 -'?l
  6. Job Frestrse                                 ',;i -'og
                                               -,io                '9!t    tt   '$0
                                                           'tl -'r i - +o       '58 -'30         8J  '68     '36    '64
   7- Manual elTortinvolved
   *                                                                                                                         '4s
                                                  .J5       .i7     .00 -'{?      'd6 -'34      '68    '70    '4s    '47
       3"tu;:-',,it;'neration'l                                                                        '49   '76    '43     '78
   9 . D e e r c cL ) fi n c e n t r v e o r K     .is      '50     '56 -'ll     '13      '05   '36
 I 0. I)egrcc of ohviotts                                                                              '47    '43    '95    '35
                                                   'l |.65          '43 -'53      '37     '18    '64
       danscr firctot                                    -'oo -'01 -'?0          'ff -'zr        '41 '45      '28    '35     '73
 l t . A c c i d e n tr a t e                      .ii
                                                                                                         confrtlence or bett€r'
 I cocfiicicnts in boldface typc are statistically signilicant at the five-F,trcrlnt Llvel of
                                                               the formula o" -               was uscd' Cocfficientsin italics
 In determining the staiisticill significance'                                   tlt--fi

 on cach of the ten variables,thcn were com-                           duction prcssure. When comfort of shop
 nutecl and the dcpartrnental means on each                             environmcnt js held constant by partial
 variahlc wcre correlatedwith departrnental                             correlation, the relationship increases to
                                                                        statistical significance,suggcsting./ht'rr€r iICCt'
 accident ratesi' Intercorrclations atnong all
                                                                        <lents  in departments with above avcrage
 variables also are shown in Tablc 1'
                                                                        variation in protluctiorrpre$surc'While all
  Rnsur,ts                                                              statistics in this study which pertain to
      Results of these correlatiotlnl studies be-                       duction preslure are somcwltat questiorrrrhlc
                                                                        in interprctrtion because of the trpparently
  tween accident rates and ccrtaitr psyc:ho-
                                                                                   opinions of rirtcrs, it is possiblc that
  losical chirrirctcristics of organizational dividcd
  ,r,r]it, or. displayed in Table l' A sumnrary                         variation in procluction pressurc tcnds to
                                                                                  accidentsby reducing monotony and
  inspection of the table suggeststhat $everal reduce
   of thc ten hypothcsesstirted receiveenough encouraging
   confirnttttion to continue thctn                      as tcnable         Sirnilarly, when comfort of shop environ-
   into rnore rigorous future rcsearches'                                mcnt is heiclconstant. the partial correlation
       Variation and Volumeol lr.otluction                      Ptes.s- between collsttltt productiotr pl'essurc
                                                                                          acciderrt rate attains (inverse)
   zre.' Neitlrer of tlresevartablcs Sccms De dcnartmental       to
                                                                            value of rrnity. If this rcsult is not arti-
   -igrrilicLLntlyrelated to irccidetrt rirtc' The a
                                                               (.- '42) factual, it may be that production prcssLrre
   cxlraorclinary llct of a rregative
                                                                                         constant encourages aleftness
   reliability coefficicut stlggcsts the chance evcn when
                                         "schools of thought" which in turn make$ ftlr safe behavior'
   identifircation of two
                                              *'pressurc to get             !1pm9!io11    Probability" Althouglt the.zero
   as to wltrrt cotrstit
                                                                                 "*tciition betweetrdepartmentrrlpro-
    things donc." Since each ol thesevariables otact                                                               is insig-
    corrclatcsubstantially               with at leastone         other motion probability and accidentrate
                                                                                    the rel'.rtionshipbrcomes significant
    variable in Tablc f it is obvious that the nillcant,
                                                                         (-.90) when comfort of shop envlrollmcnt
    split-halt' rcliability coefficicntsin these two                                                                    This
    instancesare gross underestinratcs'                                  is held constantby partial s611glsfie11'
                                                                  -'18    resultis consistcnt   with the findirrgof a signi-
       P a r t i a l c o r r e l a t i o n so f r r . r t ' s *
                                                                                   correliltion of -.40 betv''cen thesc
     and rr, rr.o ",- -'55 hclp tstimirte possible ficaflt
                                                                                  two variablcs in a New Jersey elec-
     cffccts'of incentive work nnd conrfort of safle
     shctp envit'orrtllent            on the relatiorrshipbe-             trotrics Platrt.
     tween accidents and rated variatiorr in pro-                            Cor4fortable Shol Ernirottmettt: Mctst stg-
  PSYCHOLOGICAL CLTMATE AND ACCIDENTS                                                                                                      311

  nificant ofthe findings in this researchis the                             rates. Crew work in difficrrlt opcrations (e.g..
  discovcryof a corrclirtiorrof -- .70 bctween                               moving a hot ingot 1o position in a drop
  accidcnt ratc and mean-ratedcornfr:rt ol the                               forge.andl;ller renroval of the shapedpart)
  shop environment.Partialcorrelations                       hqrld-          calls not only for reasonablysynchronous
  lng constant tiegreeof opcrational conges-                                 action and coopefation of the crew urem-
  tion (-,60), nranualeflirrt involved (^-.55),                              bers,but colrgtflnt        vigilancef'or the continrr-
 constaflt pressure (-.96), and degree of                                    ous acccpting irnd relinquishingof rcspon.
 c r e w w o r k ( - . 6 6 ) s u s t a i nt h e o r i g i n a l c o r -     sibility as well as the devclopmcnt of any
 relation.             psychologicalimptication of                           u11u5ualcircumstances. Lcss stringent co-
 this result is that approximatelyhalf of the                                operativework sometimcshas greateracci-
  variance in departmental accident rirtes is                               dent probability also in largc part perhaps
 accounted for by variance in the c()rrfort                                 bccause of the occasional transition mo-
 factor. Heavy industriesin which forgesand                                 ments in which reslrons          ihility lor coflsequen-
 foundriesare essentinl           opsrirliongincvitably                     ccs is vaguc or suspendetlbut operations
 experience      great dili'iculry irr making sorrre                        nevcrtheless         conlinue, fionr nrrlnrenturrr                 if
 job environnrent physically comfortable,                                   nothing clsc. In thosc brief scconds.ncci-
 Heat radiriting from nrolten nretals.violcnt                               dents are madc, In individual-typcwork the
 sound vihrating lrom clroF firrgcs, and                                    individual rarcly is uncertainabout respon-
 gritty earth particlesfrom foundry "molds"                                 sibility for consequences.
 are vivid cxamplcsof thc practicalobstacles                                     Jah Frestige: It is interestingto note that
 to comfortableadjustmentof (hc worker to                                   whilc this corrclation of --.23 with acci-
 thc iob.                                                                   dents is not statisticnlly significant, it is
     Workers with low "physical frustration                                 highly consistentwirh the value of -.10
 toleranc:c"unclcr c:ontlitions heat. noise. of                             found in thc 53 dcpart,ments the Camden    of
 etc., probably have somewhat higher acci-                                  Works of RCA. C-'onrbining trvo plants      the
 dent frequency than their co-workers. An                                   makesa total of 97 dcpartnrcnts                 with a total
 ircculnuliltion t>f physical frustration ten-                              of 18,000 personnel, and the relationship
 sions leatlseasily to "tlistractive" helravior                            between job prestigc and accident rate is
 which often rcsults sirirultaneously re-                  in              established as highly significant, Partial
 duced quality of work and accident. It                                    correlationevidence,            however.suggerit$              thot
 should trc cmphasizcd,howcvcr-and this                                    the relationshipis at leasrin part irrtifactunl.
qurliiication is suppuylgdby the evidence                         of       When cornfort of shop cnvironrrrcnt hcld                  is
 Mintz nnd Blum-that variance in such                                      constant,the coelicient declincsto .12. and
accident proncness individuals probably
                             of                                            when manuul effort is held constant it cte-
accoulrtsfor a minority of the variance in                                 c l i n e st o - . 1 0 . H i g h a c c i d e n td e p i r r t r n e n t s
accidentratcs.ln othcr wr.rrds, "norrnal"     thc                          are characterized by less physical cornlbrt
wotker irs well as the "pronc" worker is                                   and a demand for grcatcr manual clfort;
pronc to thcse sunreunsafle               behaviors under                  incidentally, they are of lowcr avrrage
idcntical frustrriting conditions, but at a                                prestigevalue, but this fact seemstt> have
slightly bcttcr thrcslrold of tolerance. A                                Iittle causal hearing upon accident rate.
reasonable conclusion under such cviclence                                     Manual Efibrr: The glealer the degrec of
Rppearsto bc that majar prevuttalive e.ffort                              manual effort involvcd in successfulperfor-
(iln tilast lrrotitably be expendedin pprking to                          mancc ofthe typical job, the greater the tcn-
alleviate as much as possihlethe averall phv-                             derrcy toward acciclerrts(r : .47lr. This
sical frustrations v,hith impinge on ull per-                             rclationship pcrsists (rr, ,o., : .96) whcn
sonne/. Srrclr cllirrt cirn reasonably be cx'                             '.pressure get
                                                                                           to       things done" is held con-
pected to reduce the accident ratcs of both                               stirnt by pirrtial rorrelation. As in the dc-
the more-proncand lcss-pronc                   per$onnel.                 mand for constant vigilancc in crew work,
    Degree ef Crev' Work: Departments in                                  the demand for physical enerlry sonrctimes
which the typical job perflormnncc in a                  is               exceeds       efticientlydirectedsupply. Workcrs
  $rew" often have above-rverase accident                                 trained in economicalexpendifureof effort
                                                          PSYCHOLOCICAL            APPROACHES

                                                                          too much and facilitating
as well as in control of physical forces may their attention
ledusc the frequency       ol such accidents'         their involvcmcnt in otller hazards. Wherr
                                  "Degr^eeof ope- comfort of shop environment is hcld con-
     Onerational Congeslion:
                                                              (this is a questionablc ptocedure in
rational crngestion" irnplies frequency- stant                                    "cornl:ort" atrd "dan-
proximity of rnoving        objects and persons' this instance becausc
                                                                       may be regarded as almost
 Such activitics in the psychological peri' ger factor"
 phery of thc irnmcdiate task         may connote synonymous in certain key departments),the
                                                                                to a partial r of -'04
 greater annoyancc-distractionvalue than zero order r declirres
 actual physicalthreat to the worker, hut tnay        bctween obvious dangcr thctor and depart-
                                                                accident rate.
 contribute 11: .45) to unsafe behavior of mental
 personnel     A lirnitation to this interprstation
                                                       SUMM^RY                  :
 ls the fact that the correlation declincsto '06
                                            is held      Against a criterion of five years of lost-
 when comfort of shop errvironrnent
                                                             accident rates in the 44 shop depart-
 constant by partial correlation' In view of time                                                possible
 this limitation, a causalinterprelation      of the ments of a tractor factory, tcn
                                                       explanatory    hypotheses were evaluated by
 zero order r is not warranted'
                                                                of data from a special rating panel'
      Inrcntive f4zork;Departments with incen- means
                                                                       exactillg and precise dirta are
 tiuffiit      muv on thc avcrnge incur a fcw Until rlorc
                                                                    all conclusions ofl this research
 more accidentsthan do departnents on the obtained,
  "straight-time" payment system' This is must be consitleredtentative; however'rwith-
                                                          these limitations, certain trcnds appear'
  ,ugg**t*d by thc original correlation of '28 in
  (noi significant) which is increased to '35            1. Results obtaincd on variation in pro-
                                                                  pressureand con$tant production
  when promotion probability is held con- duction
                                                       pfessufe are inconclusive.
  stant. While evidence not availirble test
                            is                to
                             that when more work         2. The hypothesis that promotion proba-
  the point, it is possible                                                                   appears to
  is being pcrltrrmed    (as it usually is under an bility encouragessafe behavior
                        more accidents   will occur' be supported'bY        this studY.
  Thus, an increasein acr':idents      under incerr-      3. Comfortable shop errvironmcnt ap-
                                                                                         to be a major
  tive conditions might       bc a function of in' pears from this rcsearch
                                                                        of safe behavior.
   creased productivity or productive activity determinant
   ("mc,tionlessbodies" rarely htrve accidents)'          4. Dcgrec of crew work is po$itively cor-
                                                                  with accidcnt rates of the factory
   bn the other hand' this finding barcly bor- relerted
   ders on statistical significance and       similar departments irr this stut{y. This is believed
                                                           be the result of morc complete and con-
   $tudy of anothcr t"actory yiclded a correla- to
                                                                   acceptance of individual responsi-
   tion of .00 between incentive system and tinuous
   deuartmental accidcnt rate'                          bility in the non-cooperativewclrk situation'
       bbt,ious Danger Fador: DePartments                  5. Job prcstigcof typical job firils to cor-
   *ittl a ".tt"ttunt obvious clangcrpresent     such rclate significantly with departmental acci-
                                                                when certain other variables are held
    as glowing molten metal have significantly dents
   higlrer accident ratcs (r : '35) than do other constant.
    deqrartments'An arrcsting iinding by Yr'               6. Thc Breater the tlegree of manual
                                                                                          on the average'
    HLintington,    the Safety Supervisor' on the effort involved, the higher,
    naturc of thc accidents thesedepartments is
                               in                           the departmcntal accident rate'
    is that the accidents  usuallyarc not physical-        7. Both degree of operational congestion
                                                                          incentive work fail to predict
    ly  involved with the obvious hazard factor' and degree of
    For exarnple, in moltcn Inetal         opcratlon departmental accident rates when certain
                                                                            are held constant'
    dcpartmcnts, accident rntes are above aver- other variables                          "obvious danger
    agi but thcy rarely include burns'            The       8. The effects of an
                                                                   on departmsntal accident rates are
     obvious hazard seemsto cxcrcisc irn almost factor"
                                                                      uncettain, although obsetvational
     hypnotic effect upon personncl, dclimiting somewhat
 PSYCHOLOGICAL AFFROACHES                                                                313

 evidenceindicatesthat wherc an obvious obvious hazard seems to corrtribute to
 dangerfactor exists,the accidentswhich "unrelated"irccidents delimitingatten-
 occur tend nol to be identifiedwith the tion and encouraging
                                                            pronenc$s involve,
 obvious danger,Exi$tence an impressive ment in the non-obvious
                         of                                     hazards.

      The hypotheses   selected study, taken togrther, are assumed be indicativeof
                                for                                       to
 the psyc:hological  climate of a factory. Examination of their content, however,shows
 them lo bc a rather hctcrogeneous        mixture of physicalarrd environmsntalfactors,
 w_or\i1q   conditions,personnel     practiccs,and soc:ial  and psychological  factors,some
 of which are founcito relateto the iicciclcnt    latc. It is difiicult,however, view this
 listing as mote than a series miscellaneous
                                  of                 factors,and there is little attemrrtto
 definewhat is rneantby psychotogical         climate in so far as this relarcs accident
      In line with the foregoing criticism, the meaningsof the rated variableshave not
 leen -suffiliently analyzecl permit an interpretation thc findings.lt would haye
                              to                              of
 beenhclpful if the judgesthemselves beenstudicdto <leterrnjn*e meanineof
                                           had                               the
 theseratings to thcm. The high correlations bctweensome of the variablesrug!*rt
either that the intersorrelatsd      vrtriablcsharcd common factors or that the ratcrs
correlated    them in their own thinking. Because judgeswere factnry supervisors
who were probably already acquaintedwith the accidentrecorCsof thc rlepartmenrs
 thcy wererating, it is quite possible    that their appraisals  wcrc pcrtinently'hiascd.
     Iinally, we are told almost nothing about the 44 shop departments tcrms of  in
such characteristics quantity or quality of exposureto hazarc{,
                        as                                                  complctcncss    of
accidentreportin$, or the sex,age, experience,       and social ancl psychologicalcharac-
teristics the workmen.UnlesswE can assume
           of                                         that all 44 departnrents werealike on
these charactcristics   (which ir doubl,ful),inierdepartmentaliiiTcrencesrnighr rclare
both to varying psycrhological      climatesand to varying accidentratcswithgut therc
being any causalconnc,ction      betweenthcrtr.
    We menfionedin Chaptcr'2 that expertsive           data and much valuabletirne have
often been needlessly      clissipated insutficientl-y
                                        by                 rigorous mcthodology.This is
frequently s€en in psychologicalapproachesto accidcnts,espcciallyin work that
applies   psychological  concepts greatsubtletybut tails to pay adequate
                                  of                                             artcnrionto
suchbasic    consideralions thc sclection
                             as              ofcaseand controlgroups.      This is illustrated
by the interesting   work.that follows.

  -lrn'in M. Martus, wilma wilsrtn, Irvin Kraft, Delmar,grnarrdr'r.
                                                                  Frm!        southtrland.
     Etlith fuhulhofer

   Perhaps one of the most intensive clinical studies on psychologicalfactors in
 ccidentcausation, this work was carriedout by an interdisciplinary     of
                                                                   tearrr physi-
 ans, psychologists, psychiatrists,
                                  and social workers concernedwith accidents
 mong children.They compirlcda sampleof 23 chitdren,   agecl to I l, who had had
 at leastthrec major accidents"with a control group of 2l childrenwith "no more
3t4                                                    PSYCHOLOGICAL APPROACHES

than one rnrrior injury" and, also, with a conl.rol group of 23 enureticchildren with
"no morc
            than one seriousaccident." This last corltrol group was selected rcpre-     as
scntinga presurnably    maladjustccl     group of children. Parents       were irlso includedin
the study in an atl.ernptto invcstigateintrarhrnilyfirctors.No control group of
childrenctlmplclclyfrce fiom sclirlusaccidents                       In
                                                         was usecl. addition, thc ntethods
usedl'ol the selection the cases
                        of             and controls,includingthe rrse "voluntccrs"for
the nonenuretiscontrols,makc it cxtremclyunlikely thirt the three groups camefrom
thc samcpopulation.Consequently,           there probably also wereassociated         qualitative
and quailtitativccliffcrences their exposure risk and otl,er pcrtincntcharacteris-
                                in                    lo
tics.As a result,it is rluite possiblc(hat thc intergroupditl'erencres         oLrservecllcsultcd
in part from biasesin their selection                                                of
                                             and not frorn thc characteristics prirnary
intcrcst.This lack of dcsignsophistication in markeclcontrastwith thc approach
employed    oncellrc subjects                   lt
                               wclc sclcctcd. is an especially        unforl.unate  shortcoming
consideringthe resources      brought [o bcirr in the study of these children and thcir
    The study focuscdupon thc effectof maladjustmcrltupon the occurrence acci-            of
dents.  Thc maior research    cluestions   includedthe following:
     l. Can the child who has repeated       accidents distingLrished
                                                          bc                 from otherson the
basisof pcrsonalityfactorssuchas intelligencc,            clxotionaldisturbance,    and adapta-
tional difllculties'l
     2. Does physical functioning, health, motor cootdination, or activity type con-
tributc to thc occurrencc the accidentpattern?
     3. Do intrafanily factorssuch as broken homes,degree control excrtcd,and
emotionaldisturbance parentsdistinguish
                          of                          the accident  group from the others?
     4. Is accident lepetitionrelatedto a particulirrkind of crnotional          disturbance,   or
is it rather a response crnotiunaldisturbance
                         to                               which rnight in othcr circumstances
have led to selectionof a difl'crentbchirvioralpattern or symptom?
     The method includcclboth detailed clinical interviewsand long battcriesof
psychologicaltcsts. Thc parents were given socrial-work                interviews,including a
developntenl.  historyfor theirchildren,    and a scries psychological
                                                            of                tests.The children
were giverrphysicalcxarninations,         psychial"ric   intelviews,and the Stanford-Binet,
Rorschirch,   Lcvcl of Aspiration, Draw-a-Pcrson,         Accident Apperccption,and Spring-
field-Bcame    Walking tests.
     In general, results
                 thc         favor acceptance the hypothesis
                                                   of                   that maladjustment a  is
major factor in accidents                                                  was
                             among chilclren.The ac:cident-subjcct found to expcri-
encea Foorerover-allacl.iustment the homc anc{ school.He wasfounclto resemble
thc cuureticgroup much rnorcclosely         than the acciclcnt-free    con(rolgroup.
     This is one ol'ths mostcomprehensive        clinicalstudies childhoodaccidents--an
area which,in general,    has received    much lcssattentionthan adult accidcnts,          despite
the greatcrrclative incidence accidents this agc group. Without invoking thc
                                   of              in
dubious conccpt of accidenl.       prollclress, attcrrrptcd determinethe personality
                                                it              to
structure of the acciclcnt-liable      child by mcans of comprehensive             psychological
tcsting.We haveomittedthc dctaileddescription the testscontained the original
                                                          of                      in
79-pagereport and prcscnt only the portions concerned               with thc selcctionof the
subiects, theoretical
           the             framcwork,anclthe discussion the results.
 ACCIDENT PATTERNS IN CHILDREN                                                                                    315
 THn REcErrToF A cRANr from the United                                   of the father was not made a condition for
 S t a t e sF r r b l i c H e a l t h S c r v i c ei n 1 9 5 4 p r o -   theacceptanceofthcfarnilyirrthestudy,but
vided thE opportunity for a more extrnsive                               hc was strongly urged to participate and did
 study of thc accident-repeating                     child. White        so in a high percentage cases.
continuing to use the conceptualizations                        de-
vcloped in our Pilot Study about discrepan-                              SueJrcrs
 cies betwecn the child's self-conceptand the                                All children who served as subjects had
 coflcept which his parcnts hcld aborrt him,                              passcdtheir sixth birthday but had not yet
an atteilFt was tnirde to employ this infor-                              reachedtheir cleventhand were white. These
 mation irr a ntore comprehensive                        study of         criteria were imposedto insurethat we \vere
farnily inter-relationships. ordcr to cx-    In                           working with a relatively horrrogeneous
plore these relationshipsand arrive at an                                 group ofchildren. This particularage range
 undcrstanding of the undcrlying dynarnics                                was sclected becauseof thc importancc of
 ofthe accident-rcpeatingchild, it wirs neces-                            accidents in this age group and trccauscthe
sary to obtain a much more comprchensive                                  children of this age were suflicicntly rnaturc
pictur€ of the personality of thc parcnts as                              to be anrcnable to psychological testing of
lvcll irs of the child in his day-to-dayfunc-                             the type which we wiinted to employ. [n
tioning. In order to achievethis, the study                               addition to thesccritcria, it wirsdecidedthat
was expanded in severalrespects.                                          no child be acccptedfor participation who
    Our expsrimental design remained the                                  did not havc a Stantbr<,1-Binet of at least
same except for the addition of a second                                  80.
control group, This group consisted of                                       Acddent $yb#lts: This group was com-
enuretic children and their lhmilies. We felt                             posed of children who had sr.rflered least
that it was importartt to determinenot only                               three major accidents, last of which had
whetherthe ilccidcnt-rcpeating                     child dilTe  red       occurredwithin six months of the date of his
from the norrnal,but alsoto find out wheth-                               participationin this study. A major accident
er hc showcd charactcristicswhictr might set                              wirs dellned, for purposcs of this investiga-
himapartfromothcrcommonlyseenclinical                                     tion. irs one rcsulting in an injury which
problcms.Enureticchildrenwere chosenfor                                   required medical attcntion more than one
this group because widespread  uI                      agreement          tinre or which resultcdin hospitalization  for
that in the agc rangeof our study this prob-                              a period of 24 hours or morc. Potentialsub-
lem was indicative of adjurtment difficulties                            jects for this group were obtained by referrals
and becausc,            unlike lessclearly defined be.                    from private physiciansin the community
havior or characterrJisturbances,                     there could        and lrorrr the inspection of enrergencyroorn
bc little question as to lvhcther a given                                treatmentrecordsoflocalcoopcratinghospi-
child was properlyincluded in the classifica-                             tals. In later stages our investigation
                                                                                               of                  four
tlon.                                                                    New Orleans public schools wcre surveyed
   In order to obtain an over-nll view of the                            in an attempt to locate additional subjccts
functioning of thc farnily, it lvas rece$sary                            in this group, The parents of potential sub-
fot us to 5sgs1s           data from the fhthersof our                   jects rcccived a telephonc cilll from our
$ubjects.Thc I'nthcrs,then, as well as thc                               social worker, who at that tinre determincd
mothers of our subjects, were seen for                                   whether the child involvcd met our critcria
structuredinterviewscrrnductedbyapsychi-                                 for inclusion of the sturly. If hc did, the
atric sociirl worker and were given a battery                            family was invited to participate, being
of psychological tcsts designedprirnarily to                             offercd i1surnpletepsyctrologicaland physic-
rcveal personality dynamics. I)articipation                              al cvaluation of the child as an incentive.

        l-Reprinted, .vith permission,from Monograph.r the Su'iet.t'
                                                     ol            .fur Re,^tttthin t'hitfl
        lDavelopment,  Vol.25, No,2, ScrialNo.76, 1960, l7-19,39-54.
                                                        pp.             Published the I
        L                     Socictyfor Research Child l)evelopment.
                                                in                                       )
316                                                  PSYCHOLOCICAL
   Twenty-three children, including three      increase the chances for survival and pro- |
sets of siblings, conrprised the accident      motc the welfare of thc individual. An I
group. All 20 of the mothers and 12 of the     accident-repeating pattern obviouslY does I
fathers also parlicipatcd in the study.        thc opposite.                                    I
   Enuretir Group: For inclusion in this           Although adaptation has always been a I
group a child was required to be prescnting    central theme in psychoanalysis. focus I
the symptom of persistent and regular          hirs, until recently, been applied to disturbed I
enuresis.It was also specifiedthat he should   behavior. Healthy adaptive mechanismsI
not have suffered morc than one nrajor acci-    remainedimplied. Adaptation implies inter- |
dent. Most of these subjects were children      action with the envilonment; thus these I
who had been ref'erredby their parents or by    patterns cannot be studicd exclusively.in I
various comrrrunity agcncies for clinical       icrms of internal forces' With this in mind I
 services, complaintsconsistingin whole
          the                                   and with the known basic dcpendency all I of
 or in part of enurcsis.Some, however, wete     childrcn upon their adult cnvironfient, we I
 drawn frtrtn thc public school survcy men-     fclt parentsshould be included in our child
tioned above.                                   behavior study. The child's need for protec-
                 children, including one pair tion, care,nourishment,      and mental stimula-
                                                      undcr the sheltering  canopy of lovc are
ofl sihlings, comprised the enuretic group' tion
All 22 of the mothers     participated in the wcll known. Hnman adaptivc ability is dc-
study, as did I I of thc fathers'               pendcnt upon thc individual's basic endow-
                                                                      to devclop memory, per-
    Control Gtoup: This group was drawn ment, his capacity
entirely ft.rm volunteers    from thc New ceptions, judgment' activity type, etc.. as
                                                                        on mentil activity such
Orlcans public schools' Thcse were children well as actions based
who had suffered no morc thau       one major as defensiveor protective mechanisms,and
                                                                            when expcriencing
injury and who had shown no disturbances thc integrative capacity
of behavior or adjustment sufficicnt to cause    fcclings, thoughts, and impulses toward
                                                                 problern solving techniques'
 referral for psychological or psychiatric motor action,
 evaluation.                                     etc. All of these factors arc part of the com'
                                                                       we call adaptation.
    The control gtoup was composed of 22 plex phenomenon
 children, including  thrcc pairs of siblings'      The child's earlicst awareflessof the re'
                                                                  survival appear in the form
 Complcte data were securcd on 17 of the quirements for
 mothers and partial data on the eightcenth'      of unpleasant-painful tensions and pleasur'
                                                                            comtnunicates his
 Sevcnteenof the fathcrs of this group also able expcricnces. He
 participatedin thc studY.                        nceds at this early stage through his neuro-
                                *                 musL:ular system, in crying and general
                                                  physical activity. If thc parental environ-
 THcoRurtclL Fs.l'r'lPwotrc                       ment responds adequately to his helplcss.
                                                                     he cxperiencesrelief from
    lf wc can discoverthe dynamicsinvolved dependent stiltc,
 in acciclentbchavior, we have    a beachhcad tension, satisfaction,     atrd pleasure.His dif-
 for cventually understanding  causation'But' fuse motor activity again returns to rclative
                                                                     growth and development,
 in order to organizeour resultsin a meaning- quiet. Through
 ful m'rnner, it will bc necessary  to present his mental activity serves to integratc and
                                                                               of his needs and
 briefly the theoretical framework within differentiate his awareness
 which our studY was conducted.                   his patternsfor adirpting to the environment.
    Just trsadaptationis hasicin thc theory of As can be seenin thc foregoing
  evolution, so is it essentialto the undsr- general framework for considering
  standing of behavioral pattcrns in the  indi- functioning, we view the origin of the so-
                                                         "ego" from an undifferentiated phase
  vidual. Healthy or adequate adaptations callcd
                                                                             psychic apparatus
  would constitutethosc dynamic pattern$of of brain function. The
  intcraction with the    environrnent which develops from this phasc and later conres
 ACCIDENT      PATTERNS       IN CHILI]REN                                            1r7
                           "ego." This
 under the control of the                 for either the individual or the interpersonal
 that rudimentary adaptationalapparatusis  unit. An unhealthy or maladaptive new
 present in the infant, although lye sec them
                                          level of equilibrium would contributc to a
 only as primitive regulatorymechanisrns; higher tcnsion level by interfering with dis.
 maturation they gradually f unction with charge oftension associated    with the achicve-
 more complex but rnore effective mecha-  ment of the appropriate gotls. The ever
 nisms. In certain instances the primitivechanging new levels of equilibrium should
 re$ponse$   nray remain but are influcncedby
                                          serve to aid in survival, relieve painflul
 maturational nrechanisms, they may be
                              or          tension,and permit plcasr-rrp   experience$  and
 replaced by the newer rcgulatory apparatus.
                                          utility valucs in goal-directed   activity. Dis-
    Our formulations regarding mental mech-
                                          lurbances in cquilibriunl systemscan thus
 anjsms in the child with an accident pattern
                                          easily lead to disturbances in reality-
 integrirte various cancepts such as those of
                                          relationships.   The ability to utiliec thought
 Hartnrann,Kris, and Lowenstein,Rado. and pr"occsrjes judgment for anticipating the
 many others who have enlarged upon       l'uturc situation constitutesthe highest de-
 Freud's basic contributions. We wished tovelopmentallevel in progressive     adaptation.
 avoid the issue of whether behavior is bio-
                                          Whcnever the tension level paralyzes or
 logical or psychologicaland concentratedon
                                          diminishesthe anticipatingmechanisras,        we
 factors that jnterac[ from perhaps congeni-
                                          would expecta r.r,cakcning    o{'rcality tdaFtd-
 tally or constitutionally acquired ftatura-                    <rf
                                          tion and a lcssening the survival value at
 tional and environmentalsource$.     The indi-
                                          the new lcvcl of equilibrium, We acceptthat
 vidual may rcact to a given situation either
                                          there arc constitution;rlfactors influencing
with previously dcvcloped modes trf adapta-
                                         thc individual's ability to tolerate anxiety
tion or by utilizing ir ncw pattern, Thus, one
                                          and his reactivity to stimuli. Fries in 1935
 may speak of "regrcssivc" patterns and   explored thc congenital activity tyFe, re-
"progressive" adaptations,
                              the latter being
                                          mnrking on the amount of activity seen in
consistent with maturation and devclop-  an infant in responsc certain stimuli. The
ment" It is understood that rcgrcssivc pat-
                                         excitability of, thc ncuromuscular system
tcrns may setvetow$rd a progressive    healthy
                                         varies with congenital, maturational, and
reality aclaptation,since memory and fan-environmental factors of which the parental
tasy activity reach into the past history and
                                         attitudes are especially important. Within
may be utilized for reality achievemcnts.It
                                         the normal range infants may show quiet,
follows that tho$e individuals who lean on
                                         moderatelyactive, and active patterns; and
regressivetechniqrresare flot doomed to aat the two extremesthc pathological vari-
lifc of unhealthy pattcrns if progressiveations of a hypoactiveand hyperactive       t),pc,
developmental tcchniques can be learned  These two extremes are considered more
and applied.                             vulnerable to severe personality disturb-
   The concept of tension in the child leading
                                         ances, although any typc may develop
to an environmental responserelieving thcpsychopathology.
child was mentioned above and indicates     It will take many years of further work by
our view of the child's tension systcrn on
                                         those in the field of child rescirrchto clarily
another level. The child-parentrelationship
                                         the relationship between congenital activity
      itutes a clynarlic whole. ln all normal
                                         lype and personalitydevelopmcnt. cannotJt
situations the relationship is "disturhed"
                                         be expectcd that thcrc is any one-to,one
at frcqucnt intervalsby rnultiplestimuli. and
                                         corresporrdence,    nor that one could study
 new equilibrium musl be achieved.Each this rclationshipin isolation fronr other in-
 dividual sitnilarly experiencesrensions fluences.Nevcrthelcss, nrust include the
  hin himself which necessitate another concept in our framcwork in terms of its
 adjustment internally. The new equilib- predisposingelfect upon adaptational mech-
 um may either bc hcalthy or unhealthy anisms.The child's activitv Datternrefersto
                                                                                  PSYCHOLOGICAL APPROACHEST

his actual way of responding under the the higher'*fear reflex" during rrraturation.
various mod ifyirrg iul'luences   durirrg matura- Fear in this instance is predomirrantly an I
tion. The child who is dillicult to sntisfyduc intellectuat reaction geared to anticipating I
to extrcmcsin activity patternswill frustrate pain from impending darnage.Rado's ttreory I
himselfland his parents in the process of of cmergency behavior passesover the dis- |
establishing a mutual relationship. With tinction betweenatrxietyand lear.l Ilis em- |
impairmcnt of child-parent functioning phasis is more on the rolc of pain, fear, and I
thcre would be a disturbance in object rcla- ragc as the alerting tesFonses                                     which prornpt
tionships and r-calitytesting,thus a grcater the individual into adaptational activities I
tenderrcyto rely on magical thinking and for the prevcntion and rcpair of dilrnage. I
personalomnipotence.       The child who caunot Thc indiviclual experiencing persistent at- |
establish    closctresswith his parent$and who tacks of anxiety will rationalize thcse into I
failri to ma$ter anxicty through irrnitationand mortrid fears, We may view this as an inter- J
identification   with his parcntsmust rely tlore fercnce with progressivc ego aciaptational
on kinesthetic irlcntification and various mechanismsbecau$c,although no dirnger
motor activity patterns. The more orre relies exists, the ego is exhatrsted by conflict,
 on hyperactivcInotor dischilrgcto re'estab- flight, and other adjustive mechanisrrrs.                                                 Self-
lish equilibriun" the mot"clikely it is that the injurious actiLlns are viewed as a pain-
 individual will move irrto situatious which eliminating respon$cembodied in the
                                                                 'principle,"                                          "get
 mrry provoke disapprdv;'rlor, if overwhelm- dance                                       the wish to                             it over
 ing, lead into drrngcrorrs    situations.           with" by rushinginto the fearedsituationor
    In contrast,the chilclwith a quict activily attacking thc synrbol to whiuh the fear is
 ptrttcrn sccms1o re-e$tablishequilibriurrr by attached.
 a grcater tendency to withdraw,
    The nnticiprrtory role of thought processes Drsr:ussIow
           "ego"                                        Matched groups of 23 accident-repeating
 during            differentiation is of utmost
 irlportance in achieving greater independ' childrcn, 23 enuretic children, and 22
                                                                                  "healthy" childrcn wete com-
 encc lro trrinrrlediate envif ofl mental stim ula- symptom-free
 tion. This concept was describcdby Freud as pared on the basis of their physic:alhcalth,
 experimental action              small quantities dcvelopmental history, psychi{t(ic intcr-
 of energy," Immediatc motor: discharge vicws, psychologicaltests.including motor
  must be delayedby interposingthought pro- function, and their pal'ent$' ltistory, atti-
 cesses defend the individual from poten' tudcs, and psyclrologictrlstatu$. The parti-
 tial damtrge and to insure his ptotection. cipaling parents inclr.rdcdthe following: for
  The greater the tendency of an individual to the accidcnt children. irll 20 of their urr"rthets
  ernploy primitivc del'enscs of itnmediate and I2 fathers;for the enureticchildrcn, all
  rrrotorclischarge, lesslikely will he bc to 22 of thcir mothers and l8 fathcrs; fbr the
                      "trial thought" processes' symptom-free group, 1? nrothers and 17
  delay trctivity by
  Thus, a high activity pattern intcrferes with fathers.
  the function ofhigher order ofdcfcnses. The           The subjects had all passed their sixth
  differentiation of reaction possihiliticsbe- t We difer with this cotrcepl itnd fecl that there is
  comes more limited, selective control is virluc in rlistinguishing belwL:en ittlxicty atrd lhar.
  impaired, and thc irrdividual rcmains en- Fear is a(tilclred to rsality situirtiol)s tntl invr)lves
  slavedby his early stirnulus-reaction    pnttcrn. judgnent.'1 hu tcar nrayr n well-foundedc i n irration-l ,
                                                     al. Anxiety is without                         objcct, vagtl                  colltcn
     The dominant role of anxiety in      its influ- a n d s u b j e c t i v c . B o t h f ' e a r a n d a t r x i c l y l e a d t t t a v o i d -
                            adaptationwas clear- erncererctions, trrrt                 in t'crr thc action tan be rational;
  enceupon progressive
                                                     in anxicty the action i$ unccrt4in atrd trtay itrcrcilse
  lv recognizedby Freud in I926' Anxiety is thc ttisturbance. Fear can terd to construc(ive ilction
  viewed as a reflex responsewith a specific and growtlr; ilnxiety oflcn leads to exhaustiolr'
                                                                                            pcrccpt analysi$ that fear is
  feeling alertirrg thc individual to danger, Piotrowski nolc$ in
                  "anxiety                            usurlly scen as a ncBativc chromatic color response;
  Rado setsthe               reflcx" as a prirnitive a n x i c t y , o n t h e o t h e r h a n d , i s n e v e r i n d i c a t e d b y t h c
  responseto pain, whictr is transformed into         chromatic color rcsFonse.
 ACCIDENT PATTERNS IN CHILI]ITEN                                                                                  319

 birthday but not yet reached thcir elcventh          To a significant degree,the mothers of thc
 and wcrc whitc, Wc sclcctcdan age range accidcntchildren reportedthfit the child was
 which would bc arnenirblcto the type of           very activc in utcro and their rlelivery was
 testing we wished to employ. No chilclren difficult and prtrlonged. Onc rriry wonder if
 were accepted participationwho did not
                lor                               the accidcntchild e.tpericnccil                br.uindamage
 have a Stanford-BincttQ ofat tcast80. Thc         due tr: a difiicult birth. physical examina-
 accidcntchildren had sufferedat least three trons including spccializcd tests did rrot
 major accitlents, the last of which had oc- confirm this possihility. F.urthcrrnore,                               lhe
 curred within six months of the date of his children showed cxccllent ruottrr tlevelop-
 participationin this study. A major accident m e n t a n d n c u l ' o r l r r r r j c uilnrtre g r a t i o n I, n t h e
 was dellned a$ one resulting in an injury Iight of our prcccd thcoretica fi.iirnework,
                                                                           ing                    I
 whir;h rsqult.d rncdical attention more than     a constitutionally high activity type would
 ofle time or which resultcdin hospitttirirtion have an incrcasedlctivity responseto the
 for a period of at least 24 hours.               streris a difficult binh. Their motor clevel-
   In the foregoirrg, we describeda series of     opment ocrcurrlrd       earlier,as scenlrarticularlV
invrstigations pcrformcd by a tcarn of            in thcir sirring u1:illorle nnd staritling.Thcy
workers frorn a variety of disciptines and respondcdto toilet training eveneirrlierthan
 integrated by group discussionsunder the our control symptoril-frec gr()up. Thc
direction of the principal invcstigator.Or.rr nrothcrsofour iiccidcntgroup reportedthcy
desire was to carry on an unbiased scarch checked on their children rnore freqr-rcntly
ftrr significant correlotions in the study of than did thc others. perhaps rs a further
children who wcre experiencing rcpeated rcflection of the anxiety areused in + pro-
seriousinjuries. We attempted to view the tectiveparcnt by a very activc child was the
nature of the interucting mechanisms of rather sftiking fincling that the accident
behavior bnth witlrin the child and between group slept in thc parcntal bedroom nruch
the child and his envir.onment,       plus inte- Jongcr than did thc other two groups. 'l'he
grating the atlditional knowlecige     frorn his greilter Iikclihoocl firr cxposure to "prirnal
life history. We hopcd to lsq.,t rrrore;rhout scene" situntions might provoke lnore
how the chiltl who nranifcsts an accident- a c t i v i t y i r r h e c p i n g w i t h t h e c h i l d ' s u s u a l
rcpcatingpatternadaptsto his liflesituation, pflttern and bccome a further source of
   Thc rrredical cxarninations   for evaluating irritation to tlrc parentsand anxiety for the
physical status were carried out in ritirny child, The persisrence slccp clisturbances
spheresduring the pitot study. Thc gcncral was indicated by rhc fact that significantly
pediatric exirminutions, electroencephalo- nrore accidentchitdrcn still rcquircd a night
grams, ophthalrnology and heirring tcsts light and were morc inclincd to havc night-
revealcd no signihcant differences betwcen marcs and to fall out of bcd occasionally.
the accident subjects and thc control               I t i s o u r i m p r c s s i o nt h a t t h e c o n g c n i t a l
symptom-free    group. In the rrrajorstudy wc activity type in thc nccidcntchiltl rontinues
limited the rncdicalhcalth cxarninations a to influcncehis adaptationalmcchanisms.
                                            to                                                                      In
generirlpediatricsurvcy unles{spccialmedir k e c p i n gw i t h t h i s p o s s i h i l i t yw c v i e w e dt h e
cal tes(swerr dermed tlet;essi{ry the pcdia- shriding rcsponscsin the Itorschach, The
trician. Following the pilot study, we adcled accidcntchildrenwereclearly d istinguishable
a sccond control group-cnuretic children fronr thc othersin their useol'hcuvyshading.
and their pilrents. We were intcrcstcd in The statisticalsignificance                      was bcyond the
comparing the mechanisms behavior and . 0 O l l e v e l .l f t h e c h i l d ' sc o n s t i t u l i o r r lllr c t i v i -
the environrnental   situation of the accident ty pattern had not beenhigh, onc coultl con-
child with both the symptom-fice "healthy ceive of the increrrscd                activity as a conrpen-
child" and the child with a known problem, $atoryand anxiety-relie                 ving rrrcchanism.              1-he
since bed-wettingwithin the age group we significanceof the terrdencyto rcspond to
studied is generallyconsiclcred    indicariveof anxiety with overt nrotor activity rvas dis-
adaptational difticulties.                       cussed in our theorctical frarnework. Thc
 320                                                          PSYCHOLOCICAL APPROACHES
                                         "ritldrrnce    anisms of thinking in magical terms and I
  individual is overwhelmed by rr
                                                                      his own invincibility and ornni- |
  responsel'to remove or changethe anxicty- belief in
  prodr,cingsituation' Under the influence         of Dotencc. Thus, his action tendcncy is inten- |
                                                                 try rcgressive tlrought patterns and I
  rction-taking rleasures there is a keener iiti.O
                                                                       by the reduction of his anxiety'
  awateness onc's strcngthin contrastwith reinfitrceil
                 of                                                                                           I
  the fccling of hclplessness tctrsion
                                    ancl        when        It is well known that there is a widc range I
   action must bc inhibited.Thus, our                   of symbiotic and anaclitic relationships I
                                                                              and child, dcpcnding uFon I
   prone"z subjects are more likely to take between mother
                    'act hastily and irlpruderrtly, the needs of thc mother and hcr child' The I
   "h^n."*, to
                                                                  the relationship,the more the child I
   since the anticipatory thought rcactions are closer
                                                               utilize his mother lor reality testingand I
   supcrseded irnpaired. This acJaptational will
   pr{ttcrn woultl corresponil to oul dcscription       mastery ol the environlnent. The more ex- I
          "regressive adaptrrtion'"                      periences childrcn havc_       with their Parents'
   6rf                                                                                                         J
        We feel that our study of the activity ihe            greatcrthc opportunity for idenlification'
                                                             ()ur pilot study revealedthat        there was
   pattcrn in thc accidcnt-pronesubject sub-
   itantiatcs Piotrowski's      theory on the signi- sreater acceptancc and affcction
   ficancc of dark shading in the Rorschach'             f,*,**.n the control mothcrs and their chil-
                                                                 whereas the accident group showed a
    His conclusion that dilrk shaclingis less dfen.
                                                                     divergence in attitudes and less
    liequently secn than light shading is con- greatet
                                                people mutual ilL:ceptnnce        and allecliorr' There was
    sistentwith the known fact that nrost                                                          and feel-
    are sufllciently reality     oriented ttt restrain also grcater unaniLnityof lhought
                  action$and adapt to the environ- ing amorrg tlrc            control groull rlothers and
                                                                 childrcn tlrirn atnong tlre crpcrtmental
     rnent in a morc cautious marlner' Piotrowski their                                                 Test
     found the dark shading        tnorc promincnt in group. TIle Accident Appcrccption
                                   epilcptics. paticnts icsignea      cluritrgor;r pilot study was appliecl
     antisocial psyclropaths-
     with head traurnatiL.   adolescents,  pioneers  of during thc rrrajorwork. Consistentwith our
                                                                                             of the accident
     all kinds, volutrteers   for dangerousmilitary aboue finditrgs,the mothers
     service, and supelior individuals        who arc childrcn vicwed thc child's bclrrrvioras dc-
                                                                         rlisobccJient,  while thc control
     1g6ders,   rcformers, aud great doers' We can ceptive or
                                                                      saw thc child's generalcotrduct as
     now acld the accidellt-pronctypc to this n,others
                                                                                       Perhaps its ir conse-
      group. lt follows thrrt the individual who hclpful or compliant.
                                                                     of this   attitude the ircciclentgroup
      requires overt rrction as his rnechanism lor ouence
                                                                                      child ns guilty or un-
      re-cstablishing  equilibriumis more willing to mnthers depicted the
                                                                                             attentitln frorn
      accept risks ind more       likety to handle his happy after lre had received
                                                          his mother.     They were also more likely to
      anxicty with expressionsol belligerence'
                                                                     overtly criticism of the child'
          Wc found the symptom-free subjects ex- e,(press
      periencitrg a bctter   total acljustment, better        Supporteclby our statisticalirnalysis,we
                                                                                   parents, particularly the
      reality orientation, a greater ability to em- found-the hcflIthy
       pathize with others, less anxicty,        and a mothcrs. revcalcd a morc positivc attitude
                                                                                  in the interviewing pro-
       supcrior social arJjustment irr the htrrne, towtud children
       school, and ovcr-all situirtions'Thcy      main- ceclurcs,     They tcndcd to want children more,
                                     "ego" integration, to have planncd childrcn' and to wirnt a
       tainetl a higher dcgrecof                  "pro-
       showing what we earlier termed                      Iarger number of children than did thc other
       gressive'.tdaptational"     patterns-               t*o groupt. Lven prior to having children,
                                                                           of the healthy group showcd a
           The accident srrbjcct' on the other hand, the mothers
       with a tendency toward werrkerreality-bound         greater liking for other children than did the
                                                                                       two groups' More of
       integratiorl,  would resortto his earliermech- inothers of the other
                                                           the rnothers      in the accident and snurctic
       2 Thistermp65soined theprincipal
                              bv              investigator group would have prefcrrcdtheir prcgnancy
        (I.M.M.) as a construct the dynamic
                                 for              system
    described herein.
                                                           at a later time. There were more mlscar*
 ACCIDENT PATTERNS IN CHILDREN                                                                                 32r

  riages among the mothers of the accident                             parents were not openly demonstrative, in
  group. There was a tendencyflor rlore ofthe                         contrast with the honresfrom whiclr the con-
  mothers of thc accident group to rep()rt                            trol pareniscame.The fathcrsof the control
 post-partum depressions.             The accidentgroup               children were more dissatisfiedwith their
  (children antl nrothcls)expresscd               emotional            levcl of educational achievcmcntthan the
 insecurity and unhappiness,whcreas the                               accident group falhers. The courtship his-
 healthy group madc a better emotional ad-                            tory of thc parents of the three groups did
 justrnenl. The mothers of the healthy group                          not reveal any significant .differences,and
 engagedin more parcntal activitieswith the                           the sexual adjustme in marriage was
 family and nore individual activities.                               similar for each group. More than half the
 Healthicr families appear to engagein acti-                          mothcrs in each group reporteddifficulty or
 vities which promote further ftrnrity health.                        lack of satistaction in sexual rclations. a
 Thcy had signific:antlymore rcligious frarti-                        figure consistent with known survcys of
 cipation and more social activity involving                          married women in our culturc.
 religiousgrolrps.Our findingsin the irrcirof                            Thc lowcr responsivcncssof certain
 family background do not bear out the                                parcnts to f,anrilyprojects is reflected in the
 reports of other investigators who empha-                            percentagc of their participation in our
 sized the authoritarian strictncss through                           study. Only 60 per cent of the fathcrsof the
 religion, We found greater evidence of                               accidentgroup wereinvolvetl in our study in
 religious intcrest in the control group.                             contrast with 82 per cent of the cnurctic
    More of the mothers of the accidc-nt                   chil-      group and 100per cent in the healthygroup.
 dren reported they cxcrcisedno guidance in                           The over-all picturc of the parents of the
 thcir children's use of thcir allowance.Both                         accidentgroup suggcsts   that they were more
 parents in thc healthy group took a more                             anxious, insecure,and nonassertive.       They
 active role in supervising            the child's school            inhibited thcir own fiction-tyFcpirtternsin
 work. On the testfor levelsof aspiration,the                        exchange acceptancc othersin society,
                                                                                for              by
control group fathers wcrc significirntlymore                        Thc parentsof thc enurcticgroup, especially
 realistic,whcrcir$the healthy mothers set a                         the fathers, were rnore donrinating, ag-
 considerably        higher level of aspirationthan                  gressive,and controlling.
did the other two groups.Frorn thc Draw-a-                               This study of the i,rccidentchild does not
Person test, the xz y*1u" obtaincd shor.ved                    no     bear out the thcorjcsof thosewho cxplnined
significantdifferenccs            among the groups of                accidcnttrehavioras pri rnarily'tuncorrscious
 mothers, but fathers of enuretics showed                            suicidc" or hostitity turned inwurcl and
significantly rnore "aggressive"traits than                          gcnerirtcdby revolt against inhibiting, au-
did othcr fathcrs.On the AAT thc accidcnt                            thoritative, or punitive pilreflts. The meirn
fathersshoweda tendency depictthc childto                            rating$ of the accidentand enureticsubjects
as making a frientlly rather than a hostile                          showcd a high degreeof similarity. . . . The
approach to other children. In contrast to                           accident group wrs significantlymore sub-
their children,the accidentparentsproduced                           missive than either the healthy or enuretic
f,cwershading responseli the Rorschachon                             controls, but not apprcciahly different from
than any othcr group and proportionately                             the control Sroup on negiltivisnr.tsoth the
more light than dark responses.ln keeping                            accident anrJ enuretic groups produced re-
w i t h t h e e a r l i e rc o m m e n t so n s h a d i n g ,t h e   cords indicativeof nn emotionaldisturbance
accidcnt parents' tendcncy secmedrelated to                          and more anxiety than the healthy group,
inhibition of action. Social histories indi-                         The majority in thesc two groups was
cated that parentsril accidentchildren carne                         evaluated clinically as behavior disorders.
from homes in which one or both parents                              Almost all the subjectsirr the two groups
werc absent.In general,             the mothersofacci-               produced fire, water, blood, and explosion
dcnt clrildren reportcd they received little                         responseson thc Rorschach, The healthy
allection in their own childhood and their                           group gave relatively bcnign content: ani-
 322                                                   PSYCHOLO(iIC]AL APPROACHES

 mal, human, household,etc. Enuretic chil- depcndency            and submissivenes$  during pha-
 dren gave a prcponderancc food respon- ses of inhihitcd action, rlounting anxicty,
 ses,bu1 only the trccident    subjects produced and depression. is undcrstandable
                                                                    It                    that the
 such dysphoric content as "derrd," "ugly,"         self'-punitivernechanism, hostility turned
 and                                                inward or suicidal impulses, conscious or
   In the area of ernotionaldisturbances       the uncon$cious,   may thcn courc into play. We
 accident sub.jccts    were rnorc like the enu- pfopo$e these dynamics as sccondary rather
 retics. l hcy both rlanif'csted appreciably than primary mcchanisnts of adaptation
 lnore anxiety anrl camc lrorrr horncs where during the irccidcnt-prorrc             'fhc
                                                                              philse.      action-
 there were lesshar:rnonyirnd rlorc unhappi- prone individual who becomes           anxious will
 rress.It is the rrrannerof clerrling   with this ofterr gcnerate rrrgeful feelings and rr desire
 hnxicty which dilTercntiirtesbetween these for irction. Our society r-ccluiresrestraint.
 groups, The accident child is the "rrction- Thus, depressionresulting frorn inhibited
 prone" individual. He is rlorc likely to dis- ragc may gencratefurthcr anxiety arrd irrten-
 chargcr  ltostility ancl anxiety through ac:tion sily thc need t'or action. Irlpulsive rrctsre-
 than to withhold thcscfcclings.He rlay pass lieve anxicty as well as serve as a defense
 into an accident-proncphase if his anxiety againstdcpression.          The dysphoriccontcnt in
 lcvcl lcrnains high due to fatrlty interaction the Rorschachwould be consistsntwith the
 with an enrotionallydisturbedenvironrnent. foregoing irnpression.It follows that the
 He then expcriencesreperrted interfcrences pleasurc-pairr        principle opcr:ates  atypically
with ltis thought pr()ccsscs     duc to tlre pre- in "action-pn.rnc"subjects,   The necd to dis-
plrnderance of         "rcgrcssive
                                     adaptrrlional charge tension through actiorr is inhibitcd
 mechanisms." FIe must rid hinrself of the not by thc nccd to avoitl physiclrldamage or
 tension by action. Individuals who irre lcss pain, but rather by the pain of anxiety. ln
 inclined toward rrrotor activity utilize other this scnseit may become a self-dcstructive
 $enscs  such as cxploration with thcir cycs to force to the organism although it is not
contact the environrncnt.The action-prone primarily motivated by the wish to becomc
individual appeitrs to have a clirrrinishcd a casualty.
a.fferent  phirsc of interaction with others. His     Safety eclucationhas shown some results,
intcrcst range ltnrrows ancl he coltccntl'ates but appcars to havc reached a point of
his attcntiorr on sclected envir-onmcntal diminishing returns in many arcas, as wit-
objects or situations, thereby excluding nessedby thc continucd high national acci-
many irnportant firctors which would in- dcnt rate. Othcr studiesas well as our own
flucncehis bchavior. His pcrceptual     field can revcal an awarcncssoi saf'etyprecantiotrsin
become so lin,ited as to cndanger his lit'e. thc accidentgroup. Our pilot tttudyrevealcd
The tendencyto usea rcgrcssive        mechanisrn no discreptrncics tl're concern for cautiorr
and its e{J'cct   upon personality integration among the groups,The nrajor study did not
oftctr produces a clinical picturc of,   "imma-
                                                   inclicateany signilicantclifi'ercnceshctween
tufity," a lrait found by othcrs in studiesof the groups orr scratch injuries. These in-
acciclcnt-prone     subjects.The action-prone juries welc frequent in all and in keeping
ar:cwilling to cndure painful or dangerous with the norrnal expectations for minor
situationsto relievetheir tension.It appears injurics in childrcn, Since the accidentchil-
to be the lcsscrevil at thc time, rather than drcn had suffered significantly tilore cuts,
thc aim ofl enduring suffcring as a self- bruises, l-rlls, and near accidents,we wcre
punitive mechanism. he dcsireto succcetJ        in not surpriscd to l'ind on the AA'f a signi-
actittn goals, if frustratcd, ran prodrrcedc- Iicant tcndency in this gfoup to fantasy
pressive moods. Since thc action-prone more serious accidentsthrrn in the casc of
usually repeat thc proce$s,      the deprcssions thc controls. However, the parents of all
whcn present arc more oftcn irrternrittent groups showed no significanttcndencies              to
than chronicr,Wc worrld expect incrcased fantasy more accidentsand thus could not
 ACCIDENT                 PATTERNS                  IN CHILDREN                                                                                              323

 be distinguished on this basis. We did not                                  more than 75 per cent were on dry roads.l
 find cvidcnce that thc accident child is The National SafctyCouncil Accident Facts
 idcntitying with an accident-proncpirrent, and insurancc corlrpirny studics provide
 nor does hc appeirr to be acting out a self: annual statisticson this same matter.
 destructivc wish orr the part of the porent.                                     Thc younger the individual, thc grcatcr
 Their closeness               and identificationwith the his tendency to rcsponcl to tcnsion with
 family appear to be weaker than in our motor action. Wc bclicve (his accounts for
 control groups. Thc farnily pattern is dis- the known fact that alnrost twice as nrilny
 turbed and contriLrutedto the child's in- vehiclcaccidents would bc cxpected(on a                                   as
 security and anxiety, but the disturbanceis numerical basis) corrre fiorrr the younger
 in no way greaterhere than in other child- driver, thosc under ?5 years of irgc.
 hood personalitydisorderssuch as cnuresis,                                       Young*r childrcn, under five, cannot
 The accident pattern appcars to be a casily control their action pattern and have
 symptom or syndr()me indicative of an limited ability to anticipatcdanger through
 errlotional distr,rrhance                 within $ inilividual thought proccsscs,Prevention would ob-
 who relics on action as his chief defensive viously bc depcndent upon the adults who
 mechanismagainst anxiety.                                                  must tcach thcrn cnution and saf'cty and
    The control of acciclcntsrequires intcr- eliminirte potentially clangcroussiturrtions
 posing anticipatory thought processes                               and from their environment,In support of this
 suppressing             potentiallydangerousirnpulses, "comrnon sensc" impression is the fart that
 The nction-prone subject rrrust learn to more than a third of the victirnsof poisoning
 lscognize his stereotyped pattern for hand- were under 5 yeirrsof agc,Tlre rnostfrcquent
ling anxietythrough impulsivcactivity, lf he accidents under l6 ycars of agc arc falls (3g
cannot interrupt thc vicious cycle opcrating pcr cent), blows and collisions (29 pcr cent),
during an accident-proncphase,he should and cutting and piercing injuries (12 per
 seek help fronr those in thc lield of mental cent). The other cause.iare small in ctrm-
h c a l t h .I n d i v i d u a l s h o h a v en o t y e t h a d a n parison, but burns (3 per cent) are most
accident, but irrc aware of thcir "action- often fatal.
 prone" behavior,should scck nrentalhealth                                       Adolescentsflre known to go through dis-
 tacilities whcnever thcir Iile situation is tinctivc pcrsonality reactions.Ego integra-
crcating excessivctension oI d0prE5si9n. t i o n i n t h e a d o l e s c e n it s w e a k e r ; a c t i o n -
Those who ars irction-prone and cannol prone bchavior rRay become more promi-
improve thcir rnental health nrust learn to nent as conscicncc,                                                           sclf-control,and thought
delay or slow thcir activity pacearrd should proccssesarc tcnrporarily impaired, The
avoid driving vclricles during pcriods of irnpact of these personality mechanisms
scvcrestress.             During 1956,38 per cent ofthe upon accidcnts is bornc out by thc vital
fhtal rrccidentsinvolved spceding vehicles statistics ofl our govcrnment, In the age
and 30 per cent involved drivers who had range flrom l0 to J4 ycars, 44 per cenr of
consunted alcoholic drinks, In our society deathsare due to accidcntsand liom l5 to
alcolrol is thc nrost common I'orm of self- l9 years, 55 per cent. The percsnlsges                                                                            then
rnedicationf'or reducing tcnsion.                                    the begin tc decretrse                          contirrLrously 3 per cent  to
itction-proncindividual conrpounds acci- at 65 year$ and over.
d c n t l i a b i l i t yb y r c d u c i n g i s t e n s i o n h r o u g h
                                            h                                    Littlc was said in the forcgoing rcgnrding
meansthat lurther incapacitate thought sex differences.his                                                      Our study had a prcpondcr-
precesses.          Our discovery that the accident
                                                                           3 Since thc proportion of l'ehicle tnilcs driven under
     ject is prinrarily the victinr of personality
                                                                           thesc circumstuces is usually Ercatrr, we achnowl-
factorsis reinforccdby suchdnta as the | 957 e g d e t h a t t h e p h y s i c a l h a z a r d a s p e c t c 4 n n d t b €
    tistics,which indicatethat approximately d i n r i n i s h e t l h y t h c s r , s t i l l i s t i c s . A l ( h o u g h o u r s t u d y
                                                                           s t r e s s e sp e r s e r l a l i l y f i r c t o r s , w e d o n o ( w i s h t o d c t e r
  5 ner cent of tattt and non-t'atirlar,rtornobilc continucd a:fforti td jDrDrove safely through studies
          rrts occurred in clcar weather ilnd o f n o n p e r s o n a l i t y I a c t o r F .
                                                       PSYCHOLOCICAL           APPROACHES

ance of male subjects. Statistics show males that a physician should at least evaluate the
are involved in more accidentsthan females problem after the proper physical carc has
throughout life. The following cxamples becninstitutedand thc patientis well enough
illustratc this point: from ages I to 4 years, to conversc cooperatively with him. The
16 per cent versus12 p$r ccnt; 5 to 9 years, opportunitics are readily available tD the
25 per cent versusl3 per cent; l0 to l4 years, medical prot'cssion to investigate the pa-
34 per cent versus l0 per cent; 15 to 19 tient's behavior Pattern.
years, 43 per cent versus 11 per cent' We can         Anxiety states act as fuel for ffotor dis'
only speculatethat this may reflcct, in part, charge, Whereas mild flashcs of anxicty
the cultural pattern of inhibiting aggtessive should prime the indivitJual for integrated
action in our girls throughout their develop- behavior to cope with a possible danger,
ment.                                              intensivc or chronic doses of anxiety Inay
     The knowledgc    gained from personality fatigue the organism becauseofaccompany-
 studies of accident subjectsmust bc applied ing physiological changes.Fatigue increases
 to all areas of activity. Although vehicle the tbeling of hclplessncss,          and the fbar of
 injuries caused 1,400,000injuries in 1957, failure feeds back to incrcasc the anxiety.
 with home, public places, and industry The younger the child, the lessIikely he will
 accounting for 4,200,000, 2,000'000' and inhibit direct motor cxpression, If the im-
  1,900,000, respectively' vehicles produced pulse to act is more intense, as in the action-
 the most deaths.The problem ofthc poten- pronc child (probably constitutionally hyper-
 tial accidentdriver is a continuingchallenge. activc), the dclay in rnotor behavior through
 The death and injury rate from vehicles is thought processesis that much more diffi-
 increasingin spite of better road engineuing cult. Parents and tEachers must understand
  and other public improvements' Our find- the child's pattern and tcach him ways of
 ings suggestfurther measutesare ncccssary solving problems by thoughtful, calmer
  to interfere with the action-prone tendencies mcans.
  which may impair the thought processes        in     Greater attention to the child's personality
  the potentitl accidentsubject. Any mcasure$       is requircd in the home. Anxious parcnts
  that will inhibit the cycle and intensify cannot teach ways of coping with anxiety
  thoughtful behavior would diminish the situations without the child's feeling the
  liability to accident occurrence' Wc know stressof the parents' anxiety. Childrerr with
  that hcavier policing has decreased   the acci- accident patterfls will requirc family partici-
  dcnt ratc in certain areas. Policing may be pation in thcir mental health progtam' Dc-
  thought ol'irs analogous the parenttspro- pressivemoods which follow in the wake ol'
  tectivelimitation of the child's conduct' The inhibited impulsivity must be expected and
   action-prone individual, burdened with ten- dealt with so that children will not continue
   sion or depression,relies on regrcssive adap- the vicious cycle of eflcl, ncw impulse action.
  tational pattcrns and is a$ limited in using The goal is not 1o convert action-prone
  his thought proccsscs would be a child. children into inhibited obsessional
  We must     thereforeprotectthosewho cannot but to increase thcir cirpacity to tolerate
  control their bchavior during stress.              stre$sand to permit spontaneity with some
      It is clear from our  studies that, where self-control and thoughtfulness.This cannot
  accidentshave tirken placc or near tccidcnts be accomplished by parents and teachers
   are occurring,thc family physician,farnily, unless they are aware of thc nretrrring        of the
   or friends of the victim should advisecloser action pattern they sec in the chilcl. With
   study of the individual's personality func' understanding of the mechanism
   tion and his   problctns' It is currctrtly rare children rnay be slowly guided into hcalthier
   that a mental health professional would be patterns of adaptation to stress.Any
   called in as a cotrsultant for those  who are is potentially capable of an irnpulsive
   being treated for a serious accident' We feel responss to tension and depression;
ACCIDENT       PATTERNS       IN CHILDREN                                              325

actiou-prone are more vulnerable becauseof     workers who utilized structured afld un-
probahly constitutionalfactorsand repflated    structured interviews and by psychologists
patterrrsof inrpulsive action during develop-  who administeredtests,somc stiindirrdized,
ment. This concept does not imply that         some designed speciflcalty fot this study.
acciderrt-proncnc$$ a fixed p€rsonality
                      is                       When results lronr the diffcrent disciplines
trait, but rather that a variety of individualswere intcgratcd,thrJaccidentchild appcared
may at varioustinresin their livesfall into an to be distinguislrablc  from others studied in
action pattern which may lead to an acci-      the frrllowing ways.
dent. The action-ptone individual would          PsJrlro/ogrlq/ rfti4'ors: The accident sub-
more likely be involved in an accidcntthan    jects would appear hcrc to be msre like the
other activity types, ulthough he may not be   enureticthan likc the symptom-free    children
an accident repeatcr,Schulzingerfound that    in thc incidencc of adjustment difficultics.
accidents due to aggressivebehavior were       While sevenof the 22 normally functioning
nearly twice as frcqucnt in "irresponsihle-   control subjects    displnyedsorle :rnxiety,all
maladjusted" individuals. We propose that     of the ?3 accidcntand the 22 cnuretic chil-
thosewho impressothersas                      dren gave evidenceof emotional problems.
are in elTcctthe                  subjects.The Diagnosesranged from behavior disorder,
individqflls who remain accide        nt-pronesome with neurotic traits, to schizophrcnia,
throughout life would be those who fail to    found in two subjectsof crrch of the cxpcri-
develop othcr defensesand solutions to        mentrrlgroups.The dcgreeor kind ofl crno-
psriods of stress. hey constitutcthe minor-
                   l                          tional disturbirncedid not distinguish the
ity. The accident problem is most likely      accidentfrom the enuretic child,
presentamong those individuals who regress       Pfi)sifldl {dclors j While mcdical examina-
to action-behavior at specific periods in     tions revealed no physical differences be-
their life.                                   tween groups, nrodeof physicalfunctioning
   We suggest that the problem of juvenile    did appear to be rclated to the incidence of
delirrqrrency may be anothet worthwhile       accidents,Accident chiltjren were deemed
area for the application of our               nrr)re activer before and after birth, and
prone" psychodynamic thcory since there       showcd earlier motor dsvelopment and good
appear to be similaritics in the modus        coordination. In addition, indications here
operantli of both accidcnt-prone and delin-   were that accidcnt children usc the moior
quent individuals. Chitdren with learning     sy$temas the primary channElfor exptession
problcnls and adequate intelligence may be    of anxiety. They would be expectedto rcac(
"action-prone" subjccts.Thus, educational     to ten$ienwith an increasc physicalactiv-
problems are another area for study in the    itv.
framework of thiS report.                        Intralanily Faqplt The relationship be-
                                              tween the accidentchild and his parcnts was
Surrlf'.tanY                                  differentin severalrcspects    from'that found
   This study, extending over a four-year in the other groups.There wcre fewer activi-
period, was designcd to irtvestigatefour ties involving the family ari il group in both
aspectsof the accident pattern in children: the accident and tbe enuretic subjects, but,
psychological, physical, intrafamily factors, whcrcasthe enureticchild's parents,particu-
and thc bchavioralresponse    involved. Chil- larly the fathers,were shown to be dominat-
dren fronr 6 to l0 years of age who had had ing, aggrcssive,     and controlling, the parents
at leastthree major accidentswere compared of the accidentchild were anxious.insecure,
w,ith a group of children manifesting ad- snd nonassertive,This diffcrence appeared
justment difficulties through enuresis,ilnd to be relatcd to thc parents'own fl6vEl6p-
with a symptom-fredgroup. Parentsof the mental history and subsequentadjustrnent
subiects  were also studied,Thc investisation difficulties. Their primary need was f,or ac-
   s conducted by psychiatrists and social ceptance,sometimesindiscriminatcly.The
326                                                      PSYCHOLOCICAL            APPROACHES

prevailing attitude in this group was ex-           as "maladjusted" and "irresponsible" which
pressed by one, "l would rathcr have a              have bcen attributed 1o those cxhibiting
million friends than one enemy."                    accident-prone behavior. Also the relation-
   Attitlenls us a Syrlptomatic Response:           ship bctween accidents and suicide were
accident pattern appears hcre to be related         reviewcd.Thesefindings wor.rld     not support
to ernotionalprohletnsbut not to a specific         the theory that accidcnt hehavior is "uncon*
diagnosticcategory.It would seem to be a            scious suicide" or hostility turncd irrward
rnanifestatiorr disturbance
                of             analogousto a       arrd gencratedby rcvolt against inhibiting
symptom $uchas enuresis. terms of learn-
                            ln                     authoritative or: punitive pirrents. We did
ing theory, accidentswould bc a rcsponseto         not find that the accidentchild is acting out
emotional disturhance, a stimulus which            self-dcstructivefantasies on the part of thc
undcr othet circumstances might evoke a            parent, Nor do the major causesappcar to
dilTerent response. The conditions under           bc linked to identillcrrtion with accident-
which this bchavior wor.rldoccur include a         prone parents.Closcness     and idcntifications
hyperactivity which may be constitutional, a       within the family appear to be wcaker in the
terrdency to exprcss tension through the           accident group,
physical activity, and disturbcd farnily              A psychodynamic theory is proposed,
relationships.                                     cncompassing the constitutional, psycho-
   Findings were comparcd with those of            logicrrl, and intrafanrily factors seerrhcrc to
other studies.In somc cascsthc intcrdiscip-        be related to accident repelitioll. Therrr-
linrrry approach I'rasextendcdthe knowledge        peutic as well as preventive mcittiures are
of what lics bcncath descriptivetcrms such         suggested,

    Empirically this study leavesmuch to be desired,but as a theoreticalanalysisit
offers a sctphisticatedtreatmentof accidentliability among chilclrcn.Il dealsquite
intcnsivelywith valious psychological     theoriesconcerningthe motivating forccs
behincl accidents, discardirrg suchopcrationally  circularformulationsof accidents   as
primarily "unconsciotts   suicicle" self-directed
                                   or              hostility.Quite directly,it placesa
maior responsibility childhoodaccidents
                      for                       upon the parcnts,and thc finding that
thc parentsol thc acciclcnt  group showedmore anxicty and insccurityundcrscores
the importanceof studyingchilcl-rearing    practices  and attitudesas important influ-
encesin the developmcnt thc accident-liable
                            of                     child.+ This {inding may, however,
havc resulted .from the accidentsand their consequences      rather:Ihan the reverse,a
possibilitywhich nrustalwaysbc considerecl retrospcctive
                                               in                work of this typc.
   The results this studyseem supportthc maladjustrnent
                of                 to                              hypothesis. unfortu-
nately,thc studyis weakened rnanynrethodological
                             by                        inadequacies: mpling,research
dcsign,controls,and reliability ancl validity of tests.Thus this stucly,like so many
other clinicallyolientedpsychological   investigations,           to
                                                         attcrnpts build a theorctical
analysis greatsophistication
         of                      upon an empirical  frameworkwhich is not completely
adcquateto sustainit. Although thc work affords some excellcnt         insightsinto the
possible dynamics psychological
                   of                factorsin the causation someaccidents, does
                                                              of                 it
not provide an adequatetcst ofthe hypothcses     $uggestcd. Suchpsychologicaltheories
needtranslationinto more operationallydefinablefactors which can be cvaluatedby
more scientillcally dcfcnsible  nrcthods.
* BackettandJohtrston(Chap.8)andReatleta/.tohavereportcdcontrolledstudiesofchildhood
pedestrian nccitletrtsthat should be consulted hy readcrs interested in the influence ofsocial factors
in rclation to acci(lents among children.
 PSYCHOLOCICAL APPROACHES                                                                327

 F A C T O R S I N M O T O R V E H I C ] L EA C ( . ] | D E N T S
     -John J. Conger,Ph.D., Herbert S. Ga.rkill,
                                                          M.D., Donaltl D. Gtart.ph,D.. Linda
       Hassell,Ph.D., Robert V. Rainev,ph.D., William L. Sawrey,.ph,D.

     Thc foltowing investigationis a sophisticatedexampleof the intensivestudy of
  psychologicalfactors in relation to traff]c accidents.The iruthors' awirfeness the   of
  necessity avoidingstatistical
            of                      biascs the selection sub_iccts in the collcction
                                            in              of          and
  of data from them is clearly indicatcdin this sturJyancl an earlier report.rr For
  example,   they statethat "assignmenl interviewers testcrs subrects done
                                            of             and          to          was
  rand(tmly'and in neithcrcasedid the examinerhaveany knowledge the suhiect's of
  accidefltstatus." In adriition, they emphasizc       the coniiderablciinportance.rt-the
  cross-validation other saml)les subieils the findingsof rcscaichof this type,
                    on                 of          of
  and they did so in their earlierwork.
      rhis study was basedon a follow-up study of I0 airmen with two or more acci-
 dents and l0 with no reportcri accidenis,all           whom had already receiveda wide
 varietyof ps,ychological   tcsts. Psychiatric intcrviewing well u, ..ruiin. psychological
 and functionill testing was also inc,luded.The primary ainr of this piojcct
                                                                                      wai to
 determinehow thesetwo groups diflcred on such variables underiyinghostility,
 aggression,   reality oricntation, convcntionality,adjustrnent,and fricnd,shii putt*rn*.
 In all, l3 evaluativc'atings-werc      made by two indcpendent *accident
 additional rating involved the use of a hypotheticaf compos"ite                    and nonl
 accidenttype and the irssignment eachof the subjects one of theservnes.
                                       of                       to
     Six out of l3 specific  variahles  discriminatcd   between hieh- and low-accident
 groups at the 0' t 0 levcl of confidsnccand four of lhescat the 0.0] levelor lower. The
accident   subjectwas more likely than the nonaccident       subjcctto displaysignificantly
poorer control of hostility, lowei:tensiontolcrance,        higher srparoiinn an*iety ani
dependency     needs,  and extremcs   ofboth egocentricity  oisociocentricity    and fantasy-
preoccupation unrellectivencss. general,the prcdictivesucqess ttre clinical
                  or                     In                                   of
evaluation    turnedout to be greatcrthan rhat of the psychiatric     interview.
     As in the caseof the studyby Marcusc,/a/., these      rcsultsare more suggestivc   than
they are definitive.The small number of cases       prec:ludcs adequatecomparisonof
thc pertinentbackgrourrd       characteristics the t*o groupj. In particular,we do not
know whether the two groups lverc qualitativelydiflerent in their actual driving
exposure    and in the completencss accuracy reportingtheir accidents.
                                       and            of                              Nor do
we kllow the extent to which thesepossiblccliffirences      *ay frauecorrelatcdwith their
Psychological    characteristics contributecl thc differential
                                  and               to                  accidenr  expepcnce
on which the work was based,        The rclativehornogeneity thc military population
from which the subjectswcrc derived, the influenceof ihe military rnilieu,'anclthe
factorsthat cJetermine    entry into suchservice  makeit dilficultto inteipretthe findings,
particularlyin relationto civiliangroups.This study,however,          givesan indicationof
          with which psychological     investigationsrlrust be pursui'd, sonrcof thc ways
]he 1119
irr which rigor must be achieved,    and thc potentialof approaches      with this emphasi..
Its authorswerecer(ainly     correctin statingthat thcir "positivelindings. . . contribute
 nateriirlly to a clcarer understandingof the relation of psychotogicalfactors to
 Lccident  susceptibility."
                                                            PSYCHOLOGICAL              APPROACHES

Ir xns r.or.rc BEEN    REcocNrzEDthat rllotor including structured psychiatric intcrviews,
vehicle accidclltsare lrot, as the tcrm im- objective and prdective personality
                                    "erring        attitude and interest tcsts, and a numbet ol'
plies, sirtply a function of what           men
call chance." It has becn demonstrated,       for psychophysical measures.
example, that $olne road$ arc safer than               As rcported previously, most of these
othcrs and that ccrtain times of the day are        measures   failed to discriminateconsistcntly
safer.  There is also strnrccvidenceto indicate with repeated samples between high-acci-
                                                                            and nonaccidentrate
that certain people are rnore likely to have dent, miclclle-accident,
accidents thirn others, and       a number of airmcn. For example, while a nrtmber of
attempt$ have bccn rnade to determine psychophysical                  tneasures,  such as discrirni-
whcther such pcrsons posscssunique per-             nttion rcaction time and depth pcrception'
 sonal characteristics which difTercntiate ancl a nurlber of objectivc
 thern from their fellowmen who do not have such as the Minncsota
 accidcnts.                                         ity inventory arrd the Thurstone teurpera-
                                                                              promising initially'
     For  example, in various studics the acci- ment sczrlt, rrppetred
 dent-repeaterand nonaccident drivers have they           failed to srrrvive in cross-validrrtion.
                                                                             however, provcd ex-
 been compared on a wide variety of psycho- One type of mea$ure'
                                                                                             of the
 physical and psychophysiologicalmeasure$' trcrnely stable, rratncly, meirsures
 on measures of personirlity structure and individual's           value sYstem.
 functioning, and orr social history variables'        Of thc six scales containedin a modifica-
                           haveyieldedeqnivocal tion ofthe Allport-Vernon-Lindzey          study of
  Many of thcsestudies
 or contradictory results, probably        partly values,developcd by Dr. Jerome Levy of the
                                                                            scales--those dealing
 because of rcal differences between the project staff, three
  populations studicd. but probably also in with          esthetic, theoretical,and religiouscon-
                                                                             discriminating, both
  many casesbecause methodologicaldefi' cerns-proved highly
  ciencies the designof the studies-includ- initially and in latcr cross-validation
  ing failure to conftol for variations in acci'     The rernaining ricales(ccotrornic. political,
                                                                            The accident subjects
  dent exposure,unreliability in the accident and social) did not.
  criteriir cmployed, iurd lack of objectivity in wcrc cotrsistently     less tlrientcd toward re-
  the rlcasurementcrf drivcr:characteristics. ligious v;rlues       and morc toward estheticand
     The   weight of the evidence hr'rwcver,
                                  ,           sug' theoretical values than the nonirccidentsub-'
  gests that in most populations accident jects. As might bc expected'thc mean
                                     from other of thc middle-accidentsrrbjects          fell in the
  repeater$ are likely to differ
  drivers in their pcrsonrrl  charactcristicsand, middle rarrgcbctween the high^accidentand
  further, thirt psycho\ogical variables ate nonaccident subjects.
   morc Iikely to prove discriminating than             On the basis of the scoreson this test, it
   measures physiological psychophysical was possibleto differentiate
               of              or
   functioning,                                      tween high-acciderrt and nonaccidcnt sub-
                                                                                 73 9{ of cases in
      The purpose of the present paper is to jects in approximately
   describcthc most reccnt findings     in a ftrur- rcpeatedsamplesttvcr severalyears. While
   year investigation the rolc of pcrsonaland this dcgree of pretlictive succcss
                       of                                                               wa$ Fllcour'
   interpersonnl  f'actors deterrnining
                           in             accident aging, thc authors felt that not cnough was
   frequcncy and sevcrity among alrlnen sta- yet known of thc personality strr'rctr'rre
   tioned   at Lowry Air Force Basein Denver' functioning of tlresesubjects interpretto
   In previous work on this project' 264 sub-         psychological  meaningof thesetest findings
   jects were administereda wide varicty of with conficlerrce.
    psychiatric and psychologicaI techniques,           The prirnary aim of the present study,

        peprinted, with permission,frorn TheJournal of the                         Medical Associationl
           -                                                              -American
        L                         / d 9 : t 5 8 l - 1 5 8 7 ,p r i l 4 ' 1 9 5 9 .
 PSYCHOLOGICAL AND PSYCHOPHYSIOLOGICAL FACTORS                                                                329

therefore, was to obtain a broader under.                                 The interviews were conducted bv two
standing of the characteristics thesetwo       of                      project mcmbers, a psychiatrist ancl a psy-
groups, The subjectswere 10 high-accidcnt                              chologist;the psychological  tcstingwas done
and I0 nonaccidentrate airrnen,sclcctedto                              by five clinical psychologists.Assignmcntof
be representative        ofthe ovcr-allaccidentand                     intcrviewers  and testers subjects
                                                                                                to         was done
n o n a c c i d c np o p u l a t i o n sE a c ho f t h e s e u b -
                   t                                       s           randorrrly, and in neithercaseclid the exam-
jects was studied intensively.High-accident                            iner have any knowlcdgc of the subjcct's
subjects were dcfined, as in the frrevious                             accident status.
  udy, as indivicluals           who lrad two or more                     Thus, the basic data employed in this
accidents      for which they wcre olicially held                     phase of the study consistedof (l) psychi.
responsiblein the past four and one-half                              atric intervicwtranscriptsand (2) test proto-
   ars,at lcastonc of whichoccurred
                                  within                              cols and individual psychologicalreports.
 he year irnrlediatelypreccdingtesting.Non-                           Thcsc data were then rated indcpenrlcntly
                                                                      for l3 variableshypothcsizcdto bc signifi-
                                                                      cantly related to accidcnt frequcncy and
   eithcr official recordsor thc individual's                         severity on the basis of prcvious work. In
  bjective report, during the past four and                           addition, two over-all personality-type   de-
one-hall-years. I)ata obtained fcll into two                          scriptionswere presented,    one hypothesized
generalcategories:   clinical studiesand ob-                          to represcnt conrposite
                                                                                    a            accidentindividual
    vc tcstlng.                                                       and the othcr a composite non*accident     in-
                                                                      dividual. Raterswere asked to rate subiects
                                                                      as more like one or the othcr of thesc two
  The major portion of the presentstudy                               personality types.
   olvcd an intensiveclinical evaluation of                              In the case of the interview, two judges
l0 high-accidcntand l0 nonaccident rate                               were employed-thc original intcrvicwer
    iects,Eachsubjectreceived structured
                            a                                         and the alternate interviewer.Similarly, in
      iatric interview aimed at obtaining                             the crrseof the psychologicalreports and
nf<rrnration  regardinghis socialhistory and                          originirltest protocols(which wcrc rated as a
      nality developnent as well as his                              composite unit), two judges wcre also em.
      nt attitudes, intcrcsts, hchavior, and                          ploycd-thc original tester and one ofl the
     rnality functionirrg.The interview was                           other testers. Prcliminilry analysisshowcda
 pe-recorded for futurc reference,                                    high degrccof i{gree  rnent betweenrilters.In
  In addition, cach subiect received a rou-                          view of this eviclenceof reliability, the
 nepsychological   examinationby one ofthe                           average   rating ofthcjudges was cnrploycdin
 inical psychologistson thc staff of thc                             all cascs, The l3 variables rated by the
 niversityof Colorado Medical School.The                             judges included the following critcria.
ollowing battery was administeredat one
    ng: thc thematic apperceptiontest (10                            Vnnreer-Es
  rds), thc Rorschach test, thc Wechslcr-                               l. Amount of Underlving Hostility: Rat-
   Ievuc intclligencc scalc (four subtcsts),                         ings of hostility (whcthcr cxprcsscdor not,
     the Sacks sentencecompletion test, .ln                          ancl regardlessofhow expressed)     were made
      casc, thc psychologist was asked to                            along a continuum from negligible to ex-
         a gcneral evaluation of thc subject,                        ce$sivc,  This variablc was includcd bccause
()ting such things as over-all assets and                            it has frequently been mentioned in thc
abilitics, basic character traits. and defcnse                       litcrature as distinguishing accidcnt from
     nismsemployed.In view of the inten.                             nonaccident subjects. It was the investi-
on ofthe invcstigators avoid as rluch as
                       to                                            gators' opinion, however,that this variable
   ible influencing the spccific contcnt of                          was not clirectlyrelatcd to accidentsuscep-
   reports no further instructions were                              tibility. Rather, it was felt that the impor-
ven.                                                                 tance of underlying hostiliry as a lactor in
330                                                      PSYCHOLOGICAL
                                                                              and respect for ncedsl
accident susceptihility could be evaluated satisfying own needs
only in relation to the indiviclual's     capircity and feelingsof others) to vcty socioccntric
                                                                            awareness inrerestitr'I
for control. This hypothcsisled to incltrsion ("over-deterrnilrcd"
of the next two variables.                           and respectfor rights and fcelings of otlters).I
                                                                             accidentsubjects   would
    2. Strength o.f l)estruttive, Hostile Im- It was prcdictcdthat                                     I
pulsesDirertttl Tov'ttd Environmenl in Re-           be more likcly to cluster at the cxtremesofl
Iation to Ahility to (loutrol            Impulsts: thc contirrur.nn, with nollacciderlt subjectsl
                                                                     around the nrid'poirrt.
Possible     ratings of hostile impr.rlscs en- tcilding to firll
                                            to                                                         I
vironmcnt rangedall the way fr:orn
                                         "hostility    6. (.'onvention(r ! : Convetrtionality rat- |
                                                                       I il
                                                                       on the extent to which the I
mocleratelyor much Ereaterthan control" to ings wcrc based
'***cs55ivel] greater control in relation to individual was convcntittnalin his consciotts             I
                                                                values.intcrcsts,and inrpulscstol
amount of I'rostility."It was predictedthat thinking,
                                                                 possihly distinguished from, for I
accidcttt subjectswould show reltrtivcly less actiorr, as
                                                                pndcrlying unconsciou$       nceds, or I
capacity for corrtrol in relation to strength example,
 of hostilitY.                                       overt behirvior in certaitr life situattons.
     3. Aggression   und Behavior:    Ratsr$ofag- Degree of conventionality or unconvell-
                                                                      rrol bc cotrfusedwith degree
 gression and behavior wcre askeclto judge tionality shoulcl
 which of thc following dcscriptions          most of over-all acljustrlent. which was riltcd
                                                                  Possible ratings ranged frorn
neatly chartcterizcd the sr.rbjcct:(a) Con- separatcly.
 sistently,   openly bclligerent;(6) r;suallycorr-   highly convcntional to highly unconverr-
                                                                     prcdicted tlrat accidcut sub-
 trolled but occasionaloutburstsof hclliger- tional, It was
                                                                                   IESS conventional
 ence;   (c) covertly hostile (passive-aggressive' jects woukl bc rated as
 critical,cirrping,indirectlyaggrcssive     toward   than nonaccident suhjscts.
 others); reillistic,flexiblc" nppropriately as-        7. Over-all Atliustment: The degree of
                                                                                      to live happily
 scrtivc; (e) passive" uuits$ertive (fear of over-all adjusttnctrt-irbility
 retaliation for expression of aggression);          and function clTectively-was ratcd by satis-
                                                                                        needs in rela-
 (/) unasscrtive, low drive, low energy; factory resolution of subject's
 (5;) tlepressed;and (fi) other" to be briefly       tion to efrvironrnental and superego de-
                                                               This should be separrrtcfrom con-
 describcd, It wrrs prcdicted that;rccident marrd$.
                                                                                               of this
 subjccts    would be morc likcly to fall in cnte- ventioflhlity ot unconventionality
 goriesa through cr nonaccident        subjccts into  adjustmcnt. Possiblc ratings rtrrrgeclliotn
                                                                             to exccllent adiustrrrent
 catcgoricsr/ throrrgh g (only one suhjcct fcll definitcly disturbecl
                                                                               accident subjccts ls a
  in cntegory fi, about which there wiLs no It was predictcd that
 prcdiction).                                         group would tend to havc a somewhatpoorer
 -                                                                                      stlbjects.
     4. Reality Otientation: Possibleratings of adjustrncnt than nonaccident
  reality orientation (realistic appraisal   of own     8. Castrationln.xritrry"Possibleratings of
                                                                                           anxiety, or
  motives ancl envirorrmentrrlftlrces ir'fl'ccting specific evidcnce of crrstrttiou
  subject-extent to which behavior is directed gcneralized feitr           of injury to self, ranged
                                                                                  a scparate category
  toward meeting subject's own enlightcncd frorn high to low, with
                                                                "Not clearly observablc in these
  self'-intercst) ranged frorn very poor (definite entitled
  breakdown, at least occasionallyin reality- data." This variablc wfls includccl
  testing) to excellent.   It was predicted that because it appeareclprominent in many of
                                                                                           lt was prc-
   nonacciclentsubjectsas a group would tend the psychological test rcports.
   to show rlot:c adcquatereality      orientation'   dicted, thorrgh withorrt too much confirlence
      5. Egocentric Vtr'tus Sotitttjefllrr:(:.' Rat- that nonrrccident subjects would score
   ings rangcd from vcry egocentric (very sclf- higher on catitl:ation ilnxiety than acciclen
                                                                                             that high
   ccntered, Iittle interest in rights of others or subjeots, the rationale being
                                                                                            the suhjec
   respect for their feelings) through a ncutral castration anxiety rrright make
                                                                                              willing to
   mid-point (neither nrarkedlyegocentricnor more consciousof risk and less
   socioccntric,   fnirly good balance betwcen engage in risk-taking bchavior.
 PSYCHOLOGICAL AND PSYCHOPHYSIOLOGICAL FACTORS                                                        331

      9. Separation Anxiety: Possihle ratings            friends; (r) some close friends, many casual
  f<rr separationanxiety (fear of loss of lovc friends; (r/) no close fi'iends, rnany casual
 and support, conccrrl with oral supplics) fricnds. It *ns prcdicted that norraccident
 ranged lion1 high to low, with a separate subjectswould be more likely to lall into
 catcgoryrntitled "Not clearly observable           in   categoriesD or c, with accident subjects
 thesc data," As irr the case of castration falling primarily in categories or r/.            a
 anxicty, this variable was included largely
 becausc appcarcdprorninentlyirr many of
              it                                         Corrrposrre h.lprvl nuals
 the psychological tcst rcports, lt was pre,                In addition to thc above variables, raters
 dictcd, again witholrt too r:ruchconfidcnce. were presented with the follorvins
 that accidentsuhjects        would $corehigher on       pcrsona.liry-rype   dcscriptions, intcnd-cd to
 separation anxiety than nonaccitlent sub- clrarLrctct'izc              hypothetical composite non-
jects, the rationale being that underlying accidcnt (type x)
                                                                             and accitlent (typc y) sub-
fcclings of lac:k rrf lovc lead to rcscntment jeEts,and wcrc asked to state which
                                                                                                      of the
which might tend to be actcd out against two typcs the subjectmost nearly
figurcs perceived as depriving.                             T)tpe l': Thc typc x sort ofperspn tcndsto
     I0. E.ttent q/'Iantasy Preot,t,upaIion:     pos-   be rathcr convcntional in his values, atti-
sible ratings ranged lrorrr excessive          preoc- tudes,and behavior.ln general,hc seems               to
cupation with lantasy satislactions          through gain satisfaction ltrr his nccds from his
a mid-point of "effectiveintcgrrrtionof fan- everydaylife irnd to
                                                                               havea fairly clearnotion
tasy and action, plarrfulness"to extrEmgly of the goals he want$
"stimulus-bound"                                                                 t<: attirin in life anrl of
                          (likely to irct directly in   how to pursue thenr. He tends gcncrally to
response to people or objects with little havc rcspcctfor the rights
                                                                                        and opinions of
fantasy or reflection).It was predictedthat other peoplc, and in turn
                                                                                       to be prerty well
accidentsubjcctswould tcnd to clustcr at the            acccptedby orhers. While he ttay be nor_
extrenre$,      wlrile nrtnaccident  SubiectS  would mally assertive defencting own rights,
                                                                          in              his
    rd tti fall around the mid-noint.                   hc is scldorn extremely clcnrancling,        domi_
    I I. Charatter Type: Ratcrs wcrc askcd to           nee'ring,or cornbativc, In gq1,na        cases,hc
 udge which of a nurlbcr of charactcr types may actuallytresonre                what lcssasscrtive     thirn
    rst ncarly describcd the subicct (if morc           an adequrrte   defense his own rights would
 lritD one catcgory was uscd, probablc rcquire.Such underlying
                                                                                     hostility as he may
          ty was indicated by using rturnbcrs have tendsto bc pretty well controllttl
                                                                                                     and is
 , 2, etc.; see table 2). While it was not felt seldomexpressed irn obvious,.acting-ou1"
   at this variahlc was likely to prove dis- fashion. lf psychoparlrology
                                                                                            is presint in
    minating, it was includeclbecause the this typc, it is rlost likely to take thc
                                              of                                                     forrns
   tention it has rcceivedin the literature.           of denial, overly strong ncedsfor conform_
    12. Ahility to Tolrratt Various Ttnsions: ity, castration anxiety and fetrr of
       ible ratings of tolerance to tensions from others, and ovcrly strong needs
  .c. withor.rt   discharging   them immediately) plcasc or placate others.
            frorl high to low, with a seprrrare           Type Y.: Whilc tbcre may be a fair amount
   tcgory entitled "Not clcarly obscrvable of individual variltion on s;rccific
    thcse dnta." lt was prcclicted that high- teristics, in general the type y pL'r$on
    :ide  nt subjects would reveal a lower to be rathcr unconventionalin his
         ity for tension tolerance.                    and villues and sornetimes in his ovcrt
    1J. Friendship Puttern,t: Finally, raters behavior. He is inclincd to be
                                                                                              rather self-
        askedto judge rvhich the f'ollowing oriented,to have dilticulty in relatingeirsily
          rics, in thcir opinion, ntost nearly and warmly to other peop(e, antl
                                                                                                    to be
        ibcd the stittus of the subject'ssocial somewhat unawarc of or insensitive
  Iationships: no r:lose        friends,few casual their rights and feelings.He miry in sorne
             (D) some close friends, fcw casual casesfeel that othcr pcoplc
                                                                                         are difficult to
332                                                     PSYCHOLOGICAL APPROACHES

                                                                      or ClLINIcnt Psv<:sot-ocIcaL
understand. He often has feelings of dis'          TABLE l.-R^TINGS
                                                   Rtponrs ron AccInlNr-RFtTATER ANI) NoNAC;(f-
satisfacti()n about his everyday life and is                        DENT CRouPs
likely not to have clearly defined goirls in                                   DIREcrloN        [.evel of
life or consistcnt, practicftl methods for                                #---_l
attaining such goals as he does have' He is
                                                                           Arci-     arci'      (il lx'low
apt to have difficrrlty in controlling his             Variable,   No,     denl      drnt          0'10)
hostility, which may otten be excessive,   and      2 Hostilitv/control   Less      More
 consequently may at times be inclirred to-            ratio              contYol   tontrol        0'02
                          "acting-out" behirv-      3 Aggression and      Aggres-   Adaptive'
 ward over-cletermincd                                 behavior           Sive      una$$ert-
 ior, either at the level of ovcrt physical                                         ive            0.05

 belligercnce or at the level of verbal ag'          5 Egoccntric-bal'
                                                         anccd -socio-
 gression-in the form of sarcasrr, carping               scntric        Extremes Balanced      0.01
 iomplaints, or destructive   comnlents'IIe is       9 Scpirration
                                                                        High     Lq*           0.01
 likely  at times to exercisepoor judgment in            anxicty
                                                    l0 Fantasy-tntegra-
  his evaluation of his envirtttrnretrtor in             tion-stimulus-           Integ-
 choosing courses t)f action' In some cases'             bound          Extrernesration         0.ro
                                                    t 2 TensiontoleranceLow       High          0.01i
  he may appear overly preoccupicd with his         l 4 Gcncral
  own fanttrsy world and rclativcly cut-off              pctsonality    TYPeY TYPcX             0.10
  from socitrl relationships with others' If       * From psychiatricinterview; flll others from psy-
  psychopathol(rgycxists in this type of indi-     chological reports.
  viduat, it appears rnost likely to take the
  forrns of impulsive, hostilcly-toned,  acting- respectfor the rights and fcclings ofothers);
                                                 (3) to be eithcr cxccssivelyprcoccupied with
  out behavior,poor reality testing,emotional                                          "stirnulus-
  lability, impaircd intellectual functioning, a fantasy satisf{ctiol'lsor extfelllely
                                                 bound";   (4) to be more fcarful of lossof love
  highly personalized. idiosyncratic fantasy,
   withdrawal from interpersonal relation-       and support (and, by inference,more angry
   ships, or Excessive immaturitY.               and rcsentful towirrd pers()n$ viewed as
                                                 depriving); and (5) to be generally less ablc
 Rnsults                                         to tolerate tension without discharging it
    As may be seen    from table l, 6 of the 13 immediately.ln addition, the accidentsub-
                                                 jects tended to be categorizcdtrrorc frcquent-
 variables proved discrinrinating bctween
 high-accident and nonaccident       subjects at ly as consistently or occasionally bclligerent
                                                                hostile. rrnd less freclucntly as
 thc 0,10 levcl of conlidence or better and or covcrtly
 4 proved significant below the 0'02 level        only rrppropriately assertive,or urras$ertive
 (i.e. lcss than 2 chances in I00 that differ- Whilc not statistically     signilicant,thcre were
                                                                        accidctrt subjects to be
 ences this largc or larger could have oc' tendencics for the
                                                                                   and generally
 curred by chance). the I I variableshypo- rated as more conventional
                                                                                   no differenccs
 thesized     as possibly discriminating, 6 better adjusted. Thcrc were
                                                                            rronaccident subjects
  showed significant differences in the pre' between accidctrt and
  dicted dircction,  either in the psychological in ratings ol reality orienttrtion, castration
  test battery, the psychiatric interview, or anxiety, or friendshiPPatterns.
  both.                                             Of the two variahles hypothesized not to
     In comparison   with nonaccident subjects, be discriminating,    neithershowedsignilican
  thc accident subjects showed a stati$tically differerrccs.    Neither hostitity alone, withoul
                                                                                       nor clinical
  significant tendency (1) to have lesscapacity referenceto capacity for control,
  for  managingor controlling hostility; (2) to character type showcd any clcar-cut
                         sclf-centeredand indif-  The distributiortol accident   trndnonacciden
  be either excessivcly
                                                                                  character types
  ferent to the rights of others or exccssively subjects arnllng the variotts
                                          "over- is shown rn table 2. Only subjcctson whom
  sociocentric (i.e. characterizedby an
                                                                        lcst raters indepcndently
   detcrmined" awart:ltt:ssof, interest in, and both psychological
PSYCHOLOGICAL AND PSYCHOPHYSIOLOCICAL FACTORS                                                   333

 agreed are included in this comparison. type descriptions resulted in correct dis-
 Agrccmeltt wirs shown in ll cascsand dis- crimination of both accident and nonacci-
 agreenlentin 2 cases"As rnay bc sccn, both           dent suhjects in 70o7iof the ctrses.Despite
 accident and nonaccidcnt subjectsare dis- the small number of subjects,                such a result
 tributed acrossa ltrrgenurnbcrof catcgories, would be expectedby chirnceless than l0
 and there is little tendencyfor cithcr accidcnt times in I00. Thus, whilc a number of rhe
 or nonacciclcrrt      subjects to concentrate in specific variables rated had somewhat
any particular catcgoryor categories.          While greater predictive value, in gerreralit would
the number of sub.jects this study is too appear that rhe$e:
                                 in                                        pcrsonality-typedescrip-
small to refute thc hypothesis that qertain tions provide a fairly adequate             picrureof the
clinical character type$ are more frequent hypothetical composite accident and non-
among accidentthan nonaccidentsubiccts. accidentindividuals.
it seemsclear that no singlecategorycould
account for a majority of either accident or OuEcrrvE Tesrluc
nonaccident suhjects,                                    In addition to the clinical studies de-
     It will be noted that the predictivesuccess scribedabove,all subjects         u,creadmini$tered
of thc clinicaI testing wns considerably a number of objcctivc mcasurcsof interest,
Breatcr than that of the psychiatric inter- intclligcnce,psychophysical              skills, flnd psy^
views, While this fact mriy well reflect real chophysiological functioning. (Sevcrul of
differencesin the predictive potential of the thesc measurcs, such as thc Wcchsler-
two measures, should be pointed out that Bcllcvueintelligence
                    it                                                      scale,were actuallyad.
electronic recording difficulties resulted in ministercdas part olthe clinicirlstudies,            and
incompletcness a nnmber o[ the interview analyzed clinically at that time, but the
transcripts,and this may well have lowered results of formal scoring will be considered
the predictive valuc of this measure.                here simply for expositionalconveniencc,)
    It should also bc obscrvcd thrt the present Each of thesc mcasures*,ill bc discussed
        ngs have not been cross-validated on briefly, togetherwith the resultsobtained.
    her sanrplcsof subjects,This should cer-            Thc mod ifictl Allport-Vernon.Lindzcy stu-
 ainly be done beforc thc presentresults can dy of valuesmeasure,             which had rcpeatedly
      acceptedwithout reservation,ft mav be proved discriminatingin prcvious work on
 hat one nr more of thesc findings will              this project, was again adrninistered.      With
         ually prove to he the result of chance, use r)f ths same formula (based on a combi-
   oweverrlt appearsextremelyunlikely that flation of religious, theoreticol, and esthetic
             idatiorrwill lhil to riupportthc main scalescores) in previoussamples,
                                                                   4s                         70il of
          of the findings, since it is extremely subjects in the present group were correctly
      ikcty that signiflcant differenceswould be identified as accident or nonaccident type.
        ined in 6 out of I I casesby charrce. This figure is significant at the 0.05 level of
   pccially with relatively small sample sizes. confidence and compares with discrimina-
    The ratings of thc two ovcr-all pcrsonality- tions of 731/oin thrce previous samples.
                                                        The Kudcr preferencerecord is basicallya
  aarr 2.-DrsrnrElJrrdN of (]'AKACTER Tyrt Rer-      vocationirl interesttcst which yietrlsnreirsures
       Avonc Ac(rrrrt r.REpErT[R AND NoruAcclnExr
                                                     of the extent of the individual's interest
                                                     along I0 dimensions(outdoor, mechanical,
                  AINI                         AN
   Hysteroid            I
                                                     computational,    scientilic,
                                                                                 persuasive. stic,
                   0       S.Neurasthenic      0 0
   Phobic          l 0     9.Paranoid          0 0   literary-musical,socialservice,   and clerical),
   Obscssive 0 2 l0.l'sychopflrhic              I 0  The Kuder tcst was includedin the prcsent
   Contpulsive I 0 l{.Passivraggressive 2      2
   Cycloid         0 0 l?.Passivcdepcnderrt I   I
                                                     battery bectruseprcvious results with the
   Schiz"oid       I I lS.Aggrcssivcreafiionl 0      study of values suggested    that other interest
   Anxicty state 0 0 14.lntellectdisordcr I o        measure$might alsc prove discriminating.
   ;l : accidcnt repeater; ,i\t - nonaccident.       As it turned out, most of the scales     did not
334                                                          PSYCHOLOCICAL APPROACHES

prove discrirninating. However, a formula            TanrE 3.-IrurEr.LrcENcE TEsT ScoREs oF AccrbENT-
                                                           RF.FEATf, R,tttn Nount:croEnr Cnours
based on tlre corlpr.rtationirl    and scientific
scales (on which accidcntsubjccts      were high)                                                                  MEANS

and the persuitsive and clerical scales (on                                                                     A+          N+
                                                     l   W e c l r s l c r d u l t i n t e l l i g e n cs c a l c
                                                                          a                             e
which nonaccident     subjects werehigh), madc              I ' r o r d t e di n t e l l i g e n cq u o t i e n t
                                                                                                   c               105.3 106.
it possibleto identify 8l fl of subjectscor-             Weightedsubtcstscorcs
                                                            Inft)rmatiolr                                           10.6 10.2
rectly as accident or nonrrccidcnt types.                  Corllrrelrcnsion                                         11.9 'l2.0
While this figure is signilicant at the 0.02                Sinrilari(ics                                           10.6 I 1.4
level.it cilnnot be acceptecl   confidentlyuntil            BlockdL:sign                                            ll.4 ll.9
                                                     2 . S h i p l e y -Ila r l t b r d v o c a b u h r y s c o r e 2 7 . 8 2 5 . 9
it has been cross-validated.
                                                     + I - ilccitlent-repeater; -- nonaccitlent.  N
   The Arnerican transit rnotor abilities test,
dcvclopcd by Krafl and his associates           in   tained on any of tlrcsc measures, Thus, it
their studiesof hus drivcrs, is intcnded as a        would appcar that thc differences between
illeasure ol. clriving aptitude. The subject is      accident and nonaccident groups which wcre
serrtecl the wheel of u mock-up bus and, in
        at                                           found   on other nteilsurcs could not be at-
re$porlseto visual light patterns <rna board         trihuted to differences in thc over-all intrl-
across the room, is required to carty out            lectual level of the two groups.
various maneuvers with brake, steering                  In previous phasesol lhis rescarch,sub-
whccl, gcarshift, and clutch. All responses          jectswererrdministered Minnesotarrrulti-
are rccortlrd electtically, ln this situation,       phirsic personality inventory (MMPt) and a
accidcnt subjccts reactctl nrorc quickly and         srrcialhistory questionnaire. Wltile neithcrof
took lesstimc to conrplete task but made
                              the                    thcsc mcasuresas a whole proved hclpful,
nore etrors than nonirccidcntsub.iccts.        All   item analysis yieldcd ir numher of itcrns
mea$Lrres   w6re significhnt irt thc 0.05 level of   which in then-rselves appcrrred promisingbut
confidenccor better.                                 which rccluircd cross-vrrliclationhcfore they
   All sr.rhjects were examined for psycho-          could be confidently accepted. I.or this
physiologicrl responsiverress,       under both      reason, these items were rcpcatcd with thc
restirrg and stress conclitions (i.e., before,       prescntsampleof subjects.    Thesc itcms and
during, and aftcr prcscntationof an incom-           their levcl of signiiicancein the I955 and
plctc scntcnccstcst containing both emo-              1956samples, well as in the current study,
tionally charged and neutral topics). The            are shown in table 4. As rrravbc scen.ver
scoresobtainecl inclrrdcd rncasur€sof rcspi-         fcw of these items continue to prove dis
ration, muscletcnsiorr,psychogalvanic         skin   criminating with repeatedtesting. Thc over
reflex,heart rate.and tr comhincd lncasurc      of   whelmingly ncgative character of thesefind
over-all reactivity. While wide individual           ings, particularly ctn the MMPI, appea
dilTcrences  occurredamong subjects their irr        especially irrteresting, sincc thcsc findi
psychoplrysiological      rcsponsiveness   on all    conflict with those of other iDVestigators
rneasures,  none of the nreasures     proved dis-    somcwhat similar satnplesof subjccts.
criminating bctween accident and non-                   The Rorschachand the thcmatic a
acciclerrtsulrjccts. although thcrc was solne        ception tcst (TAT), which hacl bccn indi
tendcncy for thc accidcnt group to have a            vidually adrlinisteredto eachsubjectdur
slightly fir$tcr hcrrrt ratc during presentation     thc clinical phase of this study and in
ol the test.                                         pretcd clinically at that tirne, were la
   All subjectswere administered       two meas-     $cored objectivcly for a nufirber of di
ures of intelligcncc, ir portion of the              entiatine charactcristics derived from
Wechsler-Bcllevr,re     adult intclligence scale     vious samples.  The accident and nonaccide
(forrr subtcsts),and the Shipley-Hartford            groups were compi{red across scores f
vocabularyscalc,As tnay bc seenfrorrrtable           thesecharacteristics. both the R
J, no significirnt diflbrcnces between the           and TAT scores,the dilTcrcnces    bctween t
accidentand nonaccidentsubicctswcrc ob-              accidcnt and rronac'ciclent  gr()upswcre
                            OLOGICAL FAT,'ToRS                                                                                                               335

Trrsrr 4.-SuvMARy oF SrGNrFIclxt DlrrrnEFtcrs                                         project demonstrated that the ficcident-
B n T w r r w A ( : C ( D Lr - R L r F ^ T E f A N D N r l r u l r . t - r u r u r
                           N                   t
                                                                                       repeater and nonaccidcnt airmen differed
  Cnoupri (rruMMI'l+ lr',to Socr,ql Hlsrony Irrlrs
                                                                                       signilicantly lrom one another, and while
                     Irem                        Hish I9J5 1956 1957
MMPI*                                                                                 certairr isolatcd psychological mcasures
 Co te .1tur"n aln)ost every                                                          provcd highly predictivc of accident status
  wcck                                            Nt     0.05 0.20 ns+                tirr this poprrlation,the psychological                               signi,
 Nevu did irnythingtlnngerous
  just lbr the thrill of ir                        N     O.20 0.20 ns                 ficance of these measures was diflicult to
 Frer.;uently notice hand shak-                                                       interpret. It was felt that furthcr rcscarch
  ing when trying to do some-
  thing                                            N     Q.?0 0.20 0.20
                                                                                      was ncccssary            bcforc a clcalcr undcrstanding
Uscd alcohol excessively                           A     0.10 0.20 ns                 of thc psychodynamic                     relationships           cxisting
Ncvcr besn in troublc with                                                            betwcenthe structureand functioningofthe
  the litw                                           N   0.20 0.05 ns
 Whrn with a womi.n a man                                                             individu'.rl and his accident susceptibility
   n s u i r l l yt h i n k s a U o u tt t r i n g s                                  could be achieved.
   reli{tedto her scx                                A   0.01 0,20 ns
 Pray severallirnesa wc'ek
                                                                                         The prirnary aim of the present study,
                                                     N   0.20 0.01 ns
SDcjal istOry
          h                                                                           therehrrc,was to obtain u bnradcr knowl.
 Singtc vcrsus tnlrried                             N     ns 0.10 0.20               edge tli the chitrirctcristics of acciderrrand
 No. o[ brothcrs                                    N    0.01 0.l0 0.05              nonaccident suhjects.Twenty representative
 Llrink to felax                                    A         0.?0 0.10
 .|-i(ther ilttends cfiurch                          N   0.10 0.01 ns                high-accident            and nonaccidentairmen were
  I)esires to run away ftom                                                          s t u d i t : di n t e n s i v e l yr,. r s i n ig d i v i d u i r c l i n i c a l
                                                                                                                                       n                l
    hone                        A   0.01 0.20 0.10                                   psychological and psychiatrir:techniques,as
  Havc cvcr bccn arrcst+d (in-
   dcprndent ol' trallic viola-                                                      well as a number of measuresof interest,
   tidns)                       A   O.0t ns   ns                                     intelligcnce,          psychophysical               skills, and psy-
" MMPI : Minnesata Multiphasic Personality
                                                                                     chophysi         ologictrI functiorring.
f  N : Nonaccidrnt; A - Accirlent repeater.                                              It is felt that the positive findings of the
f Not signilicant below 0.20 level.                                                  prcsent -ctudy,Fal'ticutarty in the clinical
sufficiently l:rronouncedto meet the 0.10                                            area, contrihrute materially to a clearer
    I of statistical significance required,                                          understanding of the rellrtion of psyc:ho-
                                                                                     logical factors to accidentsusceptibility,                                   at
       MENT                                                                          Ieastin the populationof airmen from which
   While previous reffiarch on the present                                           thesc $amrrleswere drawn,

   -Han.s Eysen&,Ph.D.
    Sincethe ecologyof physicaland chemicaldangersis highly varied, rangingfrom
he threatsposedby avalanches earthquakes thoseposedby calbon monoxide
                                      and            to
nd arsenic, there are many ways in which unexpectedinjuries can be initiated.
      usc of this variety, it is not rcasonable expectto find many isolatedbehavioral
   othcr factrrrs which individually account for the initiation of large fractions of
   idents.It is reasonable,     howcver,to expect to fincl ntany individual .ret.r factors
 hich frcquently do lead to accidents,       despitetheir rclative unimportanccin the
     -all picture.
   A possible psychologicalfactor of this type, describedby Eysenck,is thc "in-
  luntary rest pauseduring which the personin questionceases pay attentionto
    task in hand." Although not yct unquestionatrly          incriminatedas a factor in
   idcnts,it servcsas an additional illustlalion of the variety of psychological      ap-
           to accident causation    irndresearch. is of considerable
                                                 It                            in
                                                                      interest, addition,
336                                                    PSYCHOLOG]CAL APPROACHE

as a factor which "can . . . be linked with personality           (seeDrew et al.
belowfor a related illustration).

                                                  which we tend to firrd almost inexplicable
 It ls ptlssrBr-E. . to make predictions from
                ,                                 The cxperiencedtrain driver who fails to see
well-known principlesin cxperimentalpsy-          the signal set at danger; the driver who steer
chology fwhich] relate to factors which           his motor-car straight olt thc road without
may increaseor decrease accidcnt rate in
                            the                   noticing the curve; thc radar opcrator who
general, but can also be linked with per-         ovcrlooks the vital signal on his scrccn-
sonality differences,Take as an example the       thesc are all opcrating under conditions of
fact that on the cirrly motorways constructed    monotonous stimulation lcading to in-
in the United States and elsewhere,      many    voluntary rest pauses during which no
accidcnts occurred because drivers drove         ordinary signal has a chance to reach their
 into trecsand bridgeson the straight,m()no-     consciousness.Ccneral fatigue, as wefl as
toflous roads. This could easily l'rrrvebeen     the consumption ofdepressant drugs such as
predictcdin tcrms of Hull's law of reactive      alcohol, incrcasethc number and the length
inhibition, according to which any kind of       of theseinvoluntary rest pauscs,    and conse
mental activity scts up a ftrtigue-like neural   quently the chances of fatal accidents.It is
state ("renctive inlribition") which becomcs     sometimes said that decrement in perform-
increasingly  more severeuntil it enforcesan     ancc produccd by alcohol may be quitc
involuntary rest pause during which the          slight, and that consequently there is no
person in question ceases pay attcntion to
                            to                   good rea$onfor banning pcoplefrom driving
the task in hand. These rest pauses last for     just bccausethey lravc bccn drinking. This
only a vsry short period of time, during         argument disregardsthe essentialfeature in
which inhibition dissipates; hc thcn pays        the situation, to wit, the lact that thc decre.
attention for a while until anothcr involun-     ment in preformance produced by a little
tary rest pausc comcs along, Thc cxistence of    alcohol is not a continuous one, but mcrelv
such rest pauseshas bcen demonstrated ex-        momentary. If a dangcroussituatitln ari
perimentirlly, and it lrirs even been possible   at a tirnc when no involuntary rest plruse
to lind a physiological component by re-         occurring, then the drinker nray not be
cording the brain waves on an electro-           severcly handicapped, lf, however, a dan-
enccphalograph.                                  gerous situation ariseswhile he is mor
   Now it is a gcneral principle in psychology   tarily unclergoing involunlary reritpause
that the condition be$t $uitcd for thc produc-   then an accidcnt is almost inevitable.
tion of reactive rnhibition and involuntary         We can take this arBument one
rest pausesis one of massedpractice, that is,    farther. lt is now known that extraver
repetition of idcntical stimulus-response        people, that is, people who are
sequenccs.  This is prcciselywhat happens on     convivial,fond of drinking and smoking
thc straight, undeviating motorways which        quite generally fond of cornpany, tend
precedcdour more carefully planned modern        devclop inhibition more quickly and mo
roads, and the monotonous stimulation ex-        strongly than do introverted peo
periencedhy thc drivcr on thcse old roads        wlro prcfer to keep therlselves to thc
almost inevitably produced the results that      selves,who tend to hc introspective, to
have been found.                                 fer reading to going out, and so forth.
   This notion of involuntary rest pauses        would make an extravert rnore liable to
accounts for a greaf nurnber of accidents        dents arising from the occurrenceof invol

                with permission,from Medicine, Science theZaw (London),3:l:aI6-
     J-Reprinted,                                                                 |
     L               423, | 962.Portionsof the text havebcenomitted,            J
 THE PERSONALITY OF DRIVERS AND PEDESTRIANS                                             337
tary restpausesduring his driving, and indeed   the extravert tends to start with a higher
thereis someevidence show that this is so.      levelofinhibitions, and tcnds to increase his
Elderly people also, and brain-damaged          level even more by drinking, whereas the
oneg,are nrore likely than those young and      introvert, starting at ir low level, tends to
healthy ones to su{Terfrom numerous and         drink rather less.Thus extraverts are doubly
lengthy involuntary rest pausesand, there-                                             of
                                                at risk and only a clear understanding the
fore, constitute a definite hazard on the       position is likely to discouragethem from
road,                                           driving while under thc influencc of drink,
  Unfortunatcly the effects of extraversion     even to a relatively slight cxtent.
and drink are cumulativc in the sense that                    *       *       +

    Another approachto the study ofpsychologicalfactorsfocusesupon ths decision-
making processes individuals siluations
                    of            in              involvingrisk-taking.
                                                                      The perception  of
dangerand the judgmcnt regardingthe margin of safetyrequiredbeforeonc will act
in a hazardous   situationappearto be basicfactorsin nn individual'ssafetybehavior,
as Gibson has emphasized.     Experimental   study of suchdecision-making    processesin
real-lifesituationspromisesto advanccour understanding thc conditionsunder
which individuals   expo$e  themselves, their children,and othersto danger.
    In this field the work by Cohenand his colleagues outstanding.
                                                           is            Thc lirst selec-
tion,         irnd Hazard," presents   the resultsof a study of the risk and hazard
incurred in t{riving a bus through gapsof varying sizes;the second,"Thc Risk Takcn
in crossing a Road," studies    pedestrian behaviorin regardto crossing road in the
lirceof traffic.fBy "risk-taking" the author$meanengaging a taskwithout certainry
of success. degrceof risk-takingvarieswith thc rlegree suchcertainty. "The
             The                                                 of
conceptof                 the other hand . , . is usedto refernot to what the indiviclual
 ftrnksabout the situation,but to his actualsuccess failure in performance.
                                                        or                        Therc
 ,re, of course,degrees hazardjust as thcrc are degrees risk." (SceFox, below.)
                         of                                   of

  -lohn Cohen,
             Ph.D.,E. J. Dearnaley,
                                 M.Sc.,C. E. Il{. Hansel.

    In the first study, the experimentaltcchniqueconsistedin asking the driver how
  any tinres he thought he could drive his bus betweena pair of pcles withcut
  uching them. He was then permitted to drive the bus betweenthe poles as a
  easureof his actual perfornrancc.   The resultsof this cxperimentshow that both the
   imates and the actual performanceof the expcrienceddriver$ were better than
hoseof the inexpericnced     drivers. Both groups tended to undcrestimate their capa-
ities whcn they thought the task was relativelyeasyand to overestimatcthem when
he task was hard. As a result of training, drivers showedir t€ndencyto take lessrisk
     to be involved in less hazard. The authors discussthc implicationsof thcse
    ings for thc general problem of road safety. They believc that their presentap-
      h permits the scientificstudy of psychologicalfastors under controlled condi-

 A portion of anotherdescription this study is also discussed Chap. 3.
                                                             PSYCHOLOGICAL             APPROACHES
338                                                                                                         I
                                                                              gap he would atrcmpt tol
Or.trllt FACE tr, the problem to be studied srnallest size of a
                                                                               other words, what is thel
in this papcr might seem to be oll a rather drive through? In
specific character. It      arosc, howcver, in the smallest clearancehe would allow bcyondl
                                                                             car? We do not imply thatl
colrsideration a widc variety of questittns thc width of the
that fall within the range of what we            have motorists in generirl know the exact size ofl
                                                                                                 that suchl
callcd sub.icctivc    probabitityand risk-taking' rheir vchiclcs.Indeed wc suspect
             jrlportant to recognizcthat when, precise            knowlcdge is unusual' Neverthclcss
 It seerns
                                                                       to dccide when not to drivcl
 fr)r example, ;r trran i:i driving a car or whcn they have
 a pedestrian    is crossinga toad his behLrviour through a gaP.                                             I
 has somethingirr common with behaviourin                  The task of driving through a gap in thel
                                                        trrrffic may prescnt varying dcgrces of difi-l
 a clivcrsity of other contiitions which are
                                                        cr.rlty to any lnotorist who is going to at-l
 charrrcterized an clcurcnt of uncertainty'
                    by                                                               "degree of difficulty"l
 We hrrve elsewhere tried to show              that all tempt it. We can dcfine
                                                                   of the proportion of times he thinksI
  forms of jurJgernent decisionmade when in terms
                                                   just he will succeedat any size of gap or clear-l
  the information availahleis incornplete,
                                                                The extreme dcgrce of di{ficulty would I
  as all kinds of perfornrance undertaken hy a ance.
  oersonwhen       he is nttt absolutely   ccrtaiu that be one at which hc thinks it would be quite
                                                                      to drive through, for instance, if
  ie will succccd,fall within the scope of the inrnossible
  study of subjective probability            and risk- hc bclieves the gap is less than the width of
                                                                          is thus always some subjectivL
  taking. Or"rr prcvious strtdies embrace the car. There
  changes,during the period of ttrental             dc- probability ofsuccessor failure which repre-
   velopmcnt,in.iudgements          made in statesof scnts the easeor difficulty of the task'
                                                                                    "risk" to denote any
   uncerttrinty. Thcsc statc$ of mind arise, for           We use the word
   example, in our unclerstanding           of the way- attemFt at perfonnance when thc person is
   distrilutions arc built rrp, the nature tlf not surc he will succccdor when his
   statrsticalindependc:nce      arrtl the various pro- jectivc probability is ltss than 1.0. lf we have
            involvcd in the knowledge gained i:r.          measurcof a perstttr's$tlb.iective     probabili-
   from    a sarnplcand in thc intcrpretation of ty o[ succcssor failute wc catl use it as a
   words wilh a cluantifiable      meaning'Wc have measure of risk-taking. An attempt which
                                                                                                   may not
   nlso investigated conliclencc itr decisions at $on$titutcs a risk lcrr orlc person
   different    ages, thc subjective element in signify a risk for anothcr person wlroscsub-
    gambling, and various fortns of risk-taking jectivc probnhility fcrr the attempt is 1.0.
    with and without danger.                                Thc idea of risk is thus basedon an indi-
                                                         vidual'sfeelingof his likely succcss lililur:c
    Atu on EltQulnY                                      Thc conceptof"hazard," on the othcr hand,
                                                                   ttr rcfcr not to whl{t the individual
       Consider i1situation which arisesfrequent- we use
                                                                               sitrration, but to his actlal
    ty in the    life of a motorist' There is an tlurrks about the
                                                          success failurc in perforrnance.There are,
    opening in thc traffic in t'ront of him' He                                             just as tllere are
    rnustclecicle  whctherthis gap is largeenough of course, degreesof haztrrd
                                                          degrecs of risk. ln our extmple, a motoris(
    for him to ctrivehis car through ot not' The
                                                          who thought he could always drive through
    simplest case would irrise whcn other ve-
    hiclcsin thc    traffic are stEttionary,  pcrhapsin a gap of 5 ft. plus hall an inch successf'u
                                                          woulcl not bc taking a risk if he attenptcd t0
    a queuc at traiiic signals.$spposehis car is
     5 ft. wide; woutd he drive      through an aFen' drive through. But if wc tested his driving
                                                          skill and found that hc firiledin a proportion
     ing of 5 ft. wide plus half, an inch, thus
                                                          of his attcmpts, the situation would holc
     ;rllowinga clearance only half an inch? If
     he would not attcmpt        this task, what is thc some hazard for hirn' The proportion of hit

                      with permission, oor;i:,rr;ilu.",saarch
  failures in actual performance provides a         moved apart a little at a time. At each dis_
  lncasure lrazirrdin thesecircuntstances.
            of                                      tance the driver was asked lo say how rlany
    Jn the study of pedestrianbehaviour we         tinres <rut of five he thought he could eei
 fttternpted measure
             to         levelsof risk_takine in    through without touching thc posts. in
 cvcryday situations when failurc mieht be         instructor moved thc posts apaft 6 in. at a
 rather iierious. the presentpapcr oui chief
                 Irr                               tintc, This was reduccdto 3 in. for the more
 aim is to examinc differences theserisk-
                                 in                expenencedgroup. The posts werc rnoved
 taking levcls attributablc to trainins in the     alternatelyso that the bus was irlwaysfacing
                                                   the centrc of thc gap. At cach distancethe
                                                   expenmenter asked the drivcr how manv
  ExruttulrurAt pnocr.ouRe
                                                   times out of five he though( he could eci
     Our experirnentis based on two groups of      through. This was continueduntil the driier
  trainee bus drivcrs, One gror,rp   consistedof   said he could succeedfive tirnes out of five,
  bcginners in their first week of itrstruction.      The next stepwas to deterrlincthe driver's
  and the other, though stjll under training, maximum risk-takirrg
                                                                            levej. Thc posts wefe
  had all passcd the Ministry of Transport replacedsevenfcct
                                                                         apart. lt was exnlained
  Public Servicc Vehicle Licence test in the that they would
                                                                      be moved aparr as heforc
  week beforc the experiment. l lre averase and that as soon
                                                                       as he was sure he coulcl
  duration of the truiningperiodwassix weeks. manaBeit
                                                              hc was to drive betweentherr and
     Thc experiment was carried out at the circle round
                                                                back to thc starting point. .I-he
 ManchesterCorporation Transport Depart- driver was told that
                                                                           now he rrrustactuallv
 ment'sTraining Schoolfor Bus Drivers.The drive between
                                                                    the posts but that on no
 task was to drive an 8-torrdouhle rieckerbus account must
                                                                  hc touch them. The exoeri_
 between two wottden posts 6 ft. high. Thcse menter and
                                                                an instructor movcd thc nosts
 postswere 3 in. squarein scctionand naintcd
                                                  apirrt as befq16un1i1   the driver signallcdhis
 with black and whitc bands.The difficultv of intention to
                                                                make an attempt. The distance
 the task could be varicd by altcring the iis-    api{rtwas notedand the driver drove throueh
tancc bctween the posts. The task was in- and back to the
                                                                      sr.-rrting poinr. The tjrir;er
cluded in the training coursecarried out in was allowed as
                                                                   much time as he wishecl.
the presenceof an instructor,                        Finally wc measuredthe driver's actual
    l'wo lines, l2 ft, long, were rnarkedon the performance
                                                                at differe4t distani:es between
ground in the training schoolyard, One line the posts.
                                                             Because limitation of time this
was a starting line. l-he other, parallcl to it could only
                                                              bc done for thrcc distanccs,      The
and 12 ft, away. was marked in 6 in. inter- expcnlrentcr chosc
                                                                        distanccs the basisof
vals, the posts being placed on this line. what he had
                                                               secnofthe driver'spcrformance
These marks w,eresuch that they coul<Jnot in the previous
                                                                     test. The three distances
be seen from the driver's seat. I'he exncri- chosenwcrc
                                                               suchas to givc a low, a hieh and
rnL-nt bcgan with an 8 ft. wide bus rrarked an intermediate       levelofsuccess.    lfin the case
with its fronr to rhe starring linc facing the ofany driver
                                                               thesewerenot realized made he
)entre of the gap.                               another set of attempts. Thc posts wcrc
   We first estimated thc driver's subjective placcd at the first dis1ap6.and the
                                                                                        driver tokl
:rrobability oI success differing degreesof
                          at                     to make five attempt$at that distance.This
Jifficulty.The driver was instructedto sit in was rcpcatcdfrlr the two other distances.
ris seat.Hc was told that he was to take part he wished, the driver was allowcd
                                                                                           a slrort
n a tcst which consistcdin clriving between rest between each set of
                                                                              fivc attcmrrts. For
;hetwo postsbut that first wc wishedto find eachattemptthe experimenter            and instructor
)ut how well he thorrght he could clo it.        cach stood by orre of the ptists in orrler to
   The cxperimenter     pointed out thst thc gap move thcm quickly asicle theyif       werc hit so
petween postswas too narrow tirr the bus as to avoid excessivc
          the                                                               damage. The time
o piilssthrough but that thcy woultl bc taken for the experiment varied
                                                                                         from 15
minutes for the more experiencecldrivers to case          that people  generally prefer saythatl
an hour for sotne of the beginncrs,the aver- they think they can either do or flot do thel
age time bcing half an hour. The wlrolc pro'        task? Wc did not encountet any dilliculticsl
gramme was complcted during Juuc and in the drivcrs in expressing                 their judgementsl
July, 1955.                                         on this scale.ln previous studieswith chil-l
    The length of time required by some of drcn ancl adults we have hardly ever fourrdl
the beginnerswas due to their Iack of profi- anyone who felr that such a scalc maclehisl
ciency in clutch control and consequent re- judgement artificial. We do flot thcrelorel
peated stalling of thc engine, This held up feel that irn artelhct is being imposed on the I
the cxperimentseriouslybecause put the drivers. The consistctrcy thcir estimates
                                      to                                         of
full lock on after passing betweenthc posts, bears this out.                                          I
it was necessary use half clutch. Some of
                    to                                 Second, is it possible that significantlyl
the beginnerswerc also not very proficient in different results might have been obtainedl
steering. An instructor provided as$istance had we employcd scalcsranging liom 0 tol
in bringing the bus back to the starting line 10,0 to 100,or () to 1000?          Possiblya subject
when neceririary    but did not hclp during an who          says he could succeed   live times out of]
 attempt. The experiment had to he aban- tive might sayl orl a different scale, that he
 doned in the caseof thrcc beginnersbecause would succeed95 or 99 tifies out of a hun-
 of their lack of proficiency.                       dred, In so far as the sub.lccthrrsany doubts
     Three sets ol obscrvationswere taken for        at all about his capacity, this mighl be mani-
 each subject at distancesbetween the       posts fested on a scalewith a large enough range.
 varying from T to l2 ft. The distanceswcrc On the other hand, a suh-icct spite of anyin
 in intervalsof 6 in. for the beginners    and 3 Iurking rnisgivingsmight still say he could
 in. for thc more advanccd trainees, the succeed 100 times out of 100 or 1000 otrt of
 intervals increasing frorn smaller to larger 1000. .Thc question is wlrcther thcre arc
 distances,                                          upper and lower thresholds on the scale of
     (i) At each particular distancebctween the subjective probability. Above the upper
 po$ts, the dtiver was irskedto say how many threshold the subject would be conrpletely
 timcs he thought he could get through with- certain of successand below the lower
  out touching them.                                 threshold hc would be equally certain of
     (ii) (rd Thc smallest distance between the failure; in the intermediate       zonehe would be
 posts which thc dtiver attemptcd to drive uncertain.             Thcrc might also be diffetential
  through: (D) His successor failure in this thresholds.
  attempt.                                              Third, there is the question whether a
     (iii) The proportion of succcsses driving linear rrrathematicaltrirnsforrnation of one
  through the posts as these were placed at $calcinto anothcr is psychologically
  varying distancesaPart.                             A subject who says lour out ofl five when
      Observations were made of the intervals asked to estinrateon a scale of 0 to 100
  as they incrcased in size from srnaller to might say 80 out of 100 or any value between
  Iarger distancesin the caseof (i) and (ii) but ?0 and 90 or he might say 99 out of 100. In
   not.alwirys in the case of (iii). The observa- the last case.this would be his best rneans        of
   tions were made in the order (i), (ii), (iii), so conveying an element of slight uncertainty.
   that any sxpcrience   gainedin (iii) would not Thesc questions of the equivalencc of psy-
   affcct (i) and (ii).                               chological and mrrtheilraticalscalcscan only
      A word ofelucidation is neededabout the be answcrcd by experimeutal study, which
   scale.Wc have useda scaleof 0 to 5 and thc we are now undertaking.
   subject is limited to a choice of onc of six
   value$on this scale.Severalquestions,      how.
   ever, flust be faced. First, is it natural for        (i\ Estimatesord P"rfrry@!-g!r!!
    people to expresstheir   juclgemcntsin terms MoreEI pifienr"d D,i t'",s" In Tables.l. and
   of thc type of scale we have used?Or is it the ttffitw'           f"r the two gt<lups of drivers
 RISK AND HAZARD                                                                                                                                     341

TADLEI.-MEAN Esflulrrs or Succssglwn Acrurr                                                It is evident from Figures I and 2 that the
P t n r o n u r r u t r o E l r u r x p t H r r t c E oD R t v r n s( n - l 5 )         relationshjp between estirnates and per-
                         wrDTH0F BLrs tsEET       I
                                                                                        formance remains much the same for troth
Distante                     Esli-           Sucttssful Errot  Number
                            Huted            perfum-     of
                                                                                        grouFs of drivcrs. The essential dilltrence
betwpen                                                       who wttuld
 POSIS                  Jl/aa(i.fJ?J          ilI((.f          CSli-        Qtt(fiPt    betweenthe tu'o groups consistsin the fact
Feet in.                  out of 5           out of 5          mate          laslr      that the more expelicnccd dri,',ersattain f'ull
           i            i     i          i     i         i      i       y      y
                                                                                        Euccess srnallcr diiitances, as indeed we
  7            0             0.0               0.0             0.0              I
  7            6             0.0               0.0             0.0              2       should expect, Thc c{l'ect of the training
  E            0             l.o               0.0           +r.0               4       shows itsclf in superior perl'ormancc   rather
   I           6             ?.0               1.0           -1.1.0             8
  9            0             3.0               2.2           +0.8              r0       than in more realisticjudgements of their
  9            6             4.2               3.s           +0.7              r4       own skill, This is brought out in column r'y
 l0            0             4.5               4.8           -0.3              15       of Tnhles t and It, Thc error of estimnteof
 lo            6             4.6               5,0           -0.4              t5
 ll            o             4.1               5.0           -0.3              15       the trained driversis, ifanything, largerthan
 II            6             4.9               5.0           --O.l             r5       that of the untrained. We should of cour$e
 12                          5.0               5.0             0.0             15
                                                                                        bear in rnind that at this stage, thc "cx-
                                                                                        perienced"drivers have not cornpleted    their
TrsLr IL-M ret Esrtverss or SuccrssAno Acru,lL
  PERF(IRMANcE MrtRs F.xptnnN(En flRtvEns
                                   (n - 15)                                             AyERASE
                              wtDTH ttP eu:i I rcr:T                                     SUCCES5
Distance                  Esti-       SutfttsJal              6ryr7,       Number       OUt OF j
betq,|'ta                 mated       perform-                  of        nhr, y,puld
 pdsllt                 Juf(i.,.f.fes   an(es                 esli-        atttnpl
Feet in,                 out ot 5      ont of 5               mate           la.rk                      totE
       i            i         i      i        i      i         i        v      v                      hJF rrll.
  7            0            0.0               0.0              0.0               0
  7            3            0.0               0.0               0.o            o
  7            6            0.3               0.0            .10.3             0
  7            9            1.?               0.0            . +t . 2          I                                               avtrd'a   l'@lt!$
  8            0            2.0               0.0            1.2.0             3
  8            3            2.5               l.l            +l.l              7        .
  I            6            3.4               3.0            +0.4             t0
  8            9            3.?               4.4            -0.7             13
  9            0            4.1               4.9            -0.6             14                                              (
                                                                                                      DISTANCE ETWEEH HE POSTS fI.)
                                                                                                             B      T
  9            r            4.4               5.0            -0.6             t4
  9            6            4.5               5.0            -0.5             15
  9            9            4.7               5.0            -0.3             15            FIo. l. Mean $timate of successand
 10            0            5.0               5.0               0.0           15            actual performance of inexperienced
                                                                                                      d r i v e r s( n = 1 5 ) .
                   ively, the averagenumber of times
hey think they would succeedout of five
 tefipts, their actual pcrformance in five                                              AYEIAGE
          and the nurnbcr who attempt to                                                out oF 5                         ;r'
  ve through at each distance between the

 We see from columns ii and iil that both
  estimatesand the actual perfnrrnanceof
he experienced drivers are greater than
    of thc incxperienced.
                        From colurnn iy
  can see the relationship between esti-
                   and the performance.
                                      Both groups
      to underestimate   their capaciticswhen
    task is relatively easyand to overestimate                                                       D I S T A N C E E T W E E N H E P O S r S( f i . )
                                                                                                                   E           T
   ir capacitieswhen the task is relatively
  rd.                                                                                       Flc. 2. 14*ttr estimates of successand
  Columns ii and iii of Tahles I and II                                                     actual performancc of more experienced
     :tively are plotted in FiguresI and 2.                                                            drivers{n = l5).
34?                                                          FSYCHOLOCICAL AFPROACHES

full training and that they are not yet re' enquiry we could simply observe the
garded as qualified drivers by the Man-                 that ;r pcrson allowed himself to cross the
chesterCorporation, although the Ministry road bcfore the arrival of an oncorning
of Transport regardsthetn as fully qualified. vehiclc, We could not, in thosc circum-
     ln both groups wc find tht'rtat the smallcst st$nccs, obtain a measure of lris subjective
gap when they are certain they will succccd, probability of successful performance
i.e. they do not feel tlrcy are taking any risk, asking hirn in rrdvance            whether he thought
they do in t-actinvariably succeed       with a still he wtrslikely to atrivc intact at the other side
 smaller     gap. Thc relevirntpairs of figurcs are of the roacl. lu the prescnt study we have
 l0 ft. and I ft. 6 in' for the experienoedand direct rrrcasure           of tnaxirnumrisk-takinglcvel
                                                                                         gap between the
 l2 ft. and l0 ft. 6 in. for the incxperierroed. ifl lerrns of the stnallest
                                                                                                 Thc valucs
 These arc avcrage valhcs' Tlrere are, of posts which a driver will attempt.
                                                         for thc individual drivers,incxpcriencecl         and
 cour$e' variations from driver to driver. ln
 fact eight out ofthe     inexpcrienced  and five of expericncccl, Ehownin TablesIII and lV.
 the expericnced      drivers thought thcy would Thc average ntaxinrLrntrisk-taking level tbr
                                                                                          I ft. 9 in. At this
 always succeed at a distance srnaller than the inexperienseddrivcrs is
 the one at      which they did, in fact, always value thcy cstimate in advance that thcy
 perforfi succcssfullY.                                  coulclsuccced 54 per cent of the attctupts;
                                                                                                 pcr cent of
      Frorn the point of view of safe driving the in actual fact they succeedin 40
  ideal    place for the irrtcrsectionhetween the tl'te atte$pts. The corrcsponding ligures for
  two curve$should be on thc Iine               That the experienceddrivers alc ll ft. 6 in., 60 per
  would mean that at the first magnitudc of cent and 50 per ccnt. Although thc
  the task hc would be preparedto undertake, between the vLilueslor the two
   his juclgcnrents would be realiritic and he not statisticallysignificantthey do indicate
   would be fully aware of his linritations. that the more experiencedclrivers take Iess
   Thcrcfore at any lcvel of dilliculty that he risk although their clear'tncc               is smallerthan
   would irttempt, he would bc cautious           and that of thc inexperiencetl.This is shown by
   undcrestimate his capacities. Wc find that the fact that thcir success                itr pcrfomrance is
   the point of intersection the two curvesis 50 per cent as cotnparecl with the inex-
                    -'b," i.e. the maximum level of perienced drivers. We should add that the
   closerto line
   risk taking, in the case of the trained as variation in traximum r:isk-taking level is
   comparedwith the untraineddrivers.ldeally significantly greater for the incxperienced
                                                  "b"     than fbr the exPeriencsddrivers'
   too, for maxitrrutn    safetyin driving, line
   should coincidewith a performancclevel of                 Three of thc inexperienccdand six of thc
                                 "b" is closer to tlris
    100 pcr cent. In fact line
   level in tlrc caseof thc experienccd      drivers'
   Thc contlititlns of the expcritnentdid not O U T O F 5
   involve the common hazards of the toad and
    we should tlrerefore not cxpect ntaximutn
    ri$k-takinglevelsto coincidc with a 100 per                            lzth
    cent level of performancc,The estirnates       and                  inpo$ibtc

    pcrformrrnce of the experienced driving
    instructor that we tested cxen"Iplifics      these
    conditions.(SeeFigure 3.)                                                          rirk t+king levcl
        (ii\ Maximum Ri.sft-Ifl(ing   Leve/s: In our
                              "maximum risk-taking
    used the expression
    level" in the sense of a valuc of strbjectivc
     probtbility of risk which a Fersonwill rarely,           FIG, 3. An experienced       dtiving ifl$truc'
                                               of                                 tor.
     if ever, take. In the circurrrstances that

Trnr.E III.*Slrrrr-Esr  Glp Brrwrr:n posrs wslcH                  TlsLt IV.-SFrnr-Lrsr. Cap BrTwurin lrosrs wHr(]H
INtXFEtrtLNclb L)nrvrx WouLu Arr.rMpr, Hrs lig11-                                      D
                                                                  E x p E R T E N ( : EI ) r r v t n W o u L u A T I F : M p t , H t s E $ T I -
    MATTD Su(:(:t$s lruo AcTUer. penronulruce                           u a r r o S u t : t r _ s sA N b A ( : T r J A l ,P l n r o n v a n t . t
           Snqlltslgup         Estimated    ptrfi)rmdn.:e                        Snntllestgap                               pcrlirnrznte
 Driwr      w,hit h tlriwr     put(ntdga    pr:rrr,nt.g.           Driver        xhith th'irer          pcr(ilttag(
           would a1lt'mp1                                                                                                     Ftr((ilt(rge
                                 JUc{:CJJ     Jtlcae.rs                         would dilemil             Jlladej,t            JracPJJ
                11.     in.                                                         ft'       in'
    lf          9        0        t00            BO                    lr           g         3              I r)0                40
   2       i      g      0         40             0                   2            8          9              100                 I00
   3           l0        0        t00           100                   3        '     9        3              t00
   4       ^      ?      0
                                    0             0                   4        +      8        6             r00                  t0
    5      r      g      0          0             0                   5        +      8        6                0
   6       '      9      0
                                  100             0                   6        +      E        3              40                  2Q
   7           9         6                                            '   l      r    E       0
                                   40           r00                                                          100                   0
   8           9         6         40           t00                   8+            E         0              100
   9       '      9      6
                                   60           r00                   9  r            E       3               60                  60
  l0            9        6         80            20                  1   0            9        6                0                100
  ll*           8        5         60             0                  ll.            I         9               40                 I00
  lz*           E        6         40             o         l         2             8         6               80
  ll*           9
                         0         60             0                  Il+            7         g                 0
  1    4       +    8      6                                                                                                       0
                                   40            2D                l4               8         6               80                 100
  15            9        6         40            80                15               I         9
                                                                                                                0                100
                E        9         54            40               Means             E         6               60                  50
* Thesedrivers ftriledat their                                   + These drivers failed at their maximum
                               maximum risk_taking                                                       risk-taking
levol.                                                           lcvcl.

expcrienced drivers took no risk. Of these                           The determination ofa person'smaximum
drivcrs who took no risk only one in each
                                                                  risk-taking level does lrot giy6 us comrrlete
group was not involveqiin any lrazard,that
                                                                  inforrlation abotrt his behaviour in a given
is, hc was alwiryssuccessful his maxilrurrr
                                at                                task and it rhereft:re needs to be su-pple-
risk-taking level.We must rcgard him as the                       mented by an additional measure. For
bcst clriver in his group frorn the noint of                     example,two drivers may first declarethat
view of road safety.                                             they will always succeedwhen the gap is
    Four of the inexperiencedand six of the                      increased 9 ft. One ofthem is preparedto
   perienccdwere not involved in auy hazartl.
                                                                 drive through a gap of tlris rragnitudc, but
 )nly onc in cachgroup took no risk, that is.
                                                                 the othm will only unclertake gap of g ft.
    uld not undcrtake the task at his r-naxi_                    3 in. They havc thc same maximum risk-
   um risk-trikinglevel unlesshe thouelrt he                     taking level, but their conduct is dilTerent.
   ruld always be succcssful.                                    Sinrilarly, two drivers rnay both say that they
   Ninc of the inexperienccd    and four of the                  will neversucceed a gap of I ft. 6 in. This
 xlrcricnceddrivers both took a risk trnd                        is the smallestgap which one of rhcrrrwill
        involved in a hazard,                                    attempt to rlrive through, whereasthe other
   (iii) Margin.r of' St1lcty and Hazard: The
                                                                will attempt a gap of u fl. J in.
   rxfinum flsK-taklnglevel i5 a measureof a
         's                                                         At thc two extreme$ of the scale. two
            subjective probability of successat                 persons may thus have identical levcls of
he rnagnitudeof a task which he will first
                                                                maximurn risk-taking but they w,ill not
  tempt. He tells us, in othcr words. the                       necessarily in the sarnecircunrstanccs.
   rccntageof times he thinks he will succeed                       In calculatingthe safety margin of a per_
 t the rnagnitucle thc task he will first
                      of                                        fornrancc wc subtract the smallestdistance
       rtake. This rnerrsure independentof
                             is                                 at wh jch the fly'.iys1 sure he can succee4five
  c nature of thc task and of thc phvsical                      times out of five from the nragnitude of the
    ts cnrploycd,so that dilTercnt   tasks may                  task in the distancein any given performance.
    conrpared in rcspcct of irn individual's                    This measurewill bt: in fcet and inchcs but
    -taking levels.Dili'erentindividuals
                                           may                  would have the advantage ol being con_
      be corlparcd with respectto the sameor                    tinuous. There would be only one safety
     rcnt tasks.                                                margin for every perfbrmance and only one
344                                                           PSYCHOLC}GICAL APPROACHES

performance would correspond to any                     TAbLE vI.-CLEARANcE, S.lrrrv Mnx.cIns lun Hrz-
                                                        ARLI MARcINs AT THF MAxtMUM RIsK-TAKiNG LtrvELS
iafety margin, The measure would be posi-                        ron Mont IlxFERtlNcru DRlvERs
tive when no risk is being taken, that is,                Drlver    Cleatantc     Mdr|in oJ'    Margin of
when the person thinks he could succeed     at                                     x'lrtYt      hazttrd*

att five attempts, and negative whctr some                             ft,            ft.           It.
                                                            I          0.25           0.75
risk is involved, that is wlren he thinks he                2          0.75           Q.Z5          0.00
may not succeed all five attempts'
                   in                                       3          0.?5           0.50        -0.75
                                                                                      0.00        -0.7 5
    ** .*n apply a similar rnea$ure hazard'                 4          0.50
                                                                                    - 0.50        -0.50
                                                            5          0.50
 This may be calculatedin the same way as                   6          0.25         - t.00        -0.25
 safety margin except that we subtract the                  7          0.00           0.25        -0.50
                                                                        0.00          0.00        - 1.00
 smallest rlistance at which a person could                 I
                                                                                    * 0.25        -0.25
                                                            9           0.2s
 actually succccdcvery tilnc out of fivc fiom              l0           L50         -0.50           0.50
                                                                                    - 1,25
 the distance in the performance in question'              ll           0.75
                                                           l2           0.50                        0.00
 It is the differencc betwecn what he c'ould               13         --0.25        -0.50         -0.50
 <lemonstrablydo and what hc actually does'                14           0.50        -0-21           0.00
                                                                        o.75        - 1.00          0.00
     Margins of safety and hazard calculated               l5
                                                                                    -0.25         -0.30
                                                          Means         o.43
 in this way tre shown at the maximum risk-             t Thc grcatcr the negative value, the larger the risk
 taking level,in TablesV and VI' The means              of hazard.
 of the two groups of drivcrs differ signifi'
 cantly, although no significant differcnce             ficance of his actions for him' The alternative
  appear$ between the nleans for clearance'             would merely be to t'ecord obscrvations of
  The vrrlue of f for the difference in safety          drivers' behaviour which rnay includc psy"
  marginsis 2.S2(tl'f. - 28)' p<0'01; and for           chologically heterogeneous elements' Indi-
  the difference in hazard margin t is 2'96             viduat driver$ nrust be studied separately
  (d.f. : 28),p<0.01.                                   otherwise the specificlllctors, such as aware-
                                                         nessol'risk, which influcrrce  thenr will eithcr
 REsut ts, Couclustoxs ,luo            Dtscusstox        not appcar at all     or be obscurcd in the
                                                                             from a group' lt is the
    The essentialleature in this type of investi- avcrages dcrivcd
 gation is that it enablesus        to take account of elucidation of this kind of
                                                                  which will help us to understaud
 the state of mind of a driver and of the signi- factor
                                                         suclclenor dramatic ch.rnges in tt driver's
                                                         bchaviour as may happen if he gets into a
 T,lnte V,-CI-EII{ANcE,    S,lrtrv MAR(]INS rNs Hlz-
 AnD Mlnclus nr
                    .rHr MnllvtlM Rtsx-lArtNc Levers     state of panic. It is not, for example, the
              F0R IhIxrERlENt]ED DRlvERs                 objective spcedof vehicles,by itself, wlrich is
    Drivet     Clearatce       Margin of     Murgin  of  the crucial element but the speed at which
                                t"f.".       haza,rd+    thc driver feels uneasy and at which his
                   ft.                                    performance begins to dcteriorate within a
        I          1.0            0.0         - 0.5
       2          0.0           - 1.0         * 1.5       siven constellation of circumstances. This
        3          2.0             0.5           0.0      speed varies from driver to driver,
        4       -1.0            - 1.0         *2.0
                                -1,0          -2.0           C)ur rnain specific conclusions are ag
        ]          0.0
        6          0.0             0.o                    follrrws:
        7          1.5          -2.s             0,0         (i) On the average,the more experienced
        I  ,        1.5         -2.5             0,0
        9          0.5          * 1.0            0.0      drivers had highcr minirlum margirrs of
      l0            1.5          -0.5          - L0       safcty and hazard. That is, they took less
                                 - 1.5         - 1.5
      II           0.5
                                 * 1.0         -1,0       risk and were involved in less hazard. As a
       l2          0.j
      l3        1.0              -0.5          - 1.5      group they variecllcssth&n the inexperienced
                0.5         - 1,0         - 1.0
                 r.5        - t.5
      l f
                             -o.97        - 1.00            (ii) The two groups did not differ with
    Means        0.73
                                                          respectto the size ofclearance, and the pro-
                           value, the targer the risk
  . Thc Efeilterthe neEative
                                                          portion f'ailing in perforfiance at their maxi
  or hazarcl.
    RISK AND      HAZARD                                                                          345

    mum risk-taking levels respectivelywrr the                Finally we should like to draw attention
    sanrein both groups,                                   to two implicationsof this investigation, one
       (iii) On the average both groups would             for a psychological approach antl thc other
    attempt a task when they wcre far from cer-           for the general problem of road safety. The
    tain they r,vould   succeed.                          special contribution which psychological
       (iv) Both groups tcnded to underestimate           sturlycan makc is one whiclrtakcsan entircly
    their capacities   when they thought the task         diffcrcnt point of departure from that taken
    was relativelyeasyand to overestimate      their      by the engineeror physiologist,In ellect it
   capacitieswhen they thought the task was               meansthat the observations the first nlace
   relatively hard.                                       are made liom thc subject'so*n pcrspective.
       (v) The dilTerences  befweenthe two groups         This is flece$sarybecause his behaviour is
   was flot large but we must bear in mind that           partly rr any ratc detcrminedby the way he
   what we havecalled the "expericnced" group             perceives and undcrstands the situation.
   had orrly had about six wceks'n'toretraining           Thc perspcctive will of course vary from
   than the inexpericnced     group.       trends in      subjcct to subject,What seemseasy or safe
   the differenccs arc all such as wc should             to one may secrn hard or dangcrous to
   expect to bc the outcome of training, These            another. The prcscnt study shows that a
   trends are;                                           purely psychological   approach s4n 6s.rm-
       (a) a tendency to take less risk,                 bined wirh a propcr experinrentatdesign
      (6) a tendency to be involved in lesu              and a suitable statistical evaluation of
   hazard.                                               results.
      (c) a tendency for variation in the group              Behaviourat zebracrossings    nlay serveto
   to diminish,                                          illustrate further certain irspcctsofa psycho-
      (r/) a tendency for estirnatesof perform-          logical approach,
   ance t(r becorne more realistic at the maxi-              Pedestrians Chingford complaincdthat
   mum risk-taking level and also for the                they slipped on zebracrossings.Investigators
   maximum risk-taking level to approach the            from the Road Researchlaboratory reported
   point at which they think thcy will always                  "the stripes
                                                        that                are no m('rreslippery than
   succeedand will, in fact, always succeed.            the rest ofthe road" and concludedthat the
      Thcsetrcnds arc borne out by tlre estimates       pedestrians slipped f<rr psychological rea-
   and performance of the drivcr-instructor.            sons: the stripes /ooAcrl slippery. The situ-
      (vi) The superiority of the exper.icnced          ation at a zebra crossjng when a pedestrian
  group showsitselfin iurprovement skill in
                                         of             is contemplating crossing the road and a
  steering and mtrnipulating thc bus rather             motorist is approachingis one the outcome
  than in better judgcrnent of their driving            of which dcpcndson mutual gucsses each  of
  capacities. In fact a proportion of both              other's intentiorrs rather than on physical
  group$ attcmpted the physically irnpossible           factors or legal rights. Zebra crossings can
  task of driving thc bus through a gap                 only becomesaf'eto the cxtent that elements
  narr()wer than its width. The experienced            of sub.jectiveuncertainty are rcduced,
  driver-instrucror,however, showed his su-                 As lar as thc incredibly complcx problem
I Fcriority chiefly in being able to judgc             of road safety is concerned,we take the view
laccuratelywherrit was and was not possible            that an understancling it can only conrc
  to drivc tbrough a given gap. Thus when              by meansof patientexperimcntal        study, The
ldriving a bus I ft. wide )recould not succced         pfescnt enquiry emphasizes,amrrng other
         gap of I ft. 3 in. and did not arrerllprit
lat a                                                  things, thc importance of the learning pro-
lu'hereashe could always succecdat I ft.               cess.Rcradbehaviour differs from most other
16 in. and did attempt it. It would seernad-           ordinary activitiesin that failure rnay mcan
lvisablc to include training in making judgc-          disastcr. l-carning, thcrcfore, takes place
lments of this charactcrin a widc Iariety of           without full knowledge of results. Qnly
               during the period of rraining.          rarely can wc lcarn from our f'ailures, for
                                                         PSYCHOLOGICAL          APPROACHES

example, from failure in overtaking' lt              A priori paflaceasin terms of reduced speed
follows thercf'orc that training to drive            linrits, wider roads, cars in better state of
should inclu<lesituations such as this which         repair, or holdirrgaloft ideasol'courtesyare
yield knowledge of results in safe circum-           not likely by themselves solve theseprob-
stanccs.   The emphasisin lcaruing to drive a        1errrs,Nor is the proposal that a Royal Corl-
car is 4t present on learning to nranipulate         rnission on Road Safety be appointed to
                                                     estahtishthe facts a solution since thc kind
the vehiclc. What is equally irrrportant is
                                                     of facts thrrt need to be-determined rcquire
learning to iudge the degree of hazard in a
                                          judging    a long term research programrne into the
 situation. This implies training in
              speed,distance,and so on' We           behaviourol'roird usersand could hardly be
                                                     establishedby a lrody such as a Royal Cotn-
 wonder how many motorist$, having mas'
 tercd tlreit l-{igtrway Code, will follow its       mission.
 advice to turn into thc dircction of the skid          We arc inclincd to suggestthat one of the
 when involved in onc thc lirst and perhaps          first steps to be taken in an expcrimental
                                                     progriimme should be a frank enqrrit:y
 the last time. Thc vast majority woultl be in
 a state of panic' Thc only way to avoid thi$        carried out on the lines of the Kinscy Reports
                                                     on humatr behaviour in another sphcre
 is to provide enorrghtrrrining for the correct
  resDon$e be nrade as the rtsult of lrabit'
             to                                      frar.rght lroth with risk and hazatd. Thc
  How many motorists in thc course of their.         conlidential character of the enquiry wcruld
                                                     probably disclose in a startling lashion the
  driving ncvcr take a risk in thc way we have
                                                     enotmorls risks habitually ttrken hy motor-
  defineclit'l Again, how rnarry motorists are
                                                     ists on the road' By chcckingthejudgemcnts
  there who though they never takc risks in
  our sense often get involved in gravc haz-         of a samplc of them in a way similar to our
  ards?Answers to there questionsirrc unlikely       stu<lyof bus drivers, the extent ofl hazar-dto
                                                     which they expose thernrielvcs    could also be
   to be f'otrnd in the statistics of the Ministry
                                                      estimated. enquiry on
                                                                  An              theselines hasnow
   of Transport, useful as these undouhtedly
                                                      been initiatetl.
   are for trnswering Parliamentary questions'

   -John Cohcn,Ph'D', E. J' Dearnaley, M'Sc', C' E' M' Hansel' M'A'

                                                                         in the face of
      In the second study, the behavior of pedestrians cIos8inB a load
                                   Measurementswcre made of the tine elapsing,      bc-
 oncoming trafl.-rcwas ohserved.
 i*.o, tHi pcclcstrian's first observationof an approaching:ut .u:d the arrival of the
                                                                differencesin crossing
 vehicle at the crossing. The demonstration of substantial
                                                                         the somewhat
 behnvior with age and-sexis of particular interest in connectionwith
                                   accidentinvolvement(seeHaddon       e+a/', Chap' 4)'
 similar variatiois in pedestrian
                                                                           of determin-
 The aul.horsbclicvc,rcasonably,that thesefindingsindicatea possibility
  ing where crossingcontrols ate necessary.
                                                                               study of
       These two studies constitute valuable contribution$ to the scientific
                                    in accident causation. Thcy clcmonstrate  that risk-
  cognitiveand judgmentalfactor$
              "suUlelctiveprohability" are subjcctto controlled study. Furthermoren  the
  i;fi"g anA
                                                                          a Elcat many
  model of decision-makingin risk situations may serve to org[rrize
  diversefi.ndings*on""rrli,[ individuat bchavior irr hazardousenvifoflments'Chance
                                                 jn                          asscssmcu
  factorsbecomethe subieciof controllcdstudy ternrsof thc individual's
 THE RISK TAKEN IN CROSSING A ROAD                                                              t4'l

of risk. Althoughnot enough knownahoutthssocial,
                           is                              physiological,
and otherfactors that determine assessment, approach
                              this        thi$       offers    promise
for future re$earch.

Iw ,e,pRtvrous pApERwe rrtudied subjective              In the present   study we did not obtain any
probabilitiesin risk-taking under conditions                                         in
                                                     e$timatesof likely succe$$ advance and
in which (a) the subjecthas varying amounts          the subjects wcrc unaware that thcy were
of infirrrnatiorr       thc
                  abor.rt task and about his          under observation,    The prescnce risk was
perfortnance, and (b) his cstirrrates   reltrte to   measured in terrns ol group attcnlpls to
diflerent levcls of dilliculty as rneasuredby        perform or not perform in valying circum-
his successin pcrfornrance. In the prescnt           stanccs, not in terms of the proportion of
paper wc proposc to extend further the study         individual successes.
of ccrtain aspects of (b). We shall be con-             The behaviourstudiedwas that ol people
ccrnccJwith cvcryday behaviour,                      cfossing a road frorrt a centre island to the
   In thc previous study cach subject esti-          pflvefl)rr)t, or })ide)'erJn,afialnst one strcam
matcd. in advance of his perforrnancc, his           of tralllc. It would not be far-fetched to
probable successat difl'erent levels of diffi-       assume that eftch pedestrian intentled to
culty, the difficultyor risk bcing rrreasurcd   in   crosr the road safely without being in.jurcd
terms of thc proportion of successes his   in        by a vehicle. A person's tacit beliefs ahout
actual perfornranccand not directly in terms         the chancesof safely crossing a road, rvhic-h
                                "f)ifficulty" or     wc regard as his subjectivc prob.rbility of
of units on a physicalsc'alc.
risk rvas thus defined relativcly to the indi-       salety (or injury), may he assumed fo bc
vidual, not in an absolute sense. Fot                related to thc spccd End distancc of ap-
exanrplc,in a given situation,suclras jutnp-         proaching vehiclc-s.
ing a bcam, if one roiscsthe height of the
bearn. thc lask bccorncs more dillcult to            Mnuon     oF ORSERvATToN
sornc cxtent for all subjccts. tsut for each            A crossing suitable for the present study
personthcre will be a particularheight ofthe         was chosenon a main road in a busy pirrt of
beam he will succeed jurnping, say, in 70
                        in                           Manchester.The full width of'the road was
percentof hisattempts.Instead mcirsuring
                                  of                 47 [eet, There was an island in thc ccntre <rf
di{ficulty in ternrs of the absolute hcight of       the road but no zebra crtrssing,  traliic lights
the beam,wc measured in termsof levelof
                          it                         or police control, Tra{lic density was mdder-
performance,that is.to say, thc di{iiculty of        atcly heavy. The study extcnded over a
a task for any subject was measured by the           period of one month ancl 1,189olrscrvntion$
proportion of his successesat any Biven              werc madc on 491 persons.
hcight of the beam.Onc and tlresamehcight               Four kinds of observation werc recorded
might yield dillerent levels of perfornranc'e        for each nrenrber of thc sarnple; (i) age,
for diffcrcnt subjects. Such a rneasure is           (ii) sex, (iii) specd and distnnceof vehicle,
indepentlcnt intlividualvrriations in skill;
               of                                    (iv) crossingor not crossing.
the 70 pcr ccnt p()int will vary in hcight for          Age and sex wcre beth estimated from
cliflerentindivicluals although it assunres   the    the pcdestrian'sr'rppcarance   and attire. A
szrme valuc for all the individuals as a             measurc of the speed and distance of the
mcasure of difficulty relirtivelyto the indi-        approaching vehiclc was obtirineclby re-
vidual. The $anreholds true of irny other            cording the time from the morncnt the
percentagc potnl.                                    person stepped to thc cdge of thc road and

                   with       *
           eFrintcd, permission,"* OU,                 Q t                 6
                                                xesearcA uar er Iy {London), :3: I 2tr |
      ff                              ;Il,: ial                                            J
                                                                 PSYCHOLOGICAL            APPROACHES

looked at the oncoming traffic to the mo'                   tion crossing and the varying degree of ob-
                                                            jective clangcr,as tneasuredby the speedand
mcnt when the lirst vehicle rcached the
crossing.Measurementswere takcn to the                      dislanceof thc vchicle.
ncarest sccond. Prcliminary ohservations                         We hirvc 1s51tlcd that a person's subjec-
had indicated that 'rryhcllvehicle$werc rrrore              tive probahility of safbty,that is to say' his
thrrn l0 secondsi{way pcdcstrianswithttut                   implicit belief about srrfelycrossing a tho-
exception would cross the road' Our analysis                roughfarc, is relatedto the speedand distance
 was thereforelimitetl to vehicles within that               of oncoming traffic. We may assumcfurthcr
 distance. Thus observations of whether                      that this rclationship holds not with thc
 personsdid or did not cross were mirdc in                   actual tims for the vchicle to approach but
 ielation to the spcedand distanceof vehicles                with the logarithm of the timc. This second
                                                             assumptionis pcrhapsrenclercd       more plau$-
 as meirsuredby thc lirnc taken by them to
 reach the site of the crossing.                             iblc by the fact that in a large variety of
    Pedestrians crossing in groups were ig-                  psychological Experimcnts a good fit is,ob-
 norecl.The number ol bicycleson thc road                    tained bctwecn the valuesof the indepcndent
 wa$ very small as comparcd with motor                       variable and the frcquencyofrcsponseifthe
  vehicle$and thcrcfore the observationsmrrde                values of the indcpendcnt variablc ate ex-
  of pedestrian behtrviour in relatiotr to                   pressed logarithms ol couventionalunits'
  cyclistswere not used in the presentstudy'                 Thit it illustrated, for example, in the p$y-
                                                             chophysical study of loudncss and the in-
 REsULTS                                                     tensity of sountl. It is also exemplifiedin
   (i)                                                        comparative judgments of the apparent
         ,ichtns-Tmc; W show in Table I                       duration of long intcrvals of time'
         Giidns madt when the velricle is at                      In vicw of thcse assumptionsand the fact
 dilTerentdistanceslrorrr the site of crossing,               that the pcrcentage of pcdcstrians crossing
 distance bcing expressed terms of the time
                            in                                the road incrcascswith longcr timc for the
 the vehicle takes to reach the site' C)nly one               caf to reach thc site, we should cxpect the
 Dcrsonout of 3(r4who had the opportunlty                     distributions of thcsc pcrccntages. wllen
                                                              plotted against thc logarithm of the time, to
 crossedwhcn the vehicle was 2.5 $scond$or
 lcss away from the site. When the vehicle                    ht a normal ogivc. Indeed, when the per-
 was 4 or 5 scconds away irbout htlf thc                      centagesare plotted against log time, a gootl
 people who could cross did so' When the                      fit is obtaincd, An ogive was fittcd by con-
 vehicle was about 9 or l0 sgspnfl5 awa]                       verting the actual precentirges to probit
  alrnost everyonecrosied.                                     valucs, A straight line was fitted to these
     Our task is nou' to determinethe nature of                values by the method of least squares' The
  the relationship holding between propor-
                                    the                        expcctccl probits and hencc the expected
                                                               pcrcentagcswcrc then calculttcd' Thc valu.e
                                           Cnosstllc   lx      of the tirne interval at the 50 per cent polnt ls
 TIsLE I'-NuMBER     AND Fnoponttou
           To rnu Ttmt TaKtN av rHE VuttlclE           TQ      0.66 log seconds (o ' 0.15 log seconds)'
                  R F A ( : HT H E $ r r E                         The expectedliequencieswere obtaincd by
   Tim{{+        Total        Number       Pertentage          multiplying the expected percentages by thc
              ohservations                                     total observations given in Table I' Whcn
   (.rers.)                   trussing
                 159              0             0.00
         4        1n\             I                            these expected frecluencies are ctltrrparcd
         3        184            11                            with the observetl (as given in Table II) we
         4        t59           60
                                                                obtain 12: l.8l; d.f. : 5; 0'tt0<p<0'90'
         5        lI0            65
                                 76            80.85               The observedand expectedproportions of
         6    9     4
         7    6     1            58            92.06            pcdestrianscrossing at each value o[ log
                                 37            92.50
         E    4      0
                                 34            97.t4            time are plotted in Figure l.
       9      3     5
                  140           138            98.57               (ii) Miximum Risk-taking Leve-ls:Let tts
  i The class interval in each case is *0.5   seconds'          assume that people tend to cross a r(tito
THE RISK TAKEN IN CROSSINGA ROAD                                                                               349

Trnrn II.-OBIIERVED AND E*PE(TED Nuxrtr                      TIBLB III,-Drsrnrsurrox     on M.qxlMuu E lsx-trrrrxc
cnosslrucIN REr.nrroN rur TIvti T.lxrN ron tnt
                      ro                                                           LF:vFL$
           Ventr-:Lr R.FACH SrTc
                   T()    tr-rr
                                                                           ExPeiled        Expertedpercentages
                                Ntmber          Number          Time*     perftntflges      crt.+sing ot not less
    nfie.      Log time         (rossing        trossing        (sers.)     trossing      thail th( stuled intprvdl
    (ser.r,)                   (ubservcd)      (exprtr.dJ         I            0.00                  0.0
       I         0.00               0              0              2            0-7                   Q.7
       2         {).30              I              I              3            to.E                 t0.r
       J         0.48              Z2             20                           rl.0                 ??.2
       4         0.60              60             52              5           59.5                  26.5
       5         0.70              65             6s              6           ?8.2                  18.7
       6         0.78              76             74              7           89.4                  l 1,2
                 0.85              58             56              E           94.4                    5,0
      I          0.90              37             38              9           97.3                   z,e
      9          0.e5              34             t4           > r0           98.8                    1,5
    Fl0          1.00             r38            138         ' Seefootnote Table I.
I See       TablcI.
 against approaching traffic at or below a                   represents the cumulative distribution of
certain level of srrbjectiveprobability of risk,             levels of maximum risk-taking.
that is, each person ha$ ir rnaximum risk-                      The expectedpercentagcscrossing at any
taking level. lf he believedthat therc was no                                        'Iable
                                                             time intcrval (col. ii,        lll) includc the
likelihood at all of reachingthe other side of               perccntagc that would hirve crosscd at a
the road, he would not attempt to cross. On                  shorter intrrval. For example,59.5 per cent
thc other hand, only very rarely would hc                    are expected to cross at 5 scconds. Thesc
cross a road with a senseof absolutc safety,                 would doubtlessalso cross at longer time
Presurnably there is a value of subjective                   intervals.Similarly,33 pcr cent are cxpected
probability ofl risk above which people will                 to crossat 4 or more seconds.   The diffcrence.
seldom, if ever, cross.This value of subjective             26.5 per cent, would cross rrt 5 or more
probability doubtless varies between indi-                  secondshut not at lessthan 5 seconds.        This
viduals and f<rrthe same individual at dilTer"              mean$ that 26.5 per cent have their maxi-
ent times, In large samplcsthcse variations                 nrunr risk*takinglevelat 5 seconds     (or, to be
may bc itssurnedto be smoothcd out, as in                   morc prccise,between4.5 and 5.5 seconds).
the major part of Table l. Table I therefore                These risk-taking levels are lpg normally
                                                            distributedwith a ntcan at 0.66 log seconds
                                                            (a : 0. l5 log seconds).    The maxirnurnrisk-
                                                            taking level nf 50 per cent of the pedestrian
                                                            population, in the circumstanccs      undcr dis-
                                                            cussion,is reachedwhen the vehicle will take
                                                            4.6 secondsto arrivc at the site,
H                                                              In thc next section we shall see how the
                                                            mean maximum risk-taking lcvcl varicswith
H                                                           age and sex.
                                                               (iii) 1{            Between               and

                                                            are groupcd by age and sex in Table IV. In
                                                            Table V wc show, for males and fcmalcs
                                                            separatcly, log time intervals
                                                                         the                 correspond-
                                        "';"                ing to the mean values of maximum risk-
                          r#   ,tCoHE
                                                            taking in errchsub-group,Thesewere calcu-
    Ftc. t. Percentage pedestrians
                       of             cross-                lated by litting ogivesusingthe samemethod
    ing and not crossing differentintervals
                        at                                  as previously.
          of time: best fitting ogives.                       A gooclfit to a normal ogive was found in
                                                                                   PSYCHOLOGICAL          APPROACHES

TABLE IV.-DIsTRIBUTIoN     or Act                arqn Stx nr tnu             more ancl more readily' We might expect
                     Sllr Pr'r:                                              that thc mnximum lcvcl of subjcctive risk-
                                     SEX                                     taking is much closer to an ohjcctively
      Ag*                   Male           Femule            Total                        "risk" in sittration$ whcr:cthere
                             54              20               74
15 and urder
                            142             262              404             is no danger than rvhere there is d;lngcr' Wc
3l-45                       247             230              477             havc no cvid€ncc that there is a gerreral
46-60                        E0             lll              l9l             factor of risk-trrking, either betweon persons
6l xnd over                  14              29               43
                                                                             or within Persorls.
  Total                      537             652           1,189                                                      emergc
                                                                                A number of practical suggestions
Tlsrr       V.-Mlxttt,tuu          Rtsr'tlKlNc           LtrvELs FoR         from this study' Childrcn's traffic sense
  M n r r : l n o I - L M A L LS u B - c R o U F s : M E A N V A L u l l $   might bc improvctl if they rcceivcd training
                         15 ttnd                                 6l and      in judging the specd and distance of ap-
                          uurk:r 16'-30 3I'15            46-60 ovttt
                                                                             proaching vehiclcs. Potential car drivers
   Mean values            0 ' 5 6 0 . 5 5 0.66 0.55 0.65                     nright also profit from such training, and
       d                  0 . 0 7 0 .1 4 0 . 1 2 0 . 1 0 0 . 0 1             appropriate tcsts rnight be irrtroduced'Thc
       n                   54      1 4 2 ?'47 80          14                                                            might
                                                                             risi.+atlng levels at z-ebtir-crossirrgs
    M e a n v a l r r e s 0 . 7 3 0 ' 5 5 0.67 0.68 0.76                     be studietl. thus measuring the ellectiverress
           d              0 1 0 0 ' 1 0 0.l2 0.06 0.02                        of such devices.    The training of men in the
           n               l0     762 2 1 0 l l l    29
                                                                              Scrviccs,cspccinlly in thc R.A'F', Dpcnstp
 all cases where thc size of ihe suh-group                                    a ticlcl for thc study of levelstlf maximal
 made it rlossiblc to apply a chi-scluaretest'                                optimnl) risk-taking in reltttion to the task
                                                                              and the operators.
 Whcn comparing adjacent age groups, in the
                                                                                Our conclusion$ mfly be briefly ststed as
 casc of thc males, the mean is signilicantly
 highet for the 3l -45 group than for thc
 t{ir0 gr.,up. In the case of thc fcmales, thc                                   l. A study was mirdc of 491 persons of
                                                                              varying age in relatiotl to crossiug or not
 -.un it lowest lor the 16-30 group and
 increases irfterwilrds in the highcr age                                     crossing a road against oncotning tralfic'
                          l5 or less,however,                                 1,189 observalionswere recorded'
 Eroups;thc group rrgecl
 also-has a high value. The differentres    be-                                  2. Fifty per cent ol'the sarlple who corrld
 tween the meins tbr the two sexes iil the                                    crosswhcnllt" vehiclewas irhttut 4'5 seconds
                                                                              awav did. in fact, cross.
 sarnc agc group$ irre all significant with the
  exccption of the 3l-45 age groups' (Setr
  Figure 2.)
        "ort express thesc rcsults in terrns of
  clock timc. Thc means range from 3'5
  seconcls the l6-10 male agc-groupto 5'8
  sectlndsin the 61* fcmalc irge-group'                                       7

                      AND CoNcLusIoNS
  DtscusstotqoF RESULTS                                                        u
     Thc problcm investig'.rted   here in rclation
  to oetlestrian bchaviour      is general ilr its
  relevanceto cvcryclaylife' Any individual
  cncountcring daily hazards probably tends
  to acquirc a given rnaximum risk-taking
  level in sirlilar types of situations' The                                            1 3a u d . r     ll 4r

                                          point at
   maxinrum risk-taking level is thc                                                                   Aot clouP
   which he will bcgin to undertake a
                             rulc act at a level of                                Frc. 2. Maximum risk taking valuesfor
   task. Fle will rrot as it
                                                                                   rnnlg and female sub-groups: rnean
   hazarcl grcatcr than lris tlttximttrn' As the
   Ievel of hazard decrease$ will tcnd to act
 THE RISK TAKEN IN CROSSING A ROAD                                                       351

   3. No one crossed when the vehicle was          Norr
 less than 1,5 seconds away, and evcryonc            At the time this paper was written, we
 crossedwhenlhevehiclewasmorethan lD.5            were unfortunatcly unaware of the work
 secondsclistant,                                 carried out by R, L, Moore previously
   4. Whcn thtr percentages crossing at           described in this journal (Operational Re.
different intervals of time werc plotte<l         .seart'h Quarrtrly) <:tr of relcvant studies
against thc logarithms of the time inter-         (unpuhlished)that have bccn rnade at the
vals,a good fit tr.r a normal ogive was ob-       Road Research Laboratory. It sccms that
tained.                                           our resulrs are in accord with those of
   5. The maximum risk-taking levelswhen          previpu5 investigiitors. Fcrr exanrple, Mr.
mcasurcdin tcrtls of log ssiJ11r45 norm-
                                   ate            Moorc observesthat "pedestriansare con-
ally distributcd.Tlrcsevtry frorn individuat      ccrned primurily lvith a tifie-gap and not a
to individual and prolrablyfroru situationto                     in
                                                  disiancc-gap the trallic," ]{e rrlsoremarks
situation,                                        that cach peflcstriirnpresumably has his own
   6. f)ur study indicatesthe possibitity of      critical distirncegap which seerns refcr to
deterrniningwhere cotrtrollcd crossingsarc        what we huve describcdtrs"rnoxitnutn risk-
necessary taking into con$ideration the           taking." Our finding that no one crossed
dcnsity of tralfic and the tinre availirble for   whcn thc vehicle wris less than 1.5 seconds
pedestriarrs cross.
             to                                   away seemssupportedby eirrlicr enquirers.

    Thesestudiesrepresent   highly original experimental  work on the kinds of decisions
that are ilrade hy individuals in everydaylifc. It is to bc hoped, however, that, as
knowlcdgc from such prcliminary studiesaccumulates,        future researchwill involve
more complexsituations     anclcollcctmore detaileddata concerning experimental
$uticcts and the differences thril rcitctionsto experimentalas opposcclto reallife
hazalcls.  Suchincreased  sophistication   woulclpermit an analysis thc influcnce
                                                                     of              of
various social, educational,cxperiential,physiological,and psychological        factors
upon risk-taking.
    As an e.xtl]lpleof the utilization of observationof behavior situationswhich ap-
proach thoseof real lif-e, first <;f thcsesludiesis a nrodelof experimcntal
                           the                                                  design.
The variable to be testedis carcfully definedirnd isolated,and appropriate controls
irre introduced to avoid the contaminatingeli'ectof extraneous    variables.However, it
is apparentthat the conclusions      reached are limited and cannot be appliedwithor.rt
reservationto Inorc cc)mplex      risk situations.              and major problem in
                                                  An aclclitional
work of this typeis the t'actthat thc hirzarcl usuallyqualitatively
                                               is                    ani quantitatively
differentfrom its real-worldequivirlcnt.One possibilityfor overcorningthis is through
the devclopment experiruental
                  of                                      the
                                     situations-,including useof simulators*which
are extremely  realistic.An additionalapproachis to study the rcal world, as in thc
pedestrianstudy presentedabove.

  --Joltn Cohen,Ph.D., E. J. Dearnuluy,M.Sc., C. E. M. Hansel,M.A.

    In a related experiment on the effect$of alcohol on driving, the same investi-
gators compirredthree groups of' bus drivcrs, two of which had rcceivedalcohol,
in regarcll,o their decision-rnaking a risk situatiun (driving a bus through gaps of
varying widths).
352                                                    PSYCHOLOGIC]AL APPROACHF'S

                                                                      from the finding
   The importanceof alcohol on a driver's estimateof hazardis seen
that drinking resultccl clrivc|s'bothbeingwilling
                      in                              to drive thr.ough narrowergap
and needing'awider gap to succcccl.  Thus  we seethat the dccisionto take a risk is a
                                                                     as the changesin
discrcteeffcct of alcotrol of perhaps at least irs great importance
perception and motor behavior    which it also produces'f

Arcosor, As A cAUSEof accidents on the           we mean embarking on a task in which thc
                                                 performer will not invariably succeed.       Thut
road has been widely studied. The investi-
gati()nsthat havc bccn carried out. valuirble    risk is   measured by estimates of succc$s
as they Arc, suffcr, howcver, frorn a serlous    beforc the task is attemptcd, whilst hazirrd is
                                   deterrnined   flleilsurq-d the proportion of successes
                                               the            by                                 in
limitation. They havc usrrally
eflect of alcohol on thc drivcr by rneasuring    actual pcrfbrmance. ln our provious studies
changes his reactiontime and in his skills'
          in                                     ofrisks tnd hazardstaken by inexpericnced
                                                  and experienced     drivers we found that, on
such as steering and btaking. Lrrrge quatrti-
ties of alcohol do, of course' produce de-        the avcrage, experienced drivers tended to
terioratiorr irr skill and retction time, but the (1) take lcss risk, (2) incur less hazard'
conclusionhas too hastily been drawn that         (3) vary Iess in thcir driving behaviour,
                                                  (4) be more realistic in judging their driving
bccausc no changes have been rccordcd
 after small quantities toad sirfety is unaf'     ability, and (5) bc prepared to drivc only
 fcctcd. This conclusion is not justificd, for    when they f'clt secureand were in firct safe;
 all that can be inf'erredfiom carlier inqtriries on thc other hand, an iustructor with long
 is that these skills and rcaction tifles are or  exnericncedrove without risk or hazard.
 ate not affcctcd. It cannot, for exarnple, be    That is to sayr lte trcvcr attempted a task
                                                   unlesshe was sure hc would succeed,       and he
 inferrcd that a driver is not a mendceon thc
 rofld because his reaction timcs are unirf-       neverfailcd in any task which he undcrtook'
                                                      Thesc rcsults suggested   that a cotnparison
 fecteclor becausea ttst of skill rcvealcd no
 impairrncnt. Ttre decisive feature is not the     of drivers who hnr'1irnbibed variou$ quan-
  driver's skill in itsclf but in relation to whrrttitics of alcohol with dlivers free of alcohol
  he believeslre coulcl do and what he would       might rcveal sinrilar dill'ercllccs.The presetrt
                                                   inquiry was designedto ilrea$ul'c      the influ-
  in fact undertake. Thus thc ncgative out'
  corne of these studies nlay $ignify only that    ence of alcohol on those drivers who nright
  thc methods of illea$uremerrtwere inade-         be expected to show its effects least rrirmc-
  cuate. not that thc alcohol was without          ly, bus drivers of long expcrience with re-
  cffect on the safety of the individual's         cords of safe driving' If any impairment due
 driving.                                          to alcohol could be demonstrated in such a
    Our aim, therefore, is to asscss  the cffccts group it would lollow that the consequcnces
                                                                                      ytlunger, less
 of alcohol on the more        complcx psycho' rigtl1 U. flar more seriouri in
                    and performancesinvolvcd experienced,         or lessskilled drivers' Wc hopcd
 logical processes
 in riskf which drivcts takc and in haz'ards tlrat, by introducing
                        "risk-taking" we rncan at once rnore seflsitive and more funda-
 which they   incur. By
                                                                 generalsenscas well as in a $ense
 embarking on a task without being certain m;ntal in a
                                                                                           cffects of
  of success. There  cirn be various dcgrecs o[ specific to driving, any possiblc
                                                             quantities of alcohol might reveal
 certainty of succcss and hence varying smaller
                             "incurring hazard" thcmselves.
  degrees ofrisk'taking. By

                          1958.Appendix 2 has beenomitted'                J
  f Seealso refercnce
  RISK IN DRIVING UNDER THE INFLUENCE OF ALCOHOL                                                                                                               353

  Suelncrs lltn Pnoceuunn                                               experiments were to be conducted, The
    Our subjectsconsisted of three groups of                            driver's tffsk wirs to estimate how successful
 bus drivers working for the Manchester                                 he would be in driving a bus through gapsof
 Corporation Transport Department. They                                 varying width. Having nradehis estimates     he
 had had on the average about 12 years'                                 rva$to drive the hus through parriculargaps.
 experiencc driving a bus and, in addition,
               of                                                          Two verticalposts,3 ft. (0.9 m.) high, J in.
 anothcr cight years or so of driving other                             (7.6cm,) squarc,and paintedwhite,stood on
 vehicles.  The depotswhcrc thc mcn workcd                              a linc. Thc gap trctwcen thcrn could be
 were notificd that the investigation was to                            varied and nreasured     from calibrations on
 take place, and drivers with awards for safe                           the line on which they were placcd, the
 driving were invited to volunteer to takr                              calihrationsbeing visibleonly to the experi-
 part. Tho arrangementswere made through                                menter, Beforc thc driving sessionbegan a
 lhe personnel officer and reprcsentativesof                            bus 8 ft. (2.44 m,) widc was placed at a dis-
 the trade union. The averageage of the men                             tance of 12 ft. (3.66 m.) from the girp and
 in eirch group was 45 years. There was one                             fircing its centre,
 bachelor afld one widower in the total                                    The steps in the experimental procedure
 sample; all except seven subjects had ftrnri-                          may be summarized as follo*'s:
 lies. Drinking habits among the drivers
 varied from total abstinenccto an avcraEeof                                  L Thedriversatin hisseat wastcld that
                                                                          he was shortly to drive his bus hetwecn the two
 2 0 t o 3 0 p i n t s ( 1 1 , 5 t o 1 7 l i t r e s )o f , h e e r a     posts in front of hirn. f irst, howr:ver, he was to
week; very fcw of our suhjects                   drank spirits            (cfl us how wcll he thought he could do this at
                                                                          difl'erent lizcs of gap bc-tween the tw{) posts. At
 of any kind regularly.                                                   each prcarrangcd size of-gap hdlweeil the posts
    All the drivers had worked an eight-hour                              t h e d r i v e r w a s a s k e d t r ' s a y h o w I n d | ] y t i n r e so u t
nrorningshift immcdiatclybeforethe erperi-                                6f five he could drive through without touching
                                                                          thc posts. fhc gap betwecn the pr:sts was in-
 mcnt and rcportedto u$ irt I p.m. They then                              creased by intcrvals of ? in. (5 cm.1, beginning
took lunch, without alcohol. During the                                   with fl gap of 7 ft. 4 in. (2,?-l m.) or snraller if a
rneal thcy inforrned us about their drinking                              d r i v e r t h ( ' r u g h (h e c o u l d d r i v c t h r o u g h t h i s g a p .
                                                                          This was conlinued unliJ lho driver said he could
habits and driving experience.                   The alcohol,             sucteerl tivc tir$es out of fi!e.
in the form of Scotch whisky (70" proof or                                     2. We next detcrmined the least size of gap
401" alcohol), with an equal amount of                                    through r hirh the driver would u(tudllr aIt(mpl
                                                                          to drire. Here, too, lhe post.Erverc nroved apitrt
soda-watcr, was taken from 2 to 2,30 p,m,                                 in sttges except that the gap was now increased
The atmosphere was convivial.                                             by intervals of I in. (?,5 cm.) at a time. The size
                                                                          o f t h e g a p b e t s ' e c n t h e p o s t s w a s i n c r e a s c du n t i l
    Tlrc driving scssions            began at 2.45 p.m.                                                                                  l
                                                                           l h c d f j v e r s i e n a l l c d h i i i n t e n t i r . r nd d r i v c t h r o u g h
and lastcd urrtil about 3.15 p.m. Thc blood                                it, He attenrpted to do so withc)ul t(lurhing the
samples (0.5 ml.) wcrc taken lralf an hour                                 posts and then drove rounil thc traitrirrgsite so
                                                                           as to rc(urn lo thc ritarting-point. He tlrrJvcat
later. The pnrcedureferr the control group                                 bis own sFeed.
(without alcohol) v4asexirctly the same as                                      3. Finally, we mca$urcd the driver's perform-
                                                                           ance ai diffcrsf,t sizss 5f gap. The gaps through
that for the two other groups except that                                  which he u'as itsked to drive his bus wcrc such
they proceeded to the experiment irnmedi-                                  as to rcveirl the snrallest gilps at which he would
ately after the meal and thc blood test was                                succeed (d) once or twice orrt of fivc, (b) three or
                                                                           four times out of five, and (c) five time$ out
omitted. The moderatedose of whisky corr-                                   of live,
sistedof 2 British fl. oz. (or 56.8 ml.), being
the equivalent approximately of 23 ml. of                                 The experiment$ lyere carried out at the
ahsolutealcohttl,The larger doseconsistedof                             Training School for Bus Driverri of the
6 British fl. oz. (170.4ml,) of whisky, being                           ManchesterCorporation Transport Depart-
thc equivalent approximatcly rrf 68 ml, of                              ment.
absolutealrchol,                                                          Before presenting the rcsults, we wish to
   After consuming the whisky the drivers                               make cerlain colflntentson the above pro-
wEre taken to thc sitc where the risk-taking                            cedure. First, it might be askcd whether
                                                                           PSYCHOLOGICAL         APPROACHES

different estimates might be obtained if our they were always able to drive succe$sfully,
subjectswere rrskcdto say how many times,    as shown in Table I.
say, out of ten (or a hunclred) thcy would       On the average,the drivers in all thrcc
succeed.We hale shown elsewherethat the      groups were prepared to drivc an 8-[t'
nrean c$timfltesof adults, when expressed    (2.44-m.)hus thiough a gap lessthan I ft' in
fractions                                    wiclth, the propor:tionsin each group rrndcr-
             of uni';y--that is, as subjective
probabilities arc unaffected by thc range    taking this heing tifi, 75. and 80'fi respective-
within which they arc made.Thus estimatcs    ly. [Jndet our experimental conditions,
madc within ofle tangc are transposable to   thereforc, cven the alcohol'free drivers wcre
anotlrer.Seconrjly, might be askcdwhether
                       it                    involved in hazard. The effect of the alcohol
the estirrrrrtcswould clilTerif the srrbjcctsis, however, revealedin the degreeof hazard
to say how rnany times they would fail ratherincurrecl hy the drivers in the threc groups'
than how many times they wor"rld             Three infcrences rclating to the eflect of
 Here. too, we htve r.hownthat, at any rate'  alcohol may be drirwtl frour Table I. First' as
 whcn thc rangeis                             the clrivcrs consutrted more whisky they
                      "out of five," estim;rlesof
 success and estim'ntcs failure appear to be
                          of                  wcrc prepared to drive through it nitrrower
 additive.                                    gap. Sec:orrd, they cottstttttcdmorc whisky
    A third lroint arisesin connexionwith the thcy needetl f, wider gap bet'orc thcy could
 ordcr in which the e$timates                 always succeedlthus there was rr progrcsstve
                                 were made' For
                                              deterioration in pctformance as ilore whisky
 prrrctical tcasofls, it was only convenient to
                                              was consurned.Third, it follows that the
 preseilt tlre cntire sefiesof gaps in uttentlittlg
 order of magnitudc' In principle it would    margin of hazard, as shown in the fourth
 have bcen equirlly possible to present them  coluinn, increases with greater intirks of
 irt desrentling random order of nragrritude'
                 or                           alcohol, The difference between the twcr
       ascendingorder, as we know from our    extrcmc groups is significant (t : 2'2,
 previortsexperiments,hirs thc clTectof yield- df - 17, p<0'05). Wc tnay therefore con-
 ing larger estimatesfor eaclr gap (or level ofctude that, as cotnparcd with the alcohol-
 clifficulty)than the descendingorder. The     frec clrivers,those who had taken 6 fl' oz' of
  ascending order of presentationwas also      whisky (68 nrl. of absolute alcohol) were
  employed in rclation to the least gap which  involved in considcrably greater hazard'
  the drivcr was prepared to attempt- This        Rl.r/c: Whilst grcater hazarcls werc tn'
  feature of the procedure tlust be borne in   cu.r*d by irnbibing alcohol, the risks takcn
  rnind when considerirrg     the results.     were unaffected.In othcl words' althouglr
                                                the drivcrs wcrc prcparccl drive through a
 Rtsur,ts                                       nflrrower gap, their confidence succcss
                                                                                  of         at
   Hazttrd; The first results which we wish this     gap rcmaincd utrchanged'At the nar-
 to-trrrcscnt revcal the glcatest degree of rowest gap which they attcmptcd the mean
 hazard in which our three groups of drivers cstitnatesof success tlre groups taking
 became itrvolved. This  is indicated by the 2, anct6 fl' oz' respectively   were 2'8, 2'1' and
 differerrccbetween the mean of the narrow- 2,8 timcs out of five. We carrexpress
 cst Bap which the drivcrs atternptcd and the subjectivc    probabilities by converting thcm
 mean of the narrowest gap through which to fractions of unity, thusl 0.56' 0.42,

                                                                                       : 8 rr' (2'44 ru')
                             Trsu    l.-M,tnctNs    oF.H,qzrnu (Ms'rrrqs)*Wrnrn or Bus
                                                                        NARROWEST 6AP                DIFFFRENCE
                                            NARR0WEST GAP
                                                                       AT WHICH DRTVERS              (MARGIN oF
                                            WHI.]H L)RIVERS
             QUANTITY OF                                                                                 HAzARt))
                                              A T T L M PI [ D         ALWAYS :tU(](:EtrDLD
               WTIISKY                                                                                      (4)
                  (l)                               (2)
                                                            m.                                     ln,                cm.
    B r i t . . f t .o z '    ml,         ft. in.                     It    ln,        ffi,
                                                                                                                    - r 8.ll
                                           7 il.3         2.42O        8    6.7      ?.610
              0                0                                                                  - 9-3             - 2t.6
                              56.8         7 e.6          2.315        I    6.9      ?.615
              2                                               'l'ln                              -11.5                a o f
                             170.4         7 7.8          ?            I    7,3      7,62,5
 RISK IN DRIVING UNDER THE INFLUENCE OF ALCOHOL                                                                                                                355

                                                                Trnle ll.-Mencrus                 or Sa,rcry(MrlNs)
                                                          NARRL]Wf,ST (iAP                       NARRT)WIST dAP AT WHTCH                      DIFFEEEHCE
           QUANTTTT OF                                     wHl(:ll DRlvERs                         urrvrxs 6rlir'rty' rsEy                    (MARCIN QF
               WHISKY                                       AT] TM PTID                          WOTJLD ALWAYS SUCC:EED                         SAFITY)
                   (l)                                                (2)                                       (J)                               (4)
     Brit. fl. oz.     ml.                           ft. in,                       il,             ft.    tn,          m.               tn,              tm.
              0                   0                   7 tl.3                    2.42D               I     t.3         ?.47D             -'t/]           -5     |
              2                  56.8                 7 9.6                     2.375               I     1"4            5
                                                                                                                      2.47             -3.8             * 9.6
              6                 170,4                 7 7.8                     ?.330               7    10.5         2.400            -2.7             - 6.9

 0,56 respectively.These values do not differ                                                    even thc drivers in the control group dis-
 significantly. Accordingly tlre effect of alco-                                                  played an exccss of confidence. Thcy be-
 hol was not to make drivers take rnore                                                          lieved they would succecdbcfore they could
 risk-that is, act at a lowcr lcvo-l subjcctive
                                    of                                                           actually succeed. But this cffcct wits nruch
 probability-but to assignthc same subjcc-                                                       more marked in the group which had taken
 tive probability to a rnorc diflicult tnsk. If the                                              6 fl. oz. of whisky, Thc influence of the
 difficulty of the task rcmained unchangcd                                                       alcohol thus Ied to a tendencyto overrate
they becamemore optimistic and attacheda                                                         ability in relation to perfbr.rnance,                   even in
higher subjectivcprobability to the task,                                                        our group of experienccd drivers. The fact
    Furthcrmore,on the averflge, narrow.
                                    the                                                          that two whiskieshad little or lto cllbct on
est gap (7 ft. 10.5in.-2.40 m.) at which the                                                     our highly sclected grrtup should by no
high-alcohol group belicvcd they could al-                                                       means encourage any ordinary driver to
ways succcedwas smaller than the ntrrr()\\'est                                                   indulgc himsell to that extcnt or less.
gap (7 ft. ll.3 in,--2.42 m.) which the alco-                                                        Blood Akohol in Rtlution to Whiskv Con-
hol-free drivers even ventured to attemot,                                                       sumed: Analysis of the blood sarnples
    Alcohol thus produced dangerouscHccts                                                        showed that thc rlcan level of btood rlcohol
apart from inducing the drivers to incur                                                         was 0.04 rng./ml. (o - 0.07) in the group
Ereater haz.ards,ln addition ttr influencing                                                     taking 2 fl. oz. of whisky ( - 2l ml. of abso-
the drivers to attenrpt to drive througli a                                                      lute alcohol) and 0.58 mg./ml, (o : 0.17) in
narrowcr gap, it also induccd greater confi-                                                     t h e g r o u p t a k i n g6 f l . o z . ( : 6 8 m l . o f a b s o -
dencc of success a narlowcr gap. This rnay
                   at                                                                            lute alcohol), The differencebetweenthcse
be seen from column i of Table lt. To il_                                                        tw() mcansis significant(t . 12.5,df : 39,
Iustrate: the subjectiveprohability of the                                                       p<0.001).With one cxccptiont is possible        i
high atcohol group for a gap of 7 ft. 7.9 in.                                                    to statefrom a knowledgcof his blood alco-
(?.33 m.)- the nirrrowcstthey attempted_                                                         hol whcther a driver had taken 2 or 6 fl. oz.
was 0.56,as comparedwith 0.34for the non-                                                        of u,hisky.It follows lhat the mean level of
alcohol group. Similarly. becar.rse values
                                       the                                                       blood alcohol jn the two groups is reluredro
in both colunrn 2 irnd column I of Tablc tl                                                      their respective         mFanmarginsof hazard.But
dcclinewith increasing    alcohol, the nrarsins                                                  the variation in the margin of hirzard rlrt}ra
of safety remain rnorc of lessthc same,the                                                       the groups is so large as cornpared u,ith the
smtrll variationsin column 4 not bcing sig-                                                      variation httu:eenthem that a knou,lcdgeof
nilicant.*                                                                                       the blood alcohol of any individual driver
    Estimates and                r The effect of                                                 tells us little about thc dcgrccsof hazard in
alcohol may be shrtwn in a third uranner fas                                                     which hc mighl. becomc involved.
in Talrlc lll), by c{rrlparing the size of, the
gap at which the drivr:rsbelieved    thcy would                                                  Drscussronor Rnsulrs
always succecdwith thc gap at which they                                                           It is a well-known fact that individuals
did in fact always succccd.                                                                      vary widely in their rcncticn to thc same
    In the circumstanccs this experiment
                           of                                                                    amount of alcohol,Our rcsultsbcar this out.
                                                                                                 Some of our subj6sls secrlr to hrtve been
i The ellect on spccd of driving is $jrnilar, As VernoD
                                                                                                 more adverscly affccted two whiskis5
                                                                                                                        by              than
has shown, modcratc quantities of alcohol tend to
m d k c r n o $ t d r i v e r s i r r r : r c a s eh e i r r a t c o f d r i v i n g , l h i s
                                                  t                                              others by six. Hcnse, so lar as the practical
tcrldcncy being risually uilrccognized by thcm,                                                  implications of this invcstigation are con-
356                                                          PSYCHOI,OGICAL APPROACHES

                      Tlnr-r ttf.-Estlrulres   ir,t RErArtou ro PtnroRMentcr (Mrlns)
                          N^RROWES'I CAP AT WHICH         NARI{OWEsT    CAP
                oF          nntvllts hr'Iievtrd tttsY     AT WHIC:H DITIVERS             DIFFERENCE
         wHISKy           WOULD    ALWAYS sLI(:(IEBD     ALWAYS SL](](:EEI.lED
              (l)                     (2)                         (3)                          (4)
                                             m.         [t. in,          m,             in,           tm,
  Brit. fl. az.    ml,      ft. tn,
                             8 r.3          2.470        8 6.7          ?.610          -5.4
        0           0                                                                  - 5.5         - 14.0
        2          56.E      E 1.4          2.475        I 6.9          2.615
                                                         I 7.3          2.625          - 8.8
        6         170.4      ? 10.5         2.400

cerned, we have to bear in mind not only associatedwith a reduced effect of alcohol.
deterioration in the averagc performance of      The absenceof any significant correlations
drivers in gcncral efter taking alcohol, but betwecn perfoilnance alld the age, weight,
also extremc effects on n small proportion       drivirrg experiencc,  and drinking habits of
of individuals.                                  our drivers may be attributed to the homo-
   After taking alcohol certain drivcrs at- geneity of thc sample,
temptcd to drive through cxtremely llilrrow        Thc reader may wonder whether our
gaps. No alcohol-free driver tried to drive results strggcstany level of blood alcohol
his 8-ft. (2.44-nr.)bus through a gaP lessthan which should not bc exceeded' the United
? ft. 5 in. (2.26m.) in width, whereas thrcc of States, the National Safcty Council recoln-
the group taking 2 fl. oz' (56'8 ml.) were mended that the lincs of demarcation be-
prepared to undertake this task, one ofthem      twcen levcls of blood alcohol should be as
bcing    willing to attempt a gap 1 ft. 2 in'    follows:
(35.5 cm.) narrower than his bus, his sub-        0.0-0.5       mfe
jective probability of success at this gap        0.5-1.() " possihlyunderinfluenceofalcoh
                                                  I.0-1.5 " " probrbly
 being 0.8, although his blood alcohol was
                                                  Above .
 only 0.02 mg./nrl. Two drivers after 6 fl' oz.       1.5 "  "  definitely
 (170.a ml.) attempted gaps which were
 respectivcly I ft. 2 in' (35.5 cm.) and I ft'   We do not, however, think it is justifiable,
 4 in,  (40,6 cm.) narrower th;rn the bus; the on the basis of our cxperiment, to establish
 blood alcohrtl of the former was 0.74 mg'/rnl. the interval of 0,0-0.5 as safe' We have found
 and of the lrrttcr 0.?6 mg./ml., and the sub' that thc trustworthinessof a man's jtrdgment
 jcctive probabilities crf successwcre 0.6 and of his driving skill was irlpaircd evelrafter a
 zero respectivcly.                              small quantity of alcohol producing a blood
    It is perhaps not surprising that in our alcohol concentration lower than 0.5 rng'/ml.
  very highly sclectedand expcriencedsarrrple This impairment ol judgrncnt occurred in
  ofdrivers there should have been one or two what might be described as the i/ite of
  who compensated for the effects of six drivers on the road. Furthermore, the effccts
  whiskies hy bccoming extremcly cilutious. of the alcohol were in a $euscminimal, as the
  One driver, after 6 fl. oz. (170.4 ml') of alcohol was taken immecliately        after a meal.
  whisky, refrained from making an attempt The performance          ofthe drivcrsdeteriorated'
  until the gap wa$ l0 in. (25.acm.) wider than they were involvcd in greatcr hazards, aud
  his bus, though hc believedhe could always they displayed a false conlidetrcc in their
  succeedwhen thc Eap was I in. (20.3 cm.) driving abilitics.
  widcr, and. in fact, he always succeeded          Ordinarily a driver suspected of being
  when it was 7 in.  ( 17.8cm.) wider; his blood under the influence of alcohol cannot be
  alcohol was fls rnuch as 0'60 rng./ml. The given a test of safe driving' We suggest,
  blood alcohol of thc most dangerousdriver tlrercfore, that rncasures taken to reduce
  was less than half that of the most cautious traffic accidcnts due to alcohol must forbid
  one, This relatively dangerclus driver was thc taking of alcohol firr some $pecified
  -5stone(31,8 kg.) lreavicrancl 15 yearsolder period bcfore driving, The menaceto public
  than the cautious one, though a greater safety comes particulaily from those who,
  weight and age rrright be expected to be after a slight anount of alcohol, fecl they are
BISK IN DRIVING UNDER THE INFLUENCE OF ALCOHOL                                                 35'l

sober or beiieve they are driving more              investigrtionsin differcnt countries,would
cautiously-                                         appear to be of the order of 5A/,. The posi-
   In Grcat Britain it is extremely difficult to    tion appcarsto be much the samc in many
 assess ell'cctsof alcohol on the incidence         other countrics. According to Piret there
 of trafilc accidents. Thc rcason is that the       appear$to be a conspiracyofsilence bctwcen
 police, when reprirting an accident,are cx-        the iiuthoritjr.-s,
                                                                      the lawyers, and the wit-
tremely reluctant to state that they strongly              to
                                                    nesses wrap thc cffectsin a cloak of inno-
$uspect   the driver of being under the influ-      geflce, Only thc mqrsl drastic mcasures,
ence of alcohol unlessthey chargchint with          firmly carried out, could cope with this
being drunk. It is well known that the police       appalling state ol affairs.
are very unhappy about this statc of affairs.
They are well aware tlrirt even when the                Our rnain results may be briefly summa-
driver is charged with bcing drunk he stands         rized as follows:
a good chanceof being acquittcdif hc clects             Drivcrs who took alcoholbccameinvolved
to go for trial beforea j ury at quartcrscssions.   in greaterhazardsthan alcohol-free    drivers.
   Wc have been privately told by many                  As the amount of alcohol taken wtrs in-
membersof the policc force that they have           creased,the drivers wcre prepared to drive
no confidence the tables publishcdby the            their vehiclesthrough narrowcr gaps. This
Ministry of Transport and Civil Aviation            revealed that the alcohol had adversely
and Scottish Homc Depar(ment,giving de-             affected their judgment.
tails of the -'factors whcre drivers wcrc re-           The performance o[ the drivers, as well as
garded by the police as contributing to the         their judgment, pregrcssivcly  deteriorated as
accident." From the figures in these tatlles        they consumed morc alcohol.
the influenceofalcohol flppeiirsto be a flegli-         Aftcr taking alcohol the drivers become
gible factor, rvhereas police believethat
                         thc                        morc dangerous.    although they did not take
the figures give a grossly misleading im-           grcater risks. This was due to the fact that
prcssion. Thus, for e.xample,in 1956 there          the lcvcl of confidenceabove which they
were2 16,172    road accirlents Crcat Britain.
                                in                  were prcparcd to drive their bss remained
       influence  ofdrink or a drug" was con-       unchanged.But at any given sizeof gap the
sideredby the police (according to the official     drivers, after taking alcohol, were morc con-
report) to be a contributing factor in only         fident ofsuccess.Thus thcy wcre prepared to
994 of thcse accidents-that is. Iess than           drive through nafrower gaps.
0.5'1. In 1954and 1955the proportion was               Alcohol intensitiedany driver's tcndency
approximatcly thc same..      The true propor-      to overratehis ability in relation to his per.
tion, accordirrg to various estimates and           formance.

                                          APPENDIX 1
                             Mlrcrr.rs oF HAZARD (STANDARD DEvlATloNs)
                             NARR()WESI   CAP          NARROWEST CAP            DIFFFRENCE
       OUANTTTYoF            WHIT]H DRIV[R$           AT WHIC.H DRIVERS         (ltlncrN or
         WHISXY               AI'TEMPTED             ALWAYS SU( CEEDEN            HAZART,)
              (r)                  (?)                      (l)                     (4)
  Brlt, fi. oz.       ml.    in.          tnr.        in.         (m.         in.          cm.
        0              0     ).)          14.0         1.8         4.6        5.4          13.7
        2             56.8                13.5         1.4         3.6        5.0          12.7
        6           r 70.4   6,0          1{ ?         1.7         4.3        5.8          14.7
                              Mrn(irrus OF SAFETY(STAHDAftD Drvtlrrous)
                             NARROWE,ST  OAP      NARROWEST {jAP AT WHICH       DIFFERENCE
       QUANTITY OF           wHlcH DRrvF-RS        DRlvIRs belicv(,.{ THEY      (MARCTH 0f
         WHISKY                ATTEMPTED          WOULD AI.WAYS SUC(]EED          SAFI I Y)
            (l)                    (2)                       (3)                    (.41
  Brit, fl, uz.      nl,     in,          cmr         in.          cm.                     cm,
        0             0      5,5          r4.0        6,3          16.0       4.4          I l.z
        2            56.8    5.3          | 3.5       7.9          20.0       4.6          I 1.7
        6           170.4    6.0          15.2        5.6          t4.2       5.4          13,7
358                                                     PSYCHOLOGICAL APPROACHES

                                                                    statusof the drivers
    Theseresultsare impressive view of the professionaldriving
usedas subjects,                               of alcohol to bc cvcn mofe Pronounced
                   one niight expeclthe ell'ects
                       dri"vers, would be most interestingto see_this
                                it                                     study rcplicatecl
up"" f*,, ei.perienced
with suchlessexperienced    drivers and with a samplelarge enough" perrrritcompirri-
                                                      personality, drinking.habits,  and
sonsaccordingto suchother variables sex,ag;,
                                                        replicate thc experiment   tlsing
social charactiristics. lt would also be desirableto
                decision-makingand risk-taking, such as passing_a     car?ncgotiating a
other tests o[
                                                               might also bc clevised  in
diflicult tuln, or avoiding a pelestrian. Relatctl cxpcrirncnts
relation to hazards of o"thertypes--for example;      those encountcrcd in rclirtion to
                                                               behavior' silloking' and
i*rin.rr, gambling. i,tt*tp*tt#ul relationships,aggrcssive
infcctious agents.

                                    Ph'D'' Hazel A' Long' B'Sc'
  -G. C. Drew,M.A-, W. P' Cottluhottn,

                                                                                      of alcohol
     The next selection demonstratesthe importance of studying the.efflects
 and other pharmacologically        activc agenG on inc{ivicluals varicd P$ychologica
                                                        data with re$pect ccrtain Psycho'
 make-up. This study is r]nusuolin that i-tpresents
 losical variablesin association
 '"-j;'il;';*dv-iir.                with performanceas measured a driving sirnulator'
                          .n"", or ut"ottot on a skill resemblingdriving was
                                                                 taken  in the trairring of the
 unJer highly controlted l^boratory conclitions.The care
 *--y,*l**"ri   if subjects, r'easuringand controlling thc blood alcohol
                                                         critcria is essential such experi-
 and in obscrvingthe effectsthrougl stanclard_izcd
                                                                Goldberg, in Chap 3)' For
 rnental sturlics(scc Loomis a'd *est, and Bjt:rver and
                                                                  in which.thiswas achieved'
 this reason,wehavereproduced descriptionof the ways
  Because are concerned
             we                 hertr,however,with the personalitycor.rclates changes
                                                                     scctions  dealingwith   thc
  driving bchaviorin response alcohol,we haveomittedthc
                               concentration and the results not clirectly related persona
  measurcmcnt alcohol
 ""           as          bYmlans,o.f
 iitf aircren"csLlcterrrlin"-d              "l*:::::1f.1
                                    1.b*:"'v               n'::il:Y.i:l::
                             thardrinking a cliffcrent uponthedriving
                                        hai           eft'ect
   il;;;;;;    il; ili; clear'iy                       "Extroverts" not
                        differently theteitsrrdministerecl.
                  scoring        on                               did
 habitsof  groups
                                                      glcat deal, although they were
  changceithcr theircp*"-d o. control movcll]entsa
                                                            largeincreases thc numbe
  consistcnt thcir control nlovcntcnts. thcy clidshow
             in                           but
             ,,Introverts" changedspeedsconsiderably, thorrgh it was.not possiblc to
  of €rrors.
                                                                  speedcd The errors
  distinguish  betweenthosewio sloweddown and thosc who
  for thc introvcrts .tiO noi ot*ays incrcasc,and thcir mean error scorewils significantl
  lessthan that lbr the cxtroverts'

                                                                      be known hoth of its physio-
  Srucr rsr LIRTNKTNG atcohol in one form possibleshould
  or another is so corDtnon a characteristicol' logical effccts      and of its elTects hehaviour
  the social lif'c of man in a variety of cnltures, Irr lhct, ai the prcscnt time, very little prccisc
                                                                      availablc, in spite of an ex
  it would seem impor:tant that as much as information is

                                                                               rrS8.,A porl-l
                         permission,from the Brili sh MedicalJournal,2:993,999,
        L tion of the "", ;;-ilG       fout figrrrts, and the appendixhave been omitted' J
 EFFECT OF SMALL DOSES OF ALCOHOL                                                               359

  tensive litcrature on the subject. This litera_ performance is often
                                                                                of great importance.
  ture has beenreviewed,    more or lesscomnre_ The aclditional stirlulation provided by a
  h.ensivcly, periodically sincc the eirrly pait of  ncw task ltequcntly enablcs the subject to
  this cerrtUry.                                     compensate, a short tirtre,for corrditions
     The rrroststriking f,eatureto emergc liom       which haveproducedan overallrcdLrction         in
 any such reviewis the rrarked lack ofaeree_ his efficiency,
                                                                      For this l.cason.short intcr-
  nrent [s11y6gt1 authors, amounting in riany        polated te$tsmay faiI altogether to measure
 instances to direct contradiction of ,rnt           this tleteriorationand may show normirl or
 another. This is especiallytrue of the effects even abovc norrnal
                                                                              cfliciency, Rcasonahly
 of smaller doses.pcrhaps not surprisingly, long-lastingtegts
                                                                          seem nece$sary    unlessthe
 dosesol intoxicatingstrengthhave generiliy delicicncy is
                                                                      a gross onc. Welford has
 resulted in deterior.ation in cificiency in         stressed  the inrportance,frrr thc analysisof
 almost all aspects of behaviour tested, For         skilled bchaviour, of measuring clifferent
 doscs below this levcl, however,the picture aspects of perforrrrance,
                                                                                     since rcla{ively
 is not clear,                                       gross rilr.a$urcs mrry ohscurc changesin the
    Some authors have reported deterioration way in u,hichdifibrent
                                                                               irctionsare integratcd
 in performance howcver small the tlose, in the final perftrrmance.
 some have failed to ltnd any cffecrs, whilc           The widc intlividual variations in per_
 several havc shrtwn actual improvements formance in responsc
                                                                               to alrohol remainins
 following doses of the order of one sinele when thc abovc
                                                                         factors are takcn into a..]
whisky.                                             count havebeenattributedto tcrnpcramental
    In part thesediscrepantfinclingsreflect the ditl'ercnces,
                                                                   and especiallyto rhose ditlcr-
 unrcliability of many of the results in this cnccs related to extravcfsion-.introversion.
field, but in part they arise from tl.rc corn-         f)ifferenres along this dimension have
plexity of the clTectsof alcohol.                   been noted in other contcxts, Extraverts
    For rnoderatc and intoxicatinc levels it        have bccn shown to be relativcly less con,
has bcencstablishcrl   tliat rhc cftectif alcohol cerned with [ccuracy of performance,
is related to its corrcentration in the blood- deteriorate morc
                                                                         rapidly.during continuous
and that the blooct alcohol concentrarion work, and to
                                                                    bc lesscttnsistcnt perftrrm_
obtained from a given ingcstcdcloserlerrclrds ance. It has,
                                                                        firrtherrnore, oftcn been
both on rhc body wcight ofthc rccipientand noticed, whcn giving
                                                                               depressantdrlrgs to
on the ratc at which the alcohol is absorbed psychiatricpaticnts,
                                                                             that a given dose has a
into the blor-rd.  The rate of absorption, in greatcr effect on paticnts with hysterical
turn, has been found to be dependenton the disorders thln on 4lxicty antl clepreSsivc
amount, type, and dilution of alcohol, the paticnts. A theory postrrlirting a
corltcntriof the stornach,and the drinking afiount ofcortical control for introverts
hrbits of, the individual.                          a grclrter susceptibilityof cxtravertsto de-
    Previous rcports further srrggest that, prcssant drugs duq to thc reduction
apart from a slowly dcveloped tolerance to cortical contro[ was first trut forward bv
alcohol from repeated exporiurcto it ovcr a MclJougalt ancl has recentlyheen extcndci
numbcr of ycars, kngwn as habitu.tion, by [ysenck. Experinrcnta.[                   confirmttion of
there is irlso a more rrrpid adaptation to each the greatersusceptibility        ofextraverts to the
do$econ$urlcd, so that jt has a greater ef)'ect deprcssanaction of nrnylobarbitone
                                                               t                             sodium
when the blood alcohol level is risine than has been publishcd by Shagass.
it is fallirrs.                                       Thc aims of this experiment wdrc to in-
    The nature of the task u$ed i$ also im- vestigatcthc effect of snrall dosesof alcohol
portant. Goldberg has shown that some on a complex skill, rescmhlingdriving;
   sks arc morc rcadily impaired by a given relateany changcs            found to the levelof blood
       than arc othcrs. Farniliar tasks, too. alcohol; 1osecwhctherindividualdifferenccs
         ate lcss than ones that are unlamil- in response alcohol could be explaincdin
      Thc mcthod of mcasuring change in terms of previous experienceor of temlrera-
                                                       PSYCHOI,OCICAL          APPROACHES

mental differences, when every effort had           ablc advantagesfor scoring in presentingthe
heen rrradcto minimize differcncesin blood          srrmc objective task to the subjects at all
alcohol level; and, finally, to investigatethe      stirgesof the trial,
accuracy with which blood alcohol level                The mairr a$pccts of performanae to be
could be cstimatedby rncasuring alcohol             scorcd were accul:acyof trrrcking, spet:d of
excreted in urine and in breath. This article       driving. and the control fiovelncnts of
is a preliminary rsport of the mtrin rcsults of     stcering-whecl, acceleriltor, brake, clutch,
the expcr:imcnt which was carricd out on            and geat lever. Of the latter, only steering-
behalf of thc Road lJsers' Conrmittee ap-           wheel arrd accelerator pedal rtovenrent$
pointed jointly by the Medical Research             proved worth considering. Accuracy of
Corurcil and the Road Research Board                tracking, measured in tcrms of deviations
(L)epartnrent of Scicntific and Industrial          ltom a course parirllel ttr thc lcfl-hand kerb,
 Researcl'r).Full dctails wilt be prrblished        and stccring-wheel and acceltrator pedal
later by thc Medical Research Council'              movements wcre recorded graphically, and
                                                    sirrrulttrneou$ly scored on counters,photo-
ExprRINtst*ttr\L  MertroP                           graphcd once sach lap. Separate recording
    In decidingupon the task to be used,two was made of the number of collisionsof the    "hunting"
coflsidcrationshad to bc borne in rnind: car with the sidc of the road;
that the task would nccd to be a continuous movements of the steering-wheeltoo small
                                                                                      gear changes;
onc of reasonahlylong duration, and that to changethe car's direction;
such a task rutrs     thc risk of becoming ex- identification marks ibr lap com.pletionand
trernely boring for subjects.lt was finally major corners; and a titrte rnat'k.
decided that an appalatus known as the                  The subjccts were 40 volunleers from the
"Milcs motor driving trainer" providcd the staff of thc Road Rescarch Laboratory.
                                                                                           40 ycars,
best compromise bctwcen a task, perform- Thcir ages ranged from 23 to
ance on which could       be scored adcquately, except for one subject aged 58' The mean
and one which had the motivating capacity age was 3l years. Five of the subjectswere
 of a rcirl-life sitrration.                         women. All werc in good health and hcld a
     In this apparatus the subjcct sits in a currcnt driving liccnce. The majority re-
 durnmy car, facing a trarrslucentscreenirl a portcd that they took alcohol only occirsion-
 darkened room. Behind tlre scrccn is a allY'
 "perspex" disk with a road ljccncpainted on             Thc alcohol was administeredorally as
 it. This road scene is projected, vcry much analar grade ahsolutc alcolrol, diluted to a
 magnined, on thc screen by tr lamp on the 20 fr solution and flavoured to disguise
 perspex surface. As the driver opcrirtes the alcohol content. Haggard, Grcenberg, and
 flcceleratorand steering-whcelto change his Cohcn fourrd that the toxic eft'cctof alcohol
 speed and direction of tnovetttcnt, the car varicd consideriiblyeven within the same
 appears to pr(lgrcss along the road' The kind of spirit, owing, apparently, to
 effect is reasonably reatistic, and    the task prcsence. minute quantitie$, substanccs
                                                                in                     of
                                                                                         They found
  bears sonte rescmblance clriving in that relatedto thc original distilling.
  non-drivcrs find the rnaclrine difTtcult to         that, after very carcful distilling, absolute
  cofltrol without considerable   practice,whilst alcohol was lesstoxic than any of the other
  expericnceddrivcrs have little diliiculty with      forms tried. As analargradeabsolute    alcohol
  it, Though the cenfial task is  very sinrilirr to    ,fa$ used in this experirnent.it seernsprob-
  driving. it differsfrom it in beingcornpletely able that thc resultsreported here rcpresen
  devoitl of clangcr and cmergencies.The minimrrrrr effects for thesc quantities irrrd
  track usedwas a contintlous winding circuit, concentrations. Equivalent amounts of al-
  the equivalent of slightly over one milc in cohol takcn as beer or spirits could be ex-
  length. Its rcpetitive nature was not very pected to have somewhat grcater sll'ccts.
  app'flrenl to the subjtcts, and had consider-           This investisaticrn wrrs conc.srncd with
 EFFECT OF SMALL DOSES OF ALCOHOL                                                            361

  blood alcohol concentrations of less than            for 20 minutes and then had a lO-minute
  100 mg. pcr I00 rnl. oi blood, since tlris is break, during which the blood. urine. and
  the figure recLrmmended the National breath samples werc collected. Thcre
                                  by                                                           fol-
  Safety Counc:ilof America as the limit of           lowcd three further 20-minute periods of
  "safc" and
                only "possibly under the influ- driving, and l0-minute resr pausesduring
  ence." To achieve thesc concentrations* which samples wcre takerr. Each
  doses were given of 0.00 (placebo), 0.20, mental serieslasted 2| hours.
  0.35. 0,50, and 0.65 g. of alcohol pcr kg. of          Approximately0.5 ml. of blood was taken
  body weight. In terms of thc concentrations from the thurnb on each samplingoccasion,
  usedby Cohen, Derrrnalcy,       and Hanselthese and the blood and urine sarnples were
  doses representapproxitnatety 18, 31, 44, analysed at thc South-Western
 and 57 ml. of absolute     alcolrolfor an I l-stone ScienceLthorntory, at Bristol, by thc micro-
 (70-kg.) man, Thc largestdosc is the approx-         analytic modified Cavcff 111915.rd    rccom-
 imate equi.',,alenl three pints (1,700ml") of
                     of                               mendedby the B.M.A, Corlrnittee.Samples
              beer or 5 fl. <tz.(142 ml.) of whiskv   takctr on the alcohol-freedays were usually
 for an ll-srone (70-kg.) man.                        disctrded. but occasion.rlly wcrs used as a
    In vicw of the expected wide individual           check on the analytic procedure.Rcadings
 variations, and of the diftculty of rlefining a      from three instrumcnts for the measurement
 "correct" perftrrmance
                               on such a task as of breath alcohol were recordcdat the time
 this, it wns dec:icled use each subject as his of blood sanrpling a numbcr
                        to                                               on           of occasions.
 own control. That is, the efflcct of alcohol        The breath analysisinstrumcnts use<lwere
 was measuredas the degrcc of change in              the "alcometer," the "drunkometer," and
 performancc of cach subject against his own         the "brcathalyser,"
 perfttrrrr:rncc without alcohol. T6 grinimize           The subjects were given a battery of
 practice clfects, a latin square design was personality
                                                                   t€sts, from which only the
 used, each squarc containing five subjccts measuresof extraversion
                                                                                are considere<lin
 and five doses. This sgu€trewas repeated this paper.Testsincludedthe M
                                                'fhe                                 inncsotaM ul-
 eight times with dift'ercnt subjects,               tiphasicP*rsonalityInventory,the Maudsley
 women subjects    wcrc assigned one square, Persnnality Inventory,
                                     to                                       and the Bc-rnreuter
 Each subject was tcsted on the same diry of         Persorrality Inventory,
 the wcck for five consecutive wccks. Thus
 each subject received every dose.
    Suhjects werc given Freliminary practice Ixuvrnuu             Dlrnsnerqces rN
 to familiarize them with thc ta$k, ancl in-         Rsspor'JsE ALColoL
 formation was obtained ol body weight,                 Although the latin square design used
 age,driving experience,      and drinking habits. makes compff.risonof the performances of
 (They wereaskcdnot to rlrink on the evening indiviciualsdiflicult, certain fciltures
 beforc a trial.) On thc morning of thc trial thclcsscmcrge,There is undoubtcdlyir con-
thc subjcct took a fat-free breakfast and ? to       siderablc rangc of individual variation in
2{ hours larcr, at I0 a.m., the first urine response alcohol. This is summarizcd
                                                               to                               in
sarnple was collected to provide a chcck on Tablc Il, which givesthe numLrer subjects   of
kctones and residual alcoho( and io emntv changing their score after alcohol by varying
the bladdcr. (ln a subsidiary c*perin,*ni, amounts. The changcsarc expressedas per-
urine sanrples   were takcn with a full bladder,) centagcchange llorn their own no-alcohol
He was then given his drink antj requestedto         SCOre. Tracking errof shows a wide scattef.
finish it within I0 rninures.    Alter a iurrher l0  Though the distribution is fairly hcavily
minutes he cntered the apparatus and wag skewed towards increascderrpr after alcqr,
instruc,ted drive as he norrnallv would in
             to                                      hol, l0 of the 40 subjects show reduccd
l real car, and not to stop until toid to dtr so, error. one as much as 50|i, reduction,
unle$san enrergency     arose.Thc subjectdrove subject, however. showed also one of thc
362                                                                    PSYI]HOLOGICAL APPROACHES

                                                                FEF.CENfAGt CHANGE lU SCORESArren                    Arcorrol
TABLE ll.-NuMeen            or SunEcts     SHowrNo Drrrrntut
                                                        (tlosE 4'}

            MEASURE                                                   CHANGEOF SCORE
                                           Deiled$                                 Increase
                              -50    -40     -30       -20       -10   +10       +20 +30 +40 +50 +60                       170

Ertor                           1     0      2       2       5     1   0     1        2       4       2    2     '     -
Tinrc                                            4
                                                                                                  -                        l
largest reductions in speed, and achieved a ity Inventory wcre well distributed. The
high degrceof trccurircyby driving extrcrrrcly mean wss 10.24 (sigrna 4.5.1)as compated
slowly. Timc scores,on the othcr hand, are with thc standarcl         norm of 10.94 (sigrna4.74).
                                            "no Perl'orrnrrtrccscofes werc compared fbr a
distributecllairly normally around the
change" position.                                  group ol sulrjccts scorirrg higl'r in extra-
    An attcnrpt has bccn madc to relntc these version and a group scoring low. Fig' 5
individual differenccs responsc alcohol shows the nrean dill'crcncesin tracking per-
                         in           to
to varir)Lrspersonirl charactcristics. No re- lbr:mirncebetwcen thc tesulting scven extra-
 Iatiortship was found betwccn response to verts afld nine itrtrovcrts. The extravcrt
 alcohol and dill'crcncesin initial lcvcl of group fiade more errot (p<0.05), were le$s
 skill, plevious ilriving expcLiencc,   agc, sex! cr)rrsistEnt car posil,ioning
                                                              iu                  (p<0.05), and
 or drinking habits. The subjccts varied showed a biggcr irtcrcascin error during
 widely in their initial lcvcl trf skill and in each pcriod ol'driving (p <0'01). Thc cffect
 previous driving expericnce, thc lack of of alcohol on etroi wils gretrter litr the cxtra-
 rclatiorrship bctweerr  thesc variable-sand al- vert group (p<0.05), thc extravertrihaving
 cohol elTectrnay   be statcd with sotrrcconfi- afl irveragc increase of 231i, and the intro-
  clence. The range of drinking trabits,how' verts of 670.
 ever, was srlall, stl that the cttnrparison was     Thc cxtravertsrespondedto nlcohol in a
  hetwcenthe few indivicluals   who drank once sirnilar manncr, but the introverts showed
  of twicc a week Lrndttlose who drtrrrk very mote variccl respoflses. For cxample, the
  rarely. Thc range wirs thereltrrc too sfllall to cxtrrrverts all incrcasecltheir tracking error
  dravvconclusions the lack ofrelationship. alter alcohol, whereas two of the introverts
    Scorcs obtaincd, for 35 subjects,on the made less crror. This is especially clearly
  extraversion scalcof the MaudsleyPcrsonal- shown in thc cfl'cct of alcohol ou speed of

                            EXTRAVERT NTROTCO
    +                ---6   I N T R O V E RD o S E d


    z        30

              oaot                                   40   50    60    70    80                            too   tl
                                                      MINUTES AFTER INGE$TION

                                                                                   under no
             Fro. 5. Comparison of the tracking error of extraverts and introvcrts
                     alcohot and dose4 conditions'Extruvcrts n:7; introverts t?- 9'
 EFFECT OF SMALL DOSES OF ALCOHOL                                                              363

driving, The extraverts showed very little            people who all had blood alcohol concentra-
changcof speedafter alcohol, but the intro_           tions between I00 and 150 nrg. per l()0 mt.
verts changed considerably. The ncarer the            Morcover, the rangc of blood alcohol con-
introvert end of rhc scale thcy were, rhc             centrationsat .i,vhich   peoplc are ju<lgctJ in_
more speerl changcd,They subdividecj,   how_          toxiciitctl is vcry wide, Coldtterg hns shown
ever^rnto two distinct groups, rine of whjch          that the level at which 50f of pcople are
speeded very considerably,
         up                   arrd thc other         judged clinicallyto bc intoxicatecl   varieswith
sloweddown, lt has not so fnr beenpossihJe            the legal definition in differentcotrntries.  In
to relatelhis to any pcrsonalcharacierjstics.        America it has beendefinedby thc National
Fig. 6 shows the regression of change of             Saf'ctyCrruncil as I50 mg. per 100 ml., in
speedon extraversion..introversion meas-
                                    as               Swedenir is 100-l?0rllg. per 100 ml., and in
ured by tlre BcrnreuterNcurotic Invcntorv.           I)ennrark and Norway 80 mg. per 100mt,
                                                         The level of alcohol in the blood at which
    i                                                people are diagnosedclinir:allyto be intoxi-
                                                     catcd, besidcsbeing variablc, tends to be
                                                     high in conrparisonwith the conccntrirtion$
                                                     used herc. This is not surprising, since an
    {                                                cstirnateof inrpaired behaviour is madc in
                                                     the absenceof any criterion of norrnal be-
                                                     haviour for that individual, stl that the im-
    {                       Y : 3 . 8 * 0 . 1 4x
                                               6     pairment must he obvious bcfore it is de,

    =                                                tected,and sinceit is known that peoplecan
                                                     Compensate their rcducedelliciencyover
    z                                               the short Fcriorisof time during which they
           EXTIAVERT        TNTNOVFRT               are examined. It docs not follow, lrowever.
                     HEU     SC                     that. becauseno impairmenr is found in
        Frc. 6^ Regression of change of             clinical tests, rn()rc complcx skills, like
        specd after rrlcohol on extrlvert-          driving, will also bc unaffected.
        introvcrt score, Each point plot-               Thc prcsentstudy showsthat performance
           ted is mcan of five subiccts.            tregins to dcrcriorrrte with vcry low blood
                                                    alcohol concentri{tiorls,ccrtuinly of the
The regressionis linear and is significant at       order of 20-30 rng. per, 100 ml,, and that the
the 51{;lcvel. Coruparirrgthosc scoring more        dcterioration is progrcssive     and linearly re-
than 50 on the Bernreuter   with thosescorinu       lated to hlood alcohol level.There is, in this
lcssthan this in rermsof whcthcr they changi        study, no evidenceof a thresholdeffect.
speed hy rnore or lcsstlran I units give56 X2           The impairment o[ perforr.rrancc         was
significant at berter than 0.01. lt would          shown most clearly in the operation of the
appcar that personality characteristics-at         controls,As the aim of the task wrrsprimari-
least (hose o{' the extravert-introvert dimen-     ly to track, it is perhapsnot surprisingthat
sion-arc rclatcd to the effectsof alcohol.         in iittcmpting to retirin a levcl of accuracy
                                                   the operationofthe steering-wheel       shouldbe
                                                   most offected.     Eflicicncy usingthe steering-
   It is now generallyagrcedthat a rclation-       wheel is indicatedby the amount of tracking
ship existshctwcentlre conccntrationof al-         error which results.    With practiccthe amount
cohol in thc blood and thc appearanceof            of stcering-wheel     movcmentmadc goesdown
clinical signs oI intoxication.The definition      wrthout a redr.rction the accuracyof ftack-
of intoxication bascdon clinical evidence  is,     ing, whereasafter alcohol traching error in-
howcvcr, a variablc onc. Liljestrand gives         creasesdcspite an ilrcreasein steering-wheel
evidence of ir considcrable variation in           movement.This suggests         that timing of the
diagnoses madc by different clinicitrnson          stecrinll-whcsl rnovcnrcnts was unsct, The
364                                                      PSYCHOLOGICAL APPROACHES

decreasein consistency ol' the movements with blood alcohol levels of 60 to 80 mg.
required to ucgotiatea corner supportsthis per 100 ml.
view. Thc importance ol timing of control              ln vicw of the known difficulty of clinical
movemcllts in a complex skill           has been estimatesof impairmcnt in such situations,
                                                                                                he is
stressedpreviously by Bartlett, afld adverse and of the individual's capacity, if
                                                                                    is             to
conditions such as fhtigue have heerrshown capableof realizingthat he under test,
to have a nrarked effcct      on timing.            compensateternpomrily for lossof efficiencv,
                                                                              alcohol levcl would
     Thc work of Cohen trl a/. is interesting estirnation of blood
additional cvidence.While the task uscd         in seern the rnost direct way of assessingim-
                                                                                    alcohol excreted
this present study rcasonably reflects the pairment. Estirrrationof the
                                                                                rrllows a close ap-
tracking aspectof driving, it is almost com- in urine and in breath
 plctely free ol hazardand lisk{aking. These proxirnation to blood rrlcohol. In these cx-
                                                                                agrced very closely
 authors have shown that an arnount of periments, urine alcohol
 alcohol not tnuch larger than thc biggest          with blood irlcol'rol. Breath analysis, whiclr
 dosc uscdin this experimentproduccda sig' ha$ enorrrrous administrativc
                                                                                       analysis, can,
 nificant increase in the l'razards in which over both blood and urine
                                                                           given a suitable instrtl-
 drivers became involvcd. From the two frorn these results,
 stuilies it is rcasonable    to as$umcthtt not ment, rrlso givc a very cl()seitpproxilnation'
 only will drivers become involved in greater           It would seem fairly clear, frotn these
 .hazrrrds  with this amount of alctlhol but they    results, that thc efficiency with which a task
 will be lessefficicntin dealing with them'          like driving is performed is likely to decr:ease
                                                                                        rises.At some
      It is suggestcd, furthcrtnore, that the progrcssivelyas hlood alcohol
  resultsreported herc     show that thc lcvcl of point thc loss ofeificierrcyis likely to be latgc
                                                                             a danger in a practical
  alcohol in the blood is a good indicator of enough to constitute
  thc extelrt of .itnptrirmentof performance' It     driving situLrtion,As this experiment has
  may be pointcd out that the blood alcohol becn carried out in a laboratory, necessarily
                                                                                       of thc motivat-
  levels reportcd by Cohen el a/. seem sur- frcc of the hazards and many
  prisingly low. Assuming their drivers          to ing fcatures of a real-lifb situation, it is not
                                                                                       to say at what
  weigh rather more than 1l stone(70 kg')' on possiblc, from these results,
  averirge, their {wo doses corrcspond fairty levcl of blood          alcohol the increasedrisk of
                            and higgest  usedhere' accidentwould bccome unacceptahle.            Some
  closelyto the stnallcst
   Their blood    samples,however,would appear evidence on this is available from the ex-
                                                                                 I)enmark, Norway,
   to have been taken rtlmost two 1'ou1s4ftet perienceof thc LJ'S'A.,
   subjectshegan to drink'       If this is so, the and Swcden, wher-erather diflcrcnt cut-off
   blood alcohol levelscolrespotrd vcry closcly Ievels are uscd.
   to thosc fountl here. They report a con-              Individual differences in response to al-
   centration   of 4 rng. irnd 5ll rng. per 100 ml' cohol appear to be related, at least in part,
   for thc two doses.At two hours comparable to personality characteristics,
   figures in this experiment      were 6 mg' and those of extravefsiofl--introvefsion.Eysenck
    62 mg, Pcr 100 ml'                                 has collected the cvidence lor the existence
        While this in no way affectsthc conclusions of this dimension of personality and has
                                                                                      woulcl be more
    reache<lby these authors about the effect of po$tulated that extraverts
    such doses on performance, it does affect susccptibleto the effectsof deprcssant
                             "sa[e" levelsof blood like alcohol. Thc cxtravertsin this experi-
    their discussion   abQut
                                                                       as a group, and tend to con-
    alcohol. At the time of the tcsts the blood mcnt behaved
                                                                       hypothesis,Thc introverts do
    alcohol levelsin their drivers would altnost firm Eysenck's
    certainly bc at least   10'20 mg' pcr 100 ml' not behaveas a group. It is of interestthat
                                                                      used agreed on the extravert
     higher than those reported, Even so, thcir all the scalcs
     exoeriments, and thcsc, $uggest      a marked end but tlid not agree on thc introvcrts. It is
     impairment tlf perfornrarrce on such tasks possible        that the almost bimodal effectnotcd
     EFFECT OF SMALL DOSES OF ALCOT{OL                                                                  365

     onchange ofspeed forintroverts reflectsthe        were usecl.Best results were obtained from
     intrusion of somc other personality charae-       thc "breathalyser." The others either showe<l
     teristic not testcd in this experiment,The       high constanterrors or were unrcliable.The
     test which showed this bimodal ctlcct most       results from the breathatyser werc good
     clearly was the Bernreuter, which uses a         enorrgh to wgrrant its consideration from a
     more "neurotic" criterion of introveriion        practical point of vicw,
     than the Maudslcy.                                     Mcan error showed an increase with in_
        In this cxperimentextravcrtsappear not                     in
                                                      crease blood alcohol, amounting to about
     to be hothercd by the extrir strcssimposed by     l6 L deterioration with a blood alcohol
     alcohol. They drive at much the samespecd        concentrationof somc 80 nrg./lQ0 ml. part
     as before, make vcry little additional cor-      at least of the increasein errr)r is to be
    rective movenrct]ts,but make tluch Ereater        explained by a significant tendency for sub-
    error, ars less consistent,and deteriorate       jects to movc towards the riglrt-handside of
    more rapidly during 2O-rninute driving            the road after alcohol anti also by lesscon-
    periods. Introverts, on the other hand,           sistentpositioning.E.rrorscoresand control
    appear to be striving to compensatelor the        moverncnt scoresalso showed a variation in
    alcohol effect,and to be anxious to demon-       time, rising and falling in a way similar to
    $tratc their efficiency. They ovcr-rcact to      that of blood alcohol.
    the situation, mrrke nrore corrective move-           Meanspeedshowednosignificantchang€,
    ments of the stecring-wheel,       ancl change   but marked indiviclual differences spced               r1
    their speednrarkcdly. some slow right trown,     aflter alcohol were found.
    presumablyin an atteurpt to achievc accu-             Control movements, as measured by
    racy, though they do not necessarily so, do      steering-wheel          movcrnent,       showedsignificant
    while others appear to be attempting to          increasisand a significantreduction in con-
    demonstratc how quickly they can drive,          sistcncy.
    again not always with a proportionate loss            Age, scx, previor.rs         driving experience,and
    of efficiency.                                   prcvious drinking habits, within the lirnits
                                                     availablc,showcd no relation to individual
I S,r""o*" -
l--.-                                                d i f l e r e r r c eis r c s p o n s e o a l c o h o l .
                                                                           n                t

I'"*i'ixff'lT--":*::l :?a definiterelationto behaviourchanges.
  alcoholon a complex
     Four ul.ol:t dosesand a placebodose             Extravertsdid not chengecithcr speedor
              The peak blood alcohol con-            control movcments    'ery much,thorighthey
I*11. :.**0.
                   the doseswere approxi-            werelessc'nsistent controlnrovenrcnti,
lmately20,40,60.and80mg.    l00ml.of blood.          but showed    largeincreases error. lntro-
          and.breath analyscswerecompared            vertschangecl   specd considerably,thoughit
| .YriT:
lwith direct blood irnalysis. Urinc alcohol          isnotpossibletodifferentiatgbu-twccnthose
          tatel            than blood, but           whoslowecl   clown those
                                                                       and       whospeerJerJup.
lRcakcd         1n^d..hieher
lbreath alcoholfollowedblood alcohol in              Controlmovemcnts incrcascd, were
                                                                          also           but
       The ratio of urine alcohol to blood           relirtivelycon$i$tent.Error rnay or may not
          was 1.252 l.
                   ;                                 increase, the mean
                                                               but           crrorscorewassigni-
    Three kinds of breath-testing apparatus          ficantly lcssthan that for extraverts.
     th:.*p"rimental-method used in this study raisesseveraltroublesomequestions
| ,
         thc transfer of findings from the laboratory to a natural setting. Lrrboratory
           and thoseparticipating field experimc;tsare more Iikely to be motivated
    drive as well aii they can, evenunder the influenccof'alcohol,than their counter-
       ln a naturirlsitua(ion,conversely,  thcy also may not be threatened the same
366                                              PSYCHOLOGIC]ALA PPITOACHES

extent by the consequences failure. In addition, the drinking driver is likely to
differ in a numbcr of respccts    from the laboratorysubjcctwho takesalcoholils part
of an                In
       experinrcnt. sorne                                                   and
                              cases formcr'sdlinking maybe exccssive associatcd
with psychological,     social,and other characteristics    which set him asidefrom other
driveis.Driving in a naturalsctting,     furthermorc,mfty haveboth a diflcrcntgoal and
a differc:nt situitional complex,anil rcal automobiles aPt to vary much more than
experimcntaldevicesin both designancl stateof rcPair' Theseproblens offer serious
challcngesin thc clevelopilcltt and use of driving simulators ancl othcr lirborirtory
    Furthermore, the number of cases       (40), typical of suchexperiments, was too small
to pcrmit the kincl of control     colnParisollson such background factors as sex' ilge,
driving expericnce,     and clrirrking labits that are necessary expand upon such
experiirenial findings. Sinrilarly,lhe sclectionof l"hesc     subjccts(volunteersfrom thc
striff of the ltrbtlrntor]) probably madc for el nafrower rangc of lrackgrourldcharflc-
teristicsand-=.more      inrportant-personalitydillcrcnccsthan those which woulcl bc
found in an actual clriving populr,tiott. Thise shortcomingsdo not detract from l'hc
value of this conl"rolledlJboratory experin:Ient, they limit the extent to which its
 resultscan be extrapolated    with  conficlence nfltural situationsof greatel:
                                                  to                            complexity
 and divcrsity

   -Rit:ltartl M. Mirhaels

    We have repeatedly                                                     in
                        emphasized importanceof viewingaccidcnts their broad
 ecologicconte;t and of avoidingdiscipliriary  palochialisnt  fllld.an_overemphirsis  on
 narro; classesof  variables.Although ulost of lhe currcnt work ilr the field falls short
 ofthis slandard,at leastpartially bioaclcr approaches increa$inglyreprcscnted,
                                                         ale                           a
 developrnent with parallels many collateral
                            in                fields.In highwaysaf'ety,    one of the ex-
 pressidns this bioaderapproach seen the application tnan-machine
           of                       is     in                  of                systems
 concepts. Theseconccptsliigely grew out of problems    in the interrclationships men
 and machines, first widely."iognTr*ain Worlcl  War lI,+ which wereoften found to be
 at leastpartially solublethroulh psychologicirl rcscarch.  Ilowever, despiteits stated
 concel'nwith htoader issues, much of the tesultirrgliterature,though of great interest,
 continucswith somcwhat dill'ercntsubjectmattcr the narrow emphasisof most
 earlier psychologically oriented accidcnt research.   This is unfortunatc, since the
 systems conceptimpliis thc ic{entificationand unbiasedcorrsideratiou all pcrtinent
 variables, whateverclass.
     Somc of this litcrature,for example, althoughit emphasizes grcat variability
 of the tasksto which drivers   must respond,tends to ovcrlook the great variability
 both in the capacity thc samedriversand pedestrians differenttimesand among
                      of                                at
 individuals usingthe roads.(This may hilve lcsultcd from the predominantearlief
                                                            which are far more homo-
 concernof workirs in this ficld with military populatiol'ls,
 geneous  than thosewith which they are now faced.It may alsg stetnsubstantially
 irom cngineers' unfamiliaritywith the remarkable  variabilityof biologicalmaterial.)
 Vehiculir failuresoften ilrc unmentioned clismissed ullirlrportantdespitethe
                                             or           as
     SYSTEMS RESEARCH IN SAFETY                                                               367

      lack of adequateresearchas to their prevalence,pnrticularly in rclation
                                                                                          to serious
      collisions. Inadequacic,s veh,icrc
                                 of             disign, though carefully consicrcred some  by
      authors(see   McFarlancl, chap, 2), arc ncglected orhcrs.vedicat variables,
                                                           by                                suchas
      vascular  accidents, also often ignorcddespite
                           are                              iheir pertirrence McCarroll and
      Haddon, chap. 3), as are suchparamountcharacteristics the systemof              as the crash-
     injury design the vchicles
                     of              involvedand thc influence post-injurymedical
                                                                   of                        careon
     the human and financialcostsof the collisionsthat do oicur..
          Morc seriousquestion's      concerflingthis literaturc irri$ewhen its prcmiscsand
     conclusions considcred
                    are             in,the light of the adecluacy thc supportingevidence.
     consider, frrr example,   the folrowingstatemcnt      from tiris selcction:  ..I.hii
     considers   safetyas one aspect the performance a man-machine
                                        ofl                   of                   system. this
     context, accidentsarc viewed as random events with no speci[iccause
                                                                                         or sets of
     causes  lhat are generalizable the system a whole. Hence,accidents
                                      to              as                              arisesimrrly
     as ti contingent   consequence system
                                      of          performance of the virriatrle
                                                                or                  error inherenl
     tn rts operirtron. - . collisions appro*ch being perfectryrandom with finite proba-
     bilitics," or' as expressed an eirlier paper:';. . . it opp"o.* reasonable
                                 in                                                    tu conc-lude
     that a largeproportion of accidents      traveno cause.  Thrit'is,they arc truly accidents   in
     the strict dictionarydclinition,an eventnot specifically       causej or predictahle."rr
         Aside from the fhct that this nright be interprctei as an cxampre of the
     rationalismthat was ciiscussed Chapter l, it ii worth noting thai what ir
                                         in                                                 alrearly
     known scicntifically   abour rhc distribuiionof highwayaccidenis the populationat
    risk arguesstrongly.against     this premiseand thJrefoie againstthe conclusionsba$cd
    upon it. For example,     Mccarroll and Haddon (chap. 3j found rhar rhe maiority of
    accident-responsible    drivers had characteristics ftrund among an apprLrpriaiely
    chosencontrol group of clrivers        similarlyexposcd risk. If thcsf .workers
                                                              to                           had also
    conccrnedthcmselves      with the difl'ercntialcharactcristics the vchiclcsancl of the
    environments which the accidents
                     in                        occurred,it ir Iikely that evenlargerdift'erences
    would have beenfound
         f)cviations from randorn expectancy similarly largc magnit,ude
                                                   of                              have also been
    found for fatal pedestrianacc:iclents Hrrddon ,,r ol_ Cf,rip. d.;. Earlier work
                                               (sle                                              by
             ana Popham,laLucas et ill.,rs Bjcrvel er 4!.,rt o,lj others, reviewed by
    Haddon,r2and the reqent,excellently          controllect work of Borkenstcin al.rr also
    s.lrolFly.suppDrt the conclusion that highway accidents are far from random in
    distribution nnd that discretescts of cauial factors rhat can be generalizecl sub-      to
    stantialportionsof the entiresysteftcan be identifiedby uppropr?ately            sophisticated
    scientificmethods.In view of theseconsiderations, is unforiunor* that Michaels
    sctentttic methods. In view of these considerations, it is unforiunate
                                                                                    that Michaels
  emRhasil|s the assumption of randomness since it does not appear
I                                                                  necessary in
              acciclcnts from the systems point of view.

       nrvlew of thc highway safety move-          factors or characteristicsof thc driver, the
        indicates that thcre have been two         vehicle,nncl the highwny that opcratc incle-
         diffcrent aPProachcs the problem
                              to                   pendently or in linear t:ombinationro gen-
lrather                                                            'l'his
    safetl. Onc assutnesthat thcre are specific    srirteaccidents,       is very rnuch the domi.

I        lnr*t*n,ed at a meeting of the American PsychologicalAsrociation, Philadelphia]l
         L September   1963,
                           The author is affiliatedwith ths tJ,S.Bureauof Public Roads. J
                                                        PSYCHOLOGICAL            APPROACHES

                                                                    used to compensate for inac-
nant frame ofreference in the field, and has that may be
been the motivation for rnuch of thc very             curaciesin judgment. This too increascs    the
small research effort       in safety' lt tends to reliability of thc systemand hcnce itrcreases
direct research into intensive analysis of its safety.
acciclents,  examination of various factors in          The two approaches to safety are jrot! of
the cornporrents      of  the system in order to coursc, mutually exclusive. A significarrt
                                                                               to indicatethat indi-
determine their causal connection to acci- body ofevidenceexists
dents, and to individual characteristics         of vidual factors do exist to generateaccidctrts.
                                                                          appear that thc r:cliability
humans as they may be linked to accidents. However, it does
                                                                                in the design of the
    Thc second approach to safety has been and efticiency inherent
sonrcwhat     more subtle,and has bccn      implicit highway transport system are the major
                                                                     of its safety. In this context'
in much of thc cngineering donc by the detcrminants
automotive industry and the highway de-               safety becomespredominantly ;rn cngineer-
                                                            problem and secondarily one of indi-
 signer.This approach considcrsriafetyas one ing
                                                                      social deviancy, or personality
 aspcct of thc perlbrmance of ir man-machine vidual failure,
 $ystoill, lrr this context, ac$identsare    vicwed' dcfects.
 as randon events with no specific causc or               This systems engineerirrg approach to
 sets ofl causesthat are generalizableto the           highway safety radically changes the direc-
                                                                           highway safety. Accident
 systern a$ a whole. Hence, accidents arise tion of researchin
 simply as       a contingent consequcnce of analysisbecomessubordinate to the
 system performance or of the variable error           of the functional perlormance of a mHn-
                                                                           From the psychological
 inherent in its operation. The problem is maclrine systcl'n.
 very    much like that of gas molecules in a standpoint this requires research into
 confined space, Collisions approach being             nature of the driving Processli'e., what the
  perfectly random with finitc probabilities' humirn is requirec{ to do in opertting this
                                                                                  imposcd upon hirn
  The probability sets may be defincd and systcmand the demands
  changed, c.g., by clranging temperature'             by the design of the system.From this liame
  pres'$urc,or volume, but any specihc colli- of refcrence, highway s;rfcty rcsearch
  sions between any pair of molecules remain           quircs not the study of acciclentsor isolated
                    "uncaused" in the statistical physical or psychological traits, but rather
  random antl
                                                                                 of thc driving task.
  sense. To state the safety problem in this thc intensivc analysis
  fashion does not mean that accidents        cannot Accident analysis,at best, may provide some
                                                                 priority for such driving research-
  be reducecl.Rathcr it metrns that to rrake order of
  large reductions in thcir occttrrcnce will               This approach to safety cleLrrlyexpresses
  require iiy$tenratic changes in the basic the biascs of an engineering
  design irnd operation of the system.                  His oricntation is to carry out an analysisof
                                                                                        are uscd. By
      In these terms, the saiety problcm for the sy$teDrs in which humans
  engiuccr becomesotrc of       developingmethods quantitatively defining the performtncc of
                                                                                     its objectives,hc
   of modifying systerl dcsign so that this thc system,and knowing
  variable error in its performance is reduced.         can then specify whether the human can be
   For example.the controlled access         highway used and. if so. httw,
   is a basicdesignchangeaimed at elirlinating             Considerable time has been expended in
   a whole tgliL-s   6f interactions that reducc the defining the approachesto safety, lbr much
                                                                                             as safety
   uniformity and predictabilityof trilllic llow. of what flollows is rneaningful
   In so <loing there has bcen achieved the             rescarch only if these assumptionsare made
   greatest accidcnt rate reduction the safety explicit. Also, employing the systems
   held has sver known. Similarly, the auto-            proach requires that the rules for defining
                                                                                          well as must
   motive engineer by significantly improving highway safety be changcd, as
    the weight-horsepower ratio in the modern be the rcscarch requircments in the
    vehicte has incretrsedits responsiveness       to    field.
    loads and givcn the driver a reserveofpower             Taking the path of analysis of the driving
SYSTEMS RESEARCH            IN SAFETY                                                        369

ta$k, safety r€search breaks down into a             bility with which they can control the systcm.
fairly obvious set <rfproblenrs.  They may be        And it is this which definesits safcty. lt is
statcd ln a seriesof four questions:                 in this contcxt that certrrinof this rescarch
   L What dimensionsof the cnvironment               m:ry bc rcvicwcd.
does the driver use in what hicrarchy to                Onc consistcntproblem for a driver is to
Iocatc himself in tinre and space?                   locate fixed objectsin his ficld of movcmcnt.
   2. How does the clriver ordcr and scale           If he is to control steering rclative to such
this information?                                    ohjects on or near the roadway he must have
   3. Ho v doesthe driver translatethis ana-         solne way to r'lctcrmine  their p<rsitionrcla-
lysis into control changesin the vehicle?            tive to his path of travel. The problcnr is
   4. What arc thc transler functions from           shown in ligure I. Frorl a pcrccptual
driver inpr.rtto the controls to thc system          standpoint thcre are several ways that a
output response?                                     driver can carry out this localizingproccss,
   Thc first of these questions is clearly a         In examining thc geometryof the sitrration,
perceptual problcm and thc one which we              an intcresting ljrct cmcrges: as a driver
in the Human FactorsGroup of the Bureau              approachcs an Qbject not direcrly in tlris
of Public Rorrdshavc been emphasizing.      It       path, that object has a component of move-
appcars to us tha( this is a critical area           rnent away from the drivcr's line of sight.
becausein the rclatively unstructured en-            Thus, all suchobjcctsovcrtakenhy a moving
vironment of the driving world, the driver           observer have a compoilent of angular
is free to referencchirnselfin tirnc and snace       velocity which is tangential to the path of
in severalways. However, what driveri do             travel.
use will ultimately determinemuch of u,hat              Such a model was testedon a test track,
is observedin driving apd, rnore imp'rta't,          in which the latcral position of the test
what will be thc ultimate accuracyanrl relia-        vehicleu,asnreasurcd   contiduouslv.Analvsis



                               A+i'+f                                           trli#
      F I G . I - E O M E T R O F T H E L A T E R A LD I S P L A C E M E N T
                G             Y                                          EFFECT
                   Frc. l. Ceornetry of the lntcral displacsnrent eftect.
370                                                        PSYCHOLOGICAL            APPROACHES

of these dirta indicatcd that the driver's            in thc vehicle he is overtakirrg. lf hc adjusts
rcsponse to rrn objcct in this path could             his specd such that this arrgular vclocity
indeed be dircctly relatcd to the component           werc maintainedat thrcshold.his own vclo-
of angulrrr  vclocity of the object.'l'rirnslating    city will decreasc    srnoothly r.rntilthc speed
the lincirr dimcnsions of the virriablcs ol           difl'crcnce is elirninated. The problem is
spccd and object placcrnent into irngular             complictrtedwhen the two vchicles are irr
velocity at the drivcr's cyc, all significant         motion, for thc following driver must adjust
differences  nnrong thcm could be eliminatcd          speed diffcrences to zero at a distrnce
incluiling intcrsuh.jcctvariability. On the           separation considerrrhly    greatcr than zero,
basis of this finding it beginsto be possible         It turns out that the systerrris only margin-
to apply a gcncration of psychological                ally stableandrcquires a virrietyof compro-
 research on thc pcrccption of motion to              misesorr the part of driversto insurercliable
drivirrg.                                             control, One consecluenceis that flow is
    Another area of interest is that of car-          con$trained and there is a finitc value for
following, whic:h concerrrsthe interactions           ilraxirlurn rate of flow or capacity of a high-
 in a stream of traffic between adjacent              wav.
vehicles. rather large amount of work has
           A                                              From a perceptual standpoint the dimen-
becrrdone by systemsengineersirr this rrrca.          sions uscd in car-following irre consistent
 Espccially noteworthy in this country is             with those used in locating fixed objects.
 that of Hcrmarr and his group at Gcneral             Work now r-urder:      way at the Burerru of
 Motors. F-clie and Foote. ol'thc Port of New         Public Roads suggcststhat stable $tcering
 York Authority, ancl more rcccntly Ohio               control also requircs the driver to scale
 State Univelsity under contract with tlre             another part of the visual envirorrmentin
 Ohio Depaltnrcntof Highwaysand the LJ.S.              terrnsof angularvelocity.Tlrus, at what is a
  Burcau of I'utrlic Roads. We ha.vc been              very early stagcofthe analysis    ofthc driving
 exarniningthe prohlcrn front the stanclpoint          task, there is an indication ofone perccptual
 of the tratureof irrfrrrmation availableto the        continuum that is LrorTrnron a driver's
 driver that he can usc to dctcct and scale            locating lrinrsclf in both tirne and spacc.
 changcs in speed and spacing, From thc                   On the hasisofsuch research      into driving
 analysisso ltr complctcdit appearsthat the            bchavior it is possible examinesorneof its
 driver rrrustshift from one perteptual pro-           irlplications relative to accidents.Taking
 cess to another in ordcr to stahilize his             thc lrroblcm of location of fixed objects,
 following. At long distances or where there           smirll crrors in judgment of angular vclocity
  arc vcry sntall speeddifferenccs   betwcenthe        cirrr irrcrcascmarkedly the probability of
  vchiclcshe will judge his position primarily         collision. Such variation is unavoidablc in
 by thc visr"ralangle subterrded the vehiclc
                                  by                   any perceptunl p.ocess.Or, if the field corr-
  he is following, This, in essence,is tlre            ditions are sr"rclt that angular velocity per-
 classicalProblcrt ofl distanccjudgmcnt,               ception is nrodified, the probability of col-
    Whctt, howcvcr, a minimum speeddiffer-             lision can also be rnarkedly increascd,     The
  ence exists,thc driver c:trn               visual
                               shif't f r-orrr         questiorr under what classcs conditi
                                                                  is                      ol
  angle detectionto angular vclocity discrimi-         rnay thc prohahility of such collisions
  nation. It so happensthat thc cquation fbr           increased?
  angular velocity relative to thc human ob-              Rernctnbering that a driver's jud
  server turn$ up as part of the equations              thtt an ob.jcctis outsidc his path of tr:av
  derived from empirical experirnents on               dependsupon his perceiving col'nponent
                                                                                      il             o
  trafhc flow. What appcarsto happcn is that           angular velocity,scveralfactors exist wlric
  as a driver ovcrtakcs another vehiclc hc              would ail'cct that perceplion. One is t
  eventuallyrcaches spacingat which he can
                      a                                 apparent size of the ohjcct: the snrallcr t
  just detect a conrponentof angular velocity           sizethe greateIits apparent vclocity. Anothc
  SYSTEMS RESEARCI.I IN SAFETY                                                                                       371

  is the shape of the objecr. Another is the                              ing along a highway. On the basis of most
  brightness the phjer;:1 its contra$[ ]vith car-following mcrdels,the
                       of                 or                                                              stability of the
  the surround.                                                           flou,dcpcnds a largccxtelrttrn the variancc
      One can structure a whole seriesof situ- in spccd in the queue and
                                                                                                         the lcngth of the
  ations hrscd upon the factors that markedly qucuc. Thus, shock-type
 affect the driver's detection of ansular veloc- ges rn spacing betu,ecn vehicles_may
 ity that would rrrodifycollisionpiobabilities. transmittcd down tlrc
                                                                                                      line whcnever an
  l n g e n e r 4 l o n c w o u l d a n t i c i p a t c h a t u n d e r acceleration imposcd
                                                        t                               is            by any onc vehiclc,
 n i g h t d r i v i n g c o n d i t i o n s ,t h e a c c u r a c ya n d Whether thcsc waves are
                                                                                                       dampetl Dut Dr itre
 r c l i n b i l i t y o l l o c a t i o n d e t e c t i o n s h o u l d propagatedwith increasing
 deterioratcnruclr more than cither presence tially depcndson the
                                                                                                  magnitudcof the srreed
 r.Jctection overtaking detection. Conse- changeirlposed, the spacing,
                                                                                                            and the sensi_
 qucntly, it is reasonable predictthat nisht tivity of the
                                         to                                             driver. When certain combinir_
 accidcnt rates for fixcct-object collisiJns tions of these flactors
                                                                                                    exist, the qucuc will
 should rise much morc than either angle, becorrrc unstahlc ancl rcar-end
 rear-cnd, or single-vehicle                    collisions, Data zlrls/occur. Of coursethc lirst vehiclcsin the
 on 10,000            accidents main rural highways qucue would not necejisarily
                                  on                                                                         or ordinarily
 indicate that thc rates ftlr these last-men- even be involved, although
                                                                                                         rhEir respon5c   ii
 tioned types of accidents incrcasedabout what set the chain in
                                                                                                  motion. tt is ironic, for
 one-hirlf fronr d*y to night, Fixcd-obiect ue hold each succeeding
                                                                                                            clriver legally
 collision rfltes.however,tripled frorn day to responsiblefor his colliding
                                                                                                           wirh a vehicle
 night. This type ol comparisonis obviously ahcad, Actunlly thcse
                                                                                                     drivcrs, under the
crude and highly equivocal.This writer lakes conditionsdescribcd
                                                                                                 here.werc purely inno-
 it, at best, as suggestivc, are almost all cent victims of a
                                             as                                                 fl1,p6x1;" systcm which
attempts to generalize frorll accident data.                             went out of control whilc they wcre a part
     Another extensionconccrnsthc situation of, it.
of two vetriclesapproaching euch othcr at                                    Another aspectof the following siluntion
night. lt is possibleto ask how the drivers is the yrroblcmof rapid
                                                                                                      overtakins of one
know whether of not thcy have clcaranqe vehicle by another.
                                                                                                 This is the caie where
 betweenrheir vehicles.lt turns out that for one vehicle comes frorn
                                                                                                     unconstrained {low
 the case whcrc an obscrvcr can see only a into a car-following
                                                                                                    siturrtion, Using a
very high brightnesssource of very small motion pcrt:eption
                                                                                               model, whcn in tirlc or
sizc, the apparent angular velocity will be whcre in distancc a driver
                                                                                                          detects rate of
tnaximizecl. thc drivcr (his nrearrs large ovcrtaking dcpcnds
                       To                                    a                                  on his own thresholdto
scparation hctwccn himself and the vchicle moventent                                and the speed    difTcrcnc:es which
lre is approaching.Thus, this night tlriving he is subjected.Assurning
                                                                                                        he hegins to act
siiuation in which most other cucsto localiz- as soon as hc detectsn relativc
                                                                                                             velocity,it is
atron arc reducedshould markedly degratte possiblcto deterrrrinc                                  thc mean dccqls1n1i.rr
the accuracyof drivers'judgment of place- he is requiredto mirke in
                                                                                                      order to insurc no
ntent and should increase                   the rrrobahilitvof collision. For a normal range of threshold,
head-onas well as shoutder-object                       collisiLns. this deceleration       function is shorvnin fieure
Again, examinationof thc increase night 2 as a function of relative velocity. es
                                                           in                                                         hay
accidcnt rates Dn main rural highways shows be seen the deceleration rcquired
that ncxt to shoulder-object                    collisionshead- non-linearlywith relativevelocity,It should
on collisions increased                 the most, that is by be notcd that decelcrationsgreatcr than
?.5 tirncs,                                                             0. I g are well within the capirbilities of most
     Turning to the car-following                       situation. vehicles,but thcy are quite uncornfortable,
     fe are two situations of interest. One con- What is notcworthy, howevcr,
                                                                                                              is that the
    rns the stability of Iinesof vehicles                    travel- driver is reqr.rired operate as a variable
372                                                     PSYCHOLOGICAL APPROACHES





                                                  of      VclocitY
                                       o= Thrcshold Anoulor

                                      R e l o t i v eV e l o c i t Y - M P H

                                requifed of a following driver when overtaking another
       Frc. 2. Mean deceler:ation

gain amplifier in ordcr to compensate for Suvulnv
increasirlg relative vclocity. This is not an    In srrmmary, two distinct views of highway
optimum use of thc human and should safcty may bc delineatcd: thc specific t-lefec
increirse the probability of such collisions, and the systcm-perfbrmancc vicwpoint,
espcciallyundcr condilions where visibility Most, thouglrnot all, of the safctymovcment
or other cues are reduced, or where any has been dominatcd by the first. However,
sizablc adaptation effects occur.              the system approach appcars to hold far
   From this discussion,it is obvious that the more promise, cspecially when we bring
 dynamics of the driving task are very central modern technologylo bcar on transportation
 determinants ofsystem safety. Hcnce' one of system design.
 the kcy objectivcs of researclron a system i$   The approach requires that safetyresearch
 to defirrc its perf'ormanceand to r-lefincthe be dirccted not at accidents but rather at
 $ource$ of variation in its operation that systen performance and from a psycho-
 causehigh variable error and, rrltimately, to logical standpointat driving behavior.By so
 find ways to rrrinimize thcm. In so doing thc doing it will bc possibleto not only define
 probability of accidents must be rcduccd'     the driving task requirementsbut also to
BEHAVIORAL APPROACHES TO ACCIDENT RESEARCH                                                373

specify the mismatching of the human to the     Researchand De velopment of the Bureau of
mechanical systctn, Fronr this tc:searchcan     Public Roads. Although the mission ofl that
comc both operating rules and criteria ftrr     oflice goes far beyoncl saf'cty, its ultimate
system modification which will increaseits      aim is the designof indivitiual transportation
safety.                                         systems which accidentsdo not ttccurand
  What has been desffibed here is the basis     in which thc rcstrictionson indivitlual lree-
of thc rcsearchprogram irr engineering psy-     donr of movemcnt because designlinrita-
chology currently under way in the Ofllce of    tions or socialcontrol may be eliminated.

    These interesting observations on the nature of the driving task and the driver's
response to changes in it are of major importance in undcrstanding the dynamics of
traflic flow and probatrly also of accidents of sonc typcs, rrs Michacls sugel;sts,and it
is to be hopcd that this very promising pioneeringwork will continue,H6wcvcr.,it is
unlikely, in view of the overwhelmingevidencethat drivers in accidentsarc not a
ranclomsalnPleof the clrivcrsat risk, that the sludy of represental.ive     drivers will
provideadequate     inforrrtation to many of the drive-r
                                 as                       charaiterisrics
accidents'  Conscquently,    therewould appcarto be no scientifically  adequgrc rudrti-
tute for the study of accidents,  mc'ticulously consiciered rclation to the characteris-
tics of the systcrrils a whole. Similarly, the achievement smooth traflic flow for the
overwhelningnra.iority driversrnaybut little influence pcrtinentcharacteristics
                          of                                 thc
of the small fractions of drivers who gct into accidents, excepa somewhatreducins
their enors throLrghreducing the difficulty of their task. Tire systcmsapproach hai
rnuchto contributcto accident    research, it sccrns
                                            but         unlikelyto contributcrnaxirnally
until it moderatcsits presentemphasis trill"fic
                                             on       llow and psychological  variables
with an intensivc  consideraticn irll potcntiallyimportantvariahtes,
                                  of                                    iicluding those
nonpsychological    variahles alreadyknown to he of importance accident
                                                                  in         causatlon,


   riy'econcludc this chapter with a selectionof some of
                                                         the most relevant remarks
from a conference behavioralapproachcs accidentresearch.
                   on                         to                   This conference.
from which wc havc quotedextensively    elsewhere this book, was uniquein that it
brought togethcra numbcr of prominenthehavioralscientists     with no prior specific
expcricnce accidentresearch orclcr to considerthe potential contributionsof
their respcctive
               disciplines the preventionof accidcnti.

   THE TAKING oF RISK                           gated for analytical purposes.Quite clearly,
 The next higher lcvcl of accident research     the lesshomogencousthc farnily of accidcnts,
 ight be thought o[as an attell]ptto idcntify   the more dilficult rvill it be to identify com-
hose causal firctors or circurnstanccsfound     mon causes,This suggeststhc nccd to do
o be comnren to m6ny accidcnts.It rcpre-        taxonomic research accidcnts order to
                                                                       on           in
 nts a kind ofresearchwhich involvesnranV       facilitate the classification accidentdata
  nceptual and rnethodologicalproblcms,         into more or lcss homogeneous       categories,
 Pcrhaps the frrst conceptual problem is        It is important to point out, however,that no
hat of'decidingwhat eventsare to be aggre-      single  "nlrturtrl" ttxcrnorny
                                                                               exists,Therc are
 Associationfor the Aid of Crippled Children, lg6l.
                                                            PSYCHOLOGICAL             APPROACHES

many possible, useful systems of classifica- degree than in the uneventful population oJ'
tion. The appropriatctress any onc of  ol                   risk situtttions.
thcsewill dcpcnd upon the hypothcses,               asso'      For cxample, playgrouttd accidcnts are
ciations, or relationshipsbeing sought or the conscquenceof an inrmense amount of
tested.                                                     playglound trctivity,Analysisof playground
    A second and related problem, and one acciclent                   data would revcalthat lrrostof thesc
which has beenpointed out often              by workers accidents occurred in the daytime. Can the
in this ficld, is that of decidingwhat gets in- daytirle situationbe viewedas a prospective
cluded in cach catcgory (i.e., what is an cause o1' playgror-urdaccidents?To an$wer
accident'l).    For any class of accidents'indi- this question it is necessaryto clcvelopsomc
vidual cvctrts can be further subclassifiedin mcans lor ttlslsuring the freclucncy with
tcrms of cost or severity, cou|sc,the morc which playground risk situations occut' by
severethe acciderrt,the rrrorelikcly it is to be night and by day.
noted and reported. For cxanrple, lead                         At this point a mcthodological ptoblem
 poisoningrlight rnngc in severityfr'om              rnild which has bccn given very little attention
 cliscorlfor-l,  which is not ordinarily reported, nrakes its a.ppcitrance.           How does one meas-
                                                                                           does one define
 to a ftrtal cvcnt, which is alrnost always ure exposure?That. is, how
 reported.+ Although this is a               conceptual the conditions that characterizea class ofl
 problem oi sotrteirrterest,it is probably not risk sitr.rations?             How does onc tleasure the
 one of trajor significirnce'lt might be sub- frequency of                   occurrenceof these risk situ-
                                                                                            a cltrssof acci-
 sumcd uncler the morc general pr:obiemof ations'l And, assumingthat
 classifyingaccidentsinto morc ot: Icsshomo- dcnt risk situatious can be clefincd
 geneou$ catcgorics for the purposc of countcd,                        how cloesonc observethe circum-
 corrmon-cituse        identificirtion.    The argument stances ils$ociatcdwith risk situations so
 for sub-classification severity might be that thesernay be compal'cd
                               by                                                          with the circum-
  strong whcn there is reason to suspect              that stancesasstlcirrted    with the accidenfs which
  scvere accidcnts po$sess diffcrcnt causal these risk situations have
  structures than miltl i)ne51 itr the r-'ase
                                       ils               of     Failurc to rccognize ancl deal with this
  automohile accidents.It rnight also be irn- problcm has resultedin an unfortunate
  portant methodotogically          whcn the sampling search situation. Analytical results which
  ofl rnild accidcntstdnds to be hcLrvily           biased possess more than spcculative
                                                                      no                           value are
  by reporting problctrts.                                   bcing constantly gcncrated. Despite the
     An cvcn more irlportant conccptualprob- seeming simplicity of thesc rescarchprob-
  lcm is orrc which has probtrbly been the le rrrs, still do not know whether men
  yrrincipal stumbting block to meaflingful saferdrivers than womcn, whetherit is more
  analy:iis accidcntdata. This is the problem drrngerous cross the strcct with the light
             of                                                           to
      "cxposure," ir prohlctn which in one form or tgain$t it, whcthcr girls arc stronger
  or ^'othcr is common tp 4lrnostall forrls of swimmcrsthan boys, or whethel aspiritris tr
  retro$pectivccausal analysis' Onc way of more deadly accident hirzard than lyc. We
  Iooking Lrtthis probleln is to cQnsidcrthat do not know whether excessive                       speed is a
  any class of accidcnt represeutsa sct of factor cotnnton to tttrnpike accideuts or
   eventsgcncratcdby a much lirrgerset of risk conrmoll to turnpike drivirrg. Despite the
   situations,tnost of which wcrc not corrspic- fact that turnpikes tcnd to have fcwer fatal-
   uously cvcntful. I he iclentification cau$es ities pcr vehicle mile than ttrdinrrry roads, we
   consists cssentially in identifying those cir- really tlo not know whcthcr
   culrstances    that were prescntin the accidcnt tribute fatalitiesor prevent them. ln short,
   casesin larger tncasttreor lllorc frequent thcrc is a rlajor problcnt in separating                  those
                                                              circunrstaucss   that are associated  rvith the
   + Surprisingly, thcre is no present reason to believe
                                                              occurrence of atr acciderrt in rr given risk
   that tros( such lirtal cases atc either reported ot cvcn
  diaBnosed. Eds.                                             situation frorl those that are flssociatcdwith
 BEHAVIORAL APPROACHES TO ACCIDENT RESEARCH                                                                    3'75

  the occurrenceof risk $ituation$. This prob- be related to the temporal ancl locational
 lem is not only a conccptualand mcthodo- distribution of accidentinvestigators                                (c.g.,rr
  logicalone; it is alsoir practicalone. lt means heavily patrollcd highway will report more
 that the rcsearch investigator in this field accidcntsthan a lightly patolled one). Un-
 c a n n o t d c p c n d s o l c l y u p o n a c c i d c n t d a t a fortunately,the solution to problemsof this
 collcctcd by some routine investigatingand kind will probably require,ccinccivingof
 reporting system.He is required1o establish accident reselrrch projccts in larger terms
 a large-scale        sarlpling nctivity in order to than has traclitionallybccn the case.
 obscrve risk situirtiorrsequivnlent to those                             A final but not unimportant methodo-
 which rcsultcd in accidentsand to collect logical dilTicultyin this type of investigrrtion
 identical circurlstantial ancl hackgrountl has to do rvith thc problcrn of statistical
 data. The nrrmberof times that this has bcc-n inference.Accidents are lrequently a$sumcd
 done in accidentresearch                investigations       can to be generatedby a sinrple lloisson pr()cess.
 probabty be cor.rnte<l tlrc firrgcrs of one For thc moment, the validity of this ilssurnp-
 hand. One must continue to expect thilt tion will not bc qucstioned.I he existence                                 of
 casual analytical studies of accidcnt data a simple Poisson procesriimplics that the
 w i t h o u t p r o v i s i o n f o r e x p o s u r es a m p l i n g variance of the sanlple ofl irccidcnt cases
 will sirnply mudclythe waters with misleird, generatedduring any given period of ob-
 ing or uninterpretrrhle               results.                       scrvation will be equal to the nccidentex-
    There nre other nrethedological                   problems pectation during that period. Under this
 in studies of this typc. One of thesc is the condition, there will he a very low order of
 problem of observation and reporring, precision associated with the estir.n.dted
 Rarely is the accident researcher                   willing or cxpcctcd value unlcss the sample is quite
 irble to set up his own observational system, largc. For cxample, with a sanrple of trne
 This rncansthat he must depend upon the huldred accidents,a 9,5-pcrce-nt                                  c-onfidence
 efforts of prof'essional invcstigators who interval covering the expectedvalue would
 routinely investigate rrnd report accidents_ rangc {tom approximatLly ttl to l2l. Only
Thcse may include physicians,policcmen, the most grosscll'ccts                                ctiuk'lbe establishcdin
coroners, or others who have professional such a case,This rncansthat any accidcnt
responsibilitics         along theseIincs. The prob- researcherintcnding to establish a spccial
lem lies in dcternriningwhat snrnplingplan systern                            ofobservationand reporting in order
is to be followed f'crrdcciding which accident to ctvcrcomc any of the above-mentioned
casesto irrvestigate, well as determining problems is laced with a fornridable dirta
what circumstantial and backgrourrcldata collcction proHgm in developinga sample
should bc ohscrvcdand rcported,                                       of suitablc sizc.
    Unfoltunirtely, the typical accident re-                            This is especially true when it is con-
$earcher not in a position to control thes$ sidcrcd that a given subclass of accidents
most crucial aspcctsof his rcsearch.As a established to providc homogeneity with
rcsult, only the most obvious and immediate respect both the risk situationand severity
iterls of prospective signilicance are re- will prohably include a rather srnall propor.
corded, and thcsc arc f'rcqucntly reported tion o[ the cases which occur in a given
incorrectly and incotnpletely.Furthcrrnnre, period of time. In fact, it may well happen
obscrvation is generally liftited to the rc- that the amount of tinie required[o accumu-
cording of qualitativetactors. Seldom does late a samplc of suitable size for anaiylis-t1l
it includc quantitativemeasurements cir- purposeswill result in risk situationswhich
cumstflnces,        iLctions,envilonmental condi- arc no longer hornogeneous                          bccauseof thc
tions, etc. Pcrhaps$onrewhat                 lessseriousas time intcrval over which thcy were collectcd.
a rcscarch pr:oblem is the fact that the                                  -HEnsr.nl H. JAcoBrt,     PH.D., pp. 7-l I.
samplingoftccidcnt cases                tendsto be biased.
For example,accidcnt rcports are likely to                                                              of
                                                                        llazaryl untl Risk: ln a series most pro-
376                                                      PSYCHOLOCICAL            APPROACHES

vocative works. John Cohen and his associ' ior in garnc thcory or learning situations.
ateslravcdcalt with tendencics risk-taking. Ccrtirinlythc issncs the trvocases
                                    of                                    in                 rcsemble
                                                                                  profitahleto point
Suchrrran,Maclver, Jacobs, Deutsch' and each other. But it may be
othcrs of us herEhavc stressed       thi$ aspectof out certain aspects of bchavior rclating to
behavior. Wc       nright profit from further risk which are very important to pcoplewho
thought about defirring     this rrncl relatedcon- deal with conceptsof safetyand prcvention.
cept$, as well as acciderrts.                           Considcr the raugc of risk-taking behav-
    Most pcople tend to usethe term rrsft with ior. Ordinarily wc assumethree qualitative
two nreanings.    First, they nreatrthc inherent classesof people: thosc who are pcr:fectly
                                                                                 with thcir eyeswide
danger in a situation. Usually the situation aware of danger and act
                                                                                   partially awarc of
or action is calledrisky, implying in a mixed open; thosc who are only
way both that therc is dauger and that the           the hilzrrrds; and those who are ignotant of
                                                                                complete safety. But
actor is behavingwith likclihood olaccident thc hazards,assuming
                                                                                            parallcl to
or injury as a result of this behavior.But if thcrc are important classeslying
no ofle is actually bchaving, pcople talk            those three lvhich mirkc up a cc)nsidcrable
about thc absolute risk rtf a situation.' group and whosebehavior might bc
                                                                                 of the circumstances
Secondly, people talk about a person's if we reirlized ssmething
taking a risk, inrplyingthat hc hasdeliberate' of their behavior decisions,
ly entered a dangeroussituation. Thcre is                Thoscwho take no risk includelarge num'
                                       "risk" is at- bcrs whotn wc are attelrpting to educatc.
little implicatiorrherc that the
 tributable to the situalion-more to the But this is a twofold education,lmplicitly we
 person'sbehavingin this adtnittedlydnngcr- recognize           thosc who, if asketl,wottld know
 ous situation.                                       that a circumstttnce hazardous' But they
    I propose that we use       two words which are not asked.either by others dr hy thcm-
 will separatethese mcanings more clcarly- selves.           Thcscpcoplcare thc targctsofgeneral
             "hazard" and "risk.'n "Hazard" campaignssaying, "Bc careful." The other
 They arc
 will be ob.iective danger or likelihotld of group, which is ignorant of hazard
              "risk" will bc subjcctive     estimatc therelirre takeri no risk. is thc one which
 of hazard.Deference       must be given to John genuinelydoes not know about thc hazard'
 Cohen,   wtro used these words irr their rcla' Such a one is thc object of attcmpts to
 tivc scnse at rlnc time (margin of hazard, eilrrcateto facts rather than to ilrrposc a
  margin of safety)and as probabilitiesof suc- persistent htbit of behLrvioror attitudc.
 cessand estirnatedsucccss another (maxi'
                                 at                      Anothcr parnllelclassirrcludcs      the person
  mum risk and maximum hazirlr'J).                     who knows that tlrere could be dirngcr but is
     Tlrc usage proposcd-that is, hnzard :             uncertain about its presencc.He rnnkes two
  danger, and risk        estimatcd hazard --can bets,whereirs one who knows the datrgcr
                                                                                                 "lf an
  becomcforrlirlizcd casily,operatiotral      deriv- rrrakesotrly one bet. The latter says,
  ativesconceived   casily,ri /a Cohen,and con- accidentcomes,I am willing to acceptit," or
                                                       "If the danger comes,I can handle it." Thc
  formity to popular tcndctrcy nraintirincd.                         "Not only do I say what the
  This may be a bctter tirke-off point than forrner says.
  either popular usagc,which is confttscd,or latter says,but I am bctting that the clanger
  Colrcn's usage, which is a little unreal in has low probability." This pttts him at a
  parts (e.g., maximum hazard ,= empirical disadvantage termsof preparation.
                                                                     in                         Among
  probability of success).                             thosc in thi$ clasti is a special group who
                 +     !      t     *                  equate uncertainty about risk with zero risk.
      Classes Risk: The concept of risk, used
             o.f                                          Another parallcl class is represented by
  in thc senseproposed here-that is, percep' the person who has no tlrought of
  tion of true hazard-has beenlittle exarnined (althorrgh if asked, hc could make a reason-
  in thc litcrature. There is, of course' con' able guess)          and is suddenlyconfronted with
  siderable literature which     clealswith percep- an unusual danger or utrltsual situation or,
  tion of probabilitiesand selectiotr bchav- pcrhaps most itnportant, a potentially dan-
 BEHAVIORAL APPROACHES TO ACC]IDENT RESEARCH                                                                     377

 gerous situation before hc can tell whether it                          mental changeofrisk and the accompanying
 is darrgerous.Such it person *'ill cliffer change in willingness to bc cxposed to
 markedly fronr the one who knows the                                    polcntial hazard. J-hese     nrattcrsare subiect
 danger and is likely to be prepared for it.                             to filany influcnces-the nalipnsl cultural
 Not only is tre not as wcll litted to deal with                         attitude, the local subcultrrralpattcrns, the
 dangcr by virtue of anticipation but, more, urges of pcer grouFsj etc, Tlre vnriability
 he must make a morncntary estirnatc of                                  must be en()rntous,  but I think thc data can
dangcr undcr stress. He must thcn dccide be made available. Here also cnters the
whcthcr to take the risk. It is likcly that the nttrrrralparadigrnof reintorccmcntin learn-
 danger is increased this kind of estimate ing theory. Reperited
                                 by                                                             positivc reinforcement
 made in ail emotional state,                                           (that is, no accident accompanyingsafe or
     When a persorr not thinking of hazard, unsafc actions) strengthensbehavior pat-
 he is essentially mcmber of thc previous tetns very nruch ancl also pcrrlits much
 class, the one with uncertainty about lhe                              broadcr generalization both attituclcand
hazard, since he acts likc him. Hc doesn't short-tcrm learrrcd rcsponses.
 act with the grcatest           conljdence, thc zero-
                                                   as                      Emqathic ,4hilit.t, o.l |litrenrs: One im-
 risk man would, but with confidencethat                                portant point u,hich, I t'eel,  has hod virtually
 shotrlddangcr arise,he can recogqize nnd no attention but should rcceivc much is the
deal wilh it.              is sornewhat        likc the person enrpathicability of thc parcnt. If a parent
 in Suchnran's          paper who is irlcrt to unusual acts toward the child as if he expects rhe
events or objccts even though he doesn't chitd to perceivedangcrsbut the chilcl has
activelyexprcss continuL,da                  scarch.The im-             not yet lcarnetl tlrcse, r,vccan say the parent
portant point here is that such a man must shqrws                               poor empathy,especially hc knows
nrakcan cstimateas to his hazard rnonrenta- somethingabout thc limitationsof the
rily and under litrcss,His actions probably But far worse, what i[ hc not only shows
difl'er m:rrkedly frorn what they woukl bc if                           poor cmpathy bLrtalso is truly ignorant of
lrc were able to judge the danger with the dcvcloprncntal                                   facts of children'spercep-
greirter deliberation. Tlre moruerrtaryesti- tion and understanding'l helicvcthat this              I                  is
mator is much morc subjectto deep-scated the casefar too often. So many times has the
habits of behavingthan the delibcrateone. v i s i t i n g n u r s c h e a r d t h e m o t h e r s a y , " H o w
as well as to trttitudeshe has picked up, For coLrld he have drr,rnk keroscrrcwith that
examplc, thc nronrentary               judge, I maintain, awful taste?" And il the wtlnran clocsn't
will be far rlorc likely to rrct as if he wcre know that this is possibleand how it is pos-
i n v u l n e r a b l eh a n h e w o u l d i f h c h n c lt i r n ct o sihlc, she has the somewhatsad cr.mpanion-
think about it. It is, of course, essentially ship of our scientists her ignorantre. hese                          'l
equal to betting that this will not be the two faclors-ignorance as to childrcn'scapa-
nronlcnt of truth for the flctor, ffii1ybg r*x1 citics and inability to a$sume                           the position of
time, but not this; maybe to sorneone                           else. the child in his own world-ilre of'rhe great-
       not to him.                                                     est importancc in rnany classes accidents
    All of thc foregoing classeswould prob- in which blame attaches the parent a$ the              to
         produce diffcrent actions at various nrajor controller o[ the acciderrt                            situation.
     :ardous points. My personal opinion                                  -Benurno H. Fox, Pu.D., pp. 50, 52-54.
  bout rclativc prevalence theseclasscsis
  l) not thinking about hazard; (2) equaring ON AnrrclpltoRy JuDCMINT
  nccrtaintyofhazard with zero risk; (.3)be-                              4" One typc of study should bc dirccted
 ief in n,t hazard; (4) clcliberateappraisal 1rf toward dctcrmining thc rclevancelor a""1-
    zard, And, of coursc.thc beliefin l)LrrFonal dent avoidanccof ccrtiiin pcrsorral                            capabili-
 nvulnerabiliryis part of all fbur classes.                            ties, attitudes,and behavioral patterns and
       What might be the reasonsfor thc nre- the possibilities devclopingthose in such   of
              of attitudes toward hazard? Onc a way as to increasethe competence the                              of
           ant consideration is the develop- person for coping with hazards,
378                                                               PSYCHOLOCICAL                APPROACHES

    As an illrrstrationlet me suggest                  that one             This is as good a place as any to interject
stratcgic      clustcr of personalvariablcsis that the caution against pcrmitting cnthusiasrr
c o n t a i n i n gw l r a t w e m a y t h i n k o f i r s a n t i c i - for devcloping accidenf-;lvoidantskills to
patory and judgnrentirl skills which entrble stunt thc growth ttfia high clegree ventr.rre-                of
thc pcrson habitually to review and assess sornenuili                                and willingncssa$ well as discip-
oncotnittg cotrtingcncies               and trtilizc these lined ability to take real risks, Mty a cretrtive
proccsses prcvcnt and lvoid or to rnobilize Providence preserve us lrom a concept of
to meet such contingcncics.                   Thus a parent lifc safcly wrappcd in cottotr hatting!
                                                                                 -Lrouenn S. Corrnnll. Jn." Pu,D..
skillful at rrtilizing anticipatory review atrd
a$sessntent        proce$se$      may see tltc       potentials                                           pp.153-15a
of a poisoning accidentto the child who is
                                                                         Or'l Ar:r:lopNr BesAvton
beginning to rerclt lor things on the table,
and, in view of these anticipatory                          possi-          There is sonre clegreeof the unexpected
bilities,lock thc poison up. In a similarway, and unavoidablcin all cvcnts,and which of
thc c:hilcl      ricling his bicyclc dowrr thc street thcsc cvcnts is callcd an accident dcpends
might bc trained in habitual anticiprtory upon thc cut-offpoint one wishesto use.For
assessrncnt that lre will be able to devotea tlre re$earcher,
                   so                                                                      this means that almost all
part of his mental activity to a previcw of events can be studied ftrr their irccidentrr
                                                                                                 "accident$" r11ust
whiit peoplemiry lre doing at the next drive- qualities-and that all                                                 he
way or intelsection,or aroutrd the col'trcr, studied in terms of their nonaccidentir
and havc reacly a course of action to meet qualities.This approach opens up fbl ana-
thcse contingencics.                                                     lysis large areas of human activity which
    Should such a cluster of capabilitiesand previously were not cnvisioncd as bcing of
their habitual excrciseprove significantin interest to the accident rcsearchcr:losing
 cutting down on accidents, would then he articles, forgetting appointmcnts, etc, Simi-
 desirableto plLrrr program of cxpelimental larly it forces thc accidcnt reserrrcher to
 stucliesaimecl at t'letcrminirrg                how bcst to challenge a great rnany of his current con-
 develop thesecir"pabilities tt'ititrpersotts cclns (e.g., adult poisonings)in tcrms of
 in their utilization.                                                    their inherent accident qualities.
     5. Researchon the rclevanceof the anti-
 cipatory judgmcntal cotlplex for acciderrt
 prcvcntion atrd cxpcrirnerrtationwith de-                                  Our analogy betweenaccidentsand dis-
 veloprlerrt and optirlrrl utilization of these ease also sheds light on two opcratiotral
 capabilities       shoulcl. ctlurse,bc directedas problems that havc bccn hcld lirrgclyrespon-
 nruch      toward parentsand parent surr()gates siblc f'or the lack of researchou irccidcnts.lt
 as toward thc child. lt is indeed obviously has bccn claimcd that accidcnt rcsealclris
 morc relevant for- the parcnt than for thc di{Iicult bccauscaccidcntsa.rcrare iurd unan
 child during the carlicst dcvelopmental ticiuatedevents.But this is irlsotrue of most
 stagcsof thc child, though it is important to discascs,                             which may occur oflly oncc in irn
 recognizethat even in the first yerrr it is individual's lifctime and then at rr tirrretha
 possiblcto do sotle trriirringof the child in cannot be arrticipated. in the caseof tn          As
 anticipatory perceptionand ad-iustrlent.                           In diseases,     we can anticipatc that accident
 this connectionit should also bc pointeclout lrai//happen br.rt not ur/rea,
 that work with parcntsshorrldincludecleter-                                 Perhaps an cverl rnore telling point is
  mination of possibilities training parents question the valitlity of defining acc
  to train their children in the anticipatory altogcther as ral'e cvents. What is real
  capabilities. Direct experilnctrtatiou wilh meant is that out of the laree nurrrbcr
  flqvclopmentLrndutilization of anticipatory unexpected and unavoidable evcnts t
  capacity in the child should obvior"rsly                         in- involve any single individual, only one
  crcase with his iucreasing independcnt two lnay rcsull ln an lnJury serl()us                                  en
  mobilitv.                                                               to require medical attention. lt is our con
 BEHAVIORAL APPROACHES TO AC]CIDENT RESEARCH                                                                    319

  tefltion that any individual is constantly                              spectshould not be a serious     deterrent  to
   having acciclents,          and that we cnn lcarn a                    rcsearch.
  great deal about this phenomen0rt study-               by                          -EoweRp A. Sr,rcrrllnN.     pH,D..
  ing all acc:idents       rather rhan only those that                                                   pp. 32,36-37.
  result in a rcportlblc iniury. This point is
  especiallyrclcvant to accident prcvention, Oll Drvrlopl,rENT ^ND LEARNTNc
  becitusc    wc are intcrestedin decreasing                     thc        In prescientific times an unexplained
  l a c k o f p r e d i c t a b i l i t ya n d c o n t r o l i n t h e series of events resulting in injury was
 situation quite irpart frorn thc clcment of                             attributed to the work of the devil or orher
 injury,                                                                 shady forccs. In primitive societies, the
      It would he cxtremely interesting to com-                          falling rock that brokc a leg was seen as
 pare accidents with othcr phenorilenain- having a definite purpose of aggression
 volving the unexpectedand unavoidable, ancestralretribution. Adults today, for
             "accidents"                                                                                            the
 such as                        in wlrich one oversleeps most part, do not indulgc in such fancics,
 Or fbrgcts an appoirrtmentor loscs an ob- but thc present-dey                                child taking a lall and
ject, In what respectscan thesc fbrms of                                blaming the floor flor hitting him reprcscnts
 behavior be classed with accidents which this animisric approach. Althoirgh it
                                                                                                                 is not
 produce bodily injury? Another important intendedhereto posit a social
empiricalquestionwould irrvolvethe dcgree old "ontogeny recrrpirulatcsphylogeny,"
to which actual accidents                  irre recruitedfrom           such juxtaposition illustratesconcretclythe
 near-acc-idcnts. therc othcr diffcrcnces lact that children pcrceiveacciderrts
                         Are                                                                                    differ-
 besidesthc injury between accidents and ently fron the way adults do, ancl that
close calJs?                                                            dren's perccption of tlrenr is a frrnction of
     The other operational point that has been achieved lcvcl of cognitive dcvclopnenr,
made concerning the difiiculty of accidcnt Fronr fhis it follows thaf there
                                                                                                           arc iunda-
researclr the impossibility of anticipating mental and qualitativc dill'crcnces
              is                                                                                             betwccn
accidents,Although this may be true f<rrtl-re chitd und adult accident potcntial_
singleindividr-rnl, certainly is not true for
                            it                                             /1fu|t1s. Child At'<:itfuntu For an adult,
the group, as witncssthe unf,ortunate                      success hirffid                                object rela-
of the National SafetyCourrcilin predicting tionships have beconre stabilized in thcir
the nunrber of trafiic deaths for each rnajor basic lanriliarity.On rhe other hand.
                                                                                                               a child
holiday period. Thc entire fielcto{,insurancc proBres$rng                            through various dcveloprnental
         upon highly accurate actuarjal pl"cdic- stagesis continually in an cnvironmentthat
 ions. Tlrus wc ntay count on certain acci- in tnanya$pccts daily newly                   is           structure<l and
        s occurringilt ccrtain times at predicted incrcasingly                       nrorecorlFlex. Wlren thc child's
          for diffcrcnt grLrups of individuals, mother cntcrs the car and d lives
                                                                                                               to thc
  vcn though we rnay not be ableto anticipnte $upermarker thc daily shoppingroutine,  for
    hethcr Mr. Jones will hc hurt in an auto- she is following thror.rgh a planned                 on              and
    obile accident next weck.                                          prcdicted course of behavior, involving a
     Howcvcr, there are fiany other areas of                           low but dcfinitc risk. Thc risk is compre-
      ial researchin which this type of predic- hcnded but prohably cognirivcly
 iorr is also impossible,            Wc r,:annot     anticipate because the history of previoussafc iden-
      individual's cornrrrittirrg crime, or get- tical cxpeditions. In other words,
                                             a                                                                 shc is
 ing a divorce,or having a child, or succeed- likely, though rninimallyaw:lre of risk,
                                                                                                                 to be
       or failing at his .iob, etc. Most medical automaticallycautious. ln caseof cmergen-
              procccds without being able to cics, there is a set of huilt-in
   udy a discase at the exact mor.nent of its conceptual-6otor mechanisrls that,
           nce. I1 would be helpful to be present without prior personal disastercxpcrience,
   hile the irccidentis occurring,and in some could be acted out or verbally cxplained,
         even this may be possible, but, in any Fur(hcr, any unnsual figure-ground vari-
     nt, the nccd to $tudy accidentsin retro- ations which imply impending danger
380                                                     PSYCFIOLOGICAL APPROACHES

more likely than not to be noted (or we            children.The question What ate the im-
would have an even heavicr accident toll)          plications thcse
                                                             of    incidents whatmight
and to dctcrmine appropriate avoidanceor           bc the relationshipbetween these minor
protcctive behavior.                               accidents maior ones?
   A nursery-school child prrt irtto the back         llli.thaps't,s,At:cidents: The child has the
yard cvcry day to play is involved in tr less      nppropriate responses well as the mot()r
complex and see    rningly more unitary activity   and rcflex capability to protect himself in
than his shoppingpirrcnt,In actuality,how-         accident-producingsituations where he is
ever,thc play activity of the child is far nrore   fanriliar with the participating agents trncl
variable and oflers fewer predictive or ex-        wherc they can be physically managcd. But
periential guidcposts: the child has only          elementsof an cnvironment arc constantly
emhryonic and macroscopic concepts of              being changed by the actions of adults,
time and causality and does not prograln for       nature, or the child himself.Adults are rnost
himself a $cqucnccof play involving more           likely to reordcr thc cnvilontncnt in ways
than two stcps. He ascribesvolitional pro-         safefor a chilcl.But natureand thc child him-
pcrticsto inanimateohjects,He is continual-        self opcratc in almost completely random
ly reaching flor new implements in the en-         fashion and may so relate objects that they
vironment and is always experimenting with         have new accident possibilities,with this
relatingthesenewimplements one anothef
                                 to                continuingchangc,thc child finds himsclf in
and to his increasing dexterity. Limited in        a seriesof evolving and new situations re-
his ability to relate antecedent events to         quiring new adaptirtions.This repeatedne-
consequences,     with expcricnccsinadequatc       cessity ncw adaptatiorrs a characteristic
                                                            fbr                   is
to many new situations, with a low threshold       of the child's world and includes maladapta-
frrr distraction and a limited attention span,     tions that often serve as trial-rrnd-error pre-
the c,lrild is likcly not to perceive potential    cur$ors to approprirrtereactirlns.Many of
danger in his relating of himself to play          these rlaladaptationsresult in physicatdis-
implcments.Jf there is suchperception, is  he      comfort or injury and thus are accident$.
likely not to remember the possibleconse-          However, they might be more appropriately
quences whcn some other attractive stimu'          calledadaptationor learningmishaps.       These
lus catches his attention. Even if he does         mishaps,together with his successful      irdap-
rccognize potential danger, he may or may          tations-products of the child's cxperimen-
not know whirt to do ahout it, depending on        tation-scrve to teach him about concrete
the nurnber of past repetitiotrsof similar         cause-effect rclationships. With sufficient
situations,  his intellectuallevel,his develop-     repctition, a basis is crcatcd for learning an'
mentirl stage, and the degree of protection        ticipatoryavoidance reordcring response$
offered in the child-rcaring practices of his         Thc implication of the ftlregoing is that
parentsitnd socialgroup. Evr:nif the problcm        mishaps are a natural arrd necessary    part of
involved a protcctivc rlotor responsethat           a child's development and result from his
wfls in his repertory,the ratio of comprehen-       own exploratory actions on an unyielding-
 sion to fear would have to bc such that he         or sometimes too yiclding-environnrcnt,
would not lose rnotor coorclinationbecause          Freedom to explore implics rnishaps,but it
 of the dominarrce of the disorganizing             also impliesfreedomto lcarn. The adult pr
effects of fear.                                    tective rolc, whcthcr it be taken by parent or
   It is no wondsr, then, that in a seemingly       by community, is to protect the child frorn
 unitary play situation the child has many          accidents,which might be undcrstood as t
 accidents,  most oflthem relativelyminor and      Tnore senollri ano rnJuflous coflsequenccli o
 unrernarkable.    This is well recognizedand      his explorirtory behavior. It might be
 even acccptcd. The bandagc nlanufacturcrs,        gested that a small child's environ
 for exirmple,  market their productsin a wide     should be made safefor rnishaps.In additi
 variety of shtrpesand colors to appcal to         to the cognitivElesson, mishapmight ser
 BEHAVIOR.AL APPROACHES TO ACCIDENT RESEARCH                                                                           381
 as a kind ofinoculation against serious acci- repeat the experienceby touching a light
 dents, mainly through incrcasing       the child's hulb, as there would have been no generali-
 confidcnccin his own ability to dcal with              zation,irnd the Iight frorl thc llickeringllirrrrc
 uncxpcctcd happenings.                                 is a quitc differentstirnulusfrom rhat of the
     Perhapsit shoukl be statednow that there steady bulb. A five-year-oldchild, on the
 will be no attempt in this paper to dcvise other hanr'1,                     toasting a marshrualk)wand
 rigid rnodels frorn lcarning theory or else- accidcntally                  burned in the proccss,             might not
 whcre accordingto which it is presumedall              lct the singlc experienceintcrfere with the
 childhood accidcnts can be understood. rapid toastingof the marshnrallow.                                      But if he
 Rather, a variety of modcls and organizing had a scriesof sur-:h                       negativereinfrrrcernenfs,
 principleswill be suggested the hope that he is likely to generalize set ofprecaution-
                                in                                                              a
 by "trying en" differentLroncept$, qucs- ary maneuversin regard to all lires.
 tion of dcvcloprnental  considerations acci-in             In othcr words, at higher developrnental
 dent prodrrctiolr rrray bc clarilied. It would         stages, whcre the quality of conceptual
be only a Procrustcan     ideal to fit any bchav- generalizationitnproves,cxploring und .*-
ioral phenornenon,     including accidcnts,     into periencing of mishaps may have grcater
onc cncompassing       model.                          effect in lenrning accident-controllinsand
    For the behaviortrlscientist,    the question avoiding behavior. But this ,,...r*uri-ly .",
is whethcr thesc mishaps do indeed play a latcs also to the strength and lrequency of
crucial rolc in thc channclingofa developing thc injury.
child's exploratory br.-havior.     As was pre-                  -MnnrlH DErrTscH,ps.D,, pp. g0_93.
 viously stated, the ovcrwhelming proba-
bility is that they do. It is felt that they have ON Euorrohr.{L FACrr)RslN A(:ctDtNrs
bcen insufhciently considered as *ignificant               Many accidents occur whcn individuals
factors in the developrt€ntal prpl:s55,               are ernotionallyupsctcither on it temporary
    It may be hypothesized    that thc quality of or more or lesschronic basis;irnd accidents
the accidcnt experience, and thercfbre its are no doubt nrore likely when peoplc are
effect, is a function of the strength of the upset. Howcvcr, being upset at one timc or
stimulus(or the severityof the resultingpain another is an incvitable cirnserruence                                       of
or injury) and of the developmental         level of I i v i n g : t h e a n s w e ri s n o t t o r e c o m m e n d h a t
thc child, in tcrrnsof his ability to generalize. people stop bcing upset but to educatc the
We might therefore postulate that a child p o p u l a t i o n n t h c v i e w t h a t i l a n i n c l i v i d u ai ls
who has never fallen downstairs is more undcr unusual stress,whcthcr crnotional or
likely to fall than one who has expetienceda physical,he or she should if zrtall possible
previousflall,A child who has fallcn and not withdraw from any activity with
hurt himself is probahly more likely to fall potential, until the stressis dissipated.psy-
than one lvho lras cxpericncedpain. More- chiatrists do see paticnts who are accident
over, a child who has fallcn on a number of prone for a parlicular type of acciclent,                                 usu-
       ions and been left with uncorlftrrtable ally rninor in nature, for specific psycho-
     n sensationsis more likely in stepping logical reasons.                    For exanrplc,a paticnt rnay
    wnstairs to perceive nervly dilTerentiated frequently incur burns on the lrantls, and
   pects of thc environmcnt, such as the there may be reason to speculatethat this
    nister, which cirn he functionally inte- repctitive burning is a neurotic way of ex.
     ted into his behavior. The phototropic piation, By and large, howcvcr, sush in-
  ne-ycar-old  child rvho reaches thc flamc $tances ofgrcater interestto psychiatrists
                                    for                             are
 n the $toveexpcrierrces   single-trial  learning: than to safety workcrs.
    hough the flamemight continueto irttract              Tlrerc are those wlro are $o seriouslydis-
   s attention.after he has once been burncd t u r b c d e m o t i o n a l l y h a t t h e y m a y h a v en r o r e
 he usual sequence tactile investigation          is l i k c l i h o o d o f g e r r i n gi n v o l v c d i n a c c i d c n t s
   hibitcd. However,a child at this agewoulti tharr the averagcindividual, Pcrhaps l0 to
                                                                                                    PSYCHOLOGICAL                                 APPR.OACHES

I 5 pcrcent of any population have fairly       summoncd. Most people can train them-
serious emotional prohlems, and no         job  selvesto watch out for dangcr signals in
                                                                                To the exterrt tlrirt
wlriclr carrics morc than minimum accident their mood nfld astiolr$'
potential shoulcl be opened to theln' How-      aggressive       feelirrgscrrn he brought ttt the
                                                                            steps can be tahen to
.u.r, n,, harcl ancl fast directions can be Ievcl of awarenes$.
                                                                                         with self-
                                                                           One diflLc:ulty
given; cach situatiorr I'rasto bc evttluated on countcract them.
                                                examitratiotr one'sfeelings
                                                                   of              atrd motivations
its o*n mcrits. Many of thesc individuals
can be retdily identified as disturbed by       is that many individuals with emotional
                                                                         largely unaware ol'their
 supervisorsot co'w(trkcrs. l{erc arc solne d i s a h i l i t i e sa r e
 examples of psychological tactors in     acci' destructiveimPulses:
dents and near-accident$:                                                                          A younB adnlt with marked personllity
                                                                                               1 s m 5 - . - 1p - l i n l c r h y t r a d e - t o l d h i s p s y c h i a t r i s t
      A young ntlrried wolllitll who sufrered from a                                           tinc day that hc h:rd becn taken ofF the press
  severe nettrosis wits awilre that ller lirst-bol n son                                       sivcn unskilled work to tlo He was trnable to
  wirs htr lhvoritc, anrl sht resented the birllr of a                                         irnderstand why this hatl happenetl' However,
  sccrtntl hoy, allhouglr hc was I wir\iliilg                                                  much later in thc inlcrvicrv hc recountc'l blandly
                                                                                 pcr-                                                                                   press
                                                                    tl l
  b e a u t i f r t l b r b y . H o w e v c r ' t h c r c s"ct i' lo rlg c ltl"t
                                                                         c hinr                how he had hctn about to tutn orl his
  s i s t c d a n d s h c h a d h c c n f in o w n t o                                         whilc his $uPctvisor ltld illl asslstltrt wcrc
  it the conclusion of visits to frietrds' and rela-                                           working on itls intcrnal machitrery Luckily'
  tivcs'hirlres. Whctr lrc wits two and a hall, the                                            solnc()rt lloticcd Ilin] and knocked his arnr
  mother lett hitl utrattendedill tlle bilckyardi lle                                           iiom thc t)owcr switch at the last second' Even
  fell oll'a retainillg watll into it lilke ilnLl wits                                          afier desctihinll this inr:ident,he was unable to
   drowned.                                                                                     asso$s it propcrly in terms of the Ereat resent-
     A father who was rather quiet and passive in                                               nlcnt he had for his supervlsor'
   naturo wirs ()pLrratingiltr clccilic saw itr hic
   mcnt- t'li5 lw(l prLr'ildolL:sccnllh()ys citlrlc lll llld                                    Othcr individualshavc very imperfect
                                                  (hctl] to
   o1q'1 s whilc bcgan to liEh(. Hti ordiJrsd                                                                                       and destfLtctive
   stop. Ihey kcpt on tighting. and in his itngcr at
                                                                                              control of their aggrcssive
   beiirFdisobeycd hc cut oll the tip oi his thumb'                                           irrrpulses   even tlrough they are to an extent
        A mrtn who resentetl his wife's tlonrination                                          awarc of them. Both those who .ltrck            aware-
   was asked by her to gct il nlassive dilrthenware                                           ness attd thosc wlro are itware hr.rt lack
   o i t c h c r f r o t r t I . h ct o p s h e l f i n t h g k i t c h e n ' $ c v c r a l
   iimcs prcviously slrc hncl ashcrl tiirrr l{} usc tho
                                                                                              control are in need of psychiatric assisttrnce.
   kitclieir stepladdcr ldr such tasks llt rushed                                                 Certain drrrgs which all'ect tlre llcrvous
   iirto the kitchen reselrtlul at being interrupted,                                         system,especially         bcverage    alcohol, irr addi-
   ncslcc{ed thc sicplaclder' reachetl up iln(l lost
   i o - n t n r l , r f t h e p i t c : h c r ,w h i c h c r a s h c d i n l d t h e
                                                                                              tion to impairing motor coordination, may
   stove and chipped ofl several squrre inclrcs ol-                                           irllow the rclcaseof aggressivc          impulscsthat
    norcelain. Onli then did lre turn alound t'i sec                                           would otherwise under conscious
                                                                                                                       bc                    control.
    iris s,tall daughtcr staring up rt hirn- I'lte girl
    hild silontly folloucd lrcr fltlrcr init) thc kitchcn                                     An exarlple of this type of pathologictrl re-
    out of chiliish curiosity- Thc pitchcr had [risscd                                        leasewas a highway catastrophein which a
      her skull bY inchcs.
                                                                                               man irnd his wife wer-ekillcd in a head-on
                                                                                               crash with another car crowded with teen-
    Everyone is heset at times with destructive
                                                                                               agerswho haclbeendrinking becr.The man,
 impulses, sotletitrre$ dircc:ted inwardly'
                                   ilgainst oth-                                               it turned or-rt,    had not climmedhis lights,and
                      directecldilTuselyirgainst                                               the tcen-age driver was teaching hirl a
 ers.and sornctimes
                                                                                               Icsson.lt is obvious (as in this cxample)that
 the whole errvironment' ln the latter in-
         (rneor Irorc innoccntbystanders    may                                                an aggressivc         impulse rrray rcsult in loss of
                                                                                               life or clestructionwhich is wildly inap-
 be  involved, Thesc impulsesare, of course,
                                                                                               propriatc in terms of thc stitnuluswhich scts
 under tnuch trcttcr co(rtrol in some individ-
                                                                                               t h e i m p u l s ei n m o t i o n .
 ualsthan they rrrein othcrs.They are largely                                                                -Jost't MrclvBt, M.D', M.P'H''
  uncotrsciotts, in many instances degrec
                hut                    a
                                                                                                                                          pp. 67-69.
  of conscious awarencsl;of thcur can be

     In closingthis chapterit is appropriatc to Point out that psycholo8ical
                                             Btiih grcaterdcpth a1d greaterbreadthare
  accidents barelyscriltched sLir|acc.
           has                   tire
   PSYCHOLOCICAL APPROACHES                                                                                                            383
   neededin addition to the conceptual,  contextual,and rnethodologic    sophistication we
   have emphasized. is n't usuailyadcqunte, example,to tiivorcepsychorogical
                        tt                       for
  variables  frorl the characteristics thc physical,
                                      of            social, anclculturirlenvi16nnrcnls  in
  which they,maybe operative. adcjition,
                                   In         many virriables   that may be crucialto our
  understanding the dynatnicsof somc types of acciijcnts-for example,attitucics
  conucctedwith dashboard-tlountcd        ikons and other religious articlescarried in
  travel, fatalistic attitudestoward environnrental
                                                  hazardsin g*cnc-r.al, dilTerences
                                                                        arr{l           in
  attitudes_ toward physical and chemical (as oppi:sccltc, diologic.al)hazards_have
  scarcely  beeninvestigated. a time when muc-h
                                At                  psychological    iesearctr trivial and
  repetitious,the accidentarea offers nrrny possibiliticsfor' trc* anclsignilicant
  in a much broadcr context.


   I, Thorndike, R. L., Tfu Hurnan Factttrsin ,4r-                                      E. Miller, R. W., "Distant Visual Acuity Lo$s
       cidentsv,ith Special Rt'ft:renc(r Aiftrali AF-      to                              Among JapiineseCranrmar School Chil-
       c r ' r / e n t sP l o j e c t 2 1 . 3 0 - 0 0 1 , R e p o r t t , t J . S .        dren: Thc Roles of Heredity and the En-
       Air Force, School ol' Aviirrion Mcdicinc.                                           virtrnment," Journal ol Chionit' l)rlreasts,
       Randolph Fieltl, Tcx., t95t.                                                                       ^
                                                                                           / 6 : 3 1 - 5 41 9 6 3 .
  2 . M c F ' a r l a n d ,R , A . , M o o r c , R . C , , a n d W a r -               9. McClcarn, E,, and Rodgers, A., ..Dif-
                                                                                                           C.                  D.
      ren, A, 8., Human Variahles in Moror Vehit.le                                        fcrencesin Alcohol Preference         Among ln-
      At'tidutls: ,4 Review <tl'tlrc Litvaturtr, Har-                                      bred Strainsof Mice," Quartt,rltJtlurirr! of
      varcl Sctrool of Public Health. 1955.                                               Studir.s Altohol, J0:691
                                                                                                       ttn                 -695, I 959.
  3. Golrtstein, L. Ct., Rrsrurrh ttn Human Vati-                                     I0- Read,J. H., Bradlcy,E. J., Morison, J. D.,
      ahlt:; in Su[t Motor llehide Operation; A                                           Lcwall,D., antl Clarke,D. A., ..TheEnidenr-
      Corrtlational              Sutttmary of Pretlirtor                 Vari-            iology and Preventionof Tratllc Aciiden(s
      ahtcs and (rircriou                   Mea:;ur$. The Driver                          Invtrlving Child Pedestrians," (unadian
      Behavior Research Project, Thc Ceoree                                               Mtdit:al A:t.vociation      Journal, g9:697_701.
      W a s h i n g t o n U n i v e r s i r y ,1 9 6 1 .                                  I 963.
J a . C r L : s s w c l l ,W . I . , , a n d F r o g g a t t , p , , T h e           ll. Congcr,J. J., Gaskill,H. $., Cl+il, D. D,,
      Caustttionof Bus Drivt .4L:t,id(nIs.                       Putrlishcd               RLtincy, V., Sawrey, L., and Turrgll,
                                                                                                       R,               W.
     for the Nullield lrrovincial Hospitals Irust                                        E. S., "Pcrsonirland InterPersonal           lrirctors
      by Orlirrd Universily Press,1983.                                                  in Motor VehicleAccidents,"Ilc Aneritan
 4, Haddon, W., Jr.,                   "The
                                                 Prevention of Acci-                     Jownal oJ Ps,ychiatr.1,,     /1J:1069-1074,     t957.
     tlents," in. ,'I'txthook of Prevt,n.tivtMtditint,                               l?,Haddon, W., Jr., 'tAlcohol and llighway
     eds. I). W. Clark i.LnrlB. MacMahon. Bos-                                           Accid€nts," in Alcohol and Rood TralJic,
     ton: Little, Browrr & Co. ln press.                                                 Proceedingsof thc Third International
5. Harrcr, C. F., "Lrse of Fsychological Inven-                                          Conference Alcohol and Road Traffic.
     tory in Writinq lnsurance frrr Youthful Malc                                        British Medical        Association. London. 1961.
    Drivers," 7 rallic Satety Rt:.teartft Review,                                   lJ. Michaels, M., "Hunran Fcctors High-
                                                                                                         R-                         in
    7:l ;5-9, 196-1.                                                                     wrty Sirfety," Trafffu Quarrcrll,, t5i586-599,
6 . Jacobs, H. LL, Anal.ytkal ttnd Mathemarirul                                          t961.
    Slndres trl Atcident Crllr.s{rlidnr Problents in                                14,Smith,H. W-, and Popham,R. E,, ..Blood
    the Stlu:tion tl' Indivitlualt on the llasis of                                     Alcoliol [.evels Rclation
                                                                                                               in         toDriving," Ca-
    Atcidcnt Behavior. Dunlap and Associatei,                                           nadiunM edical Assotittt Journal.6.f:325-
    Stamford, Conn., I 955. Privatcly distributed.                                      3 2 8 ,t 9 5 1 .
7 . M o y n i h a n , D . P . , ' t T h e I . e g a l R e g u l a r i o no f '      I5. I-ucas, Kalow,W., McColl,J.D., Criflith,
   Automobile Design." in Passargcr (lar De-                                            B. A., and Snrith, W., ,'euantitative
                                                                                                                   H.                    Stu-
    sign and Hi{htrd-l.S'a/i"ry. Association for the                                    dics of the ltelationship .Between          Alcohol
   Aid of Ct'ippled                 (lhildren, New York, irnd                           I.evgls6n4 Motor VehiCle         Accidents,"pro-
   Consuurers Union, Mount Vcrnon, Ig62.                                                Ctcdings the Secondl[ternational Confe-
  p.2t2,                                                                   renceon Alcohol and Road Traffic.Toronto;

Shared By: