Docstoc

135

Document Sample
135 Powered By Docstoc
					     Time to Party:
 A Comparative Analysis of
Holiday Drinking and Driving
                           by
               Steven A. Bloch, Ph.D.
         Automobile Club of Southern California
              Hee-Choon Shin, Ph.D.
     Institute for Survey Research, Temple University
               Susan N. Labin, Ph.D.
     Institute for Survey Research, Temple University


        Presented at the ICADTS Conference,
           August 2004, Glasgow, Scotland
Background:
• Many local/Federal agencies have programs
  to reduce DUI on special occasions
• Data amply document that drinking and
  driving levels increase on holidays
  – In California and the US, the largest numbers of DUI
    crashes were reported on July 4th and Labor Day (2000,
    2001)

• Despite the importance of this data, there is
  little systematic research on the effects of
  holidays/special occasions on DUI
Research question:

 How do holidays and special
 occasions compare in terms
 of levels of drinking and
 driving?
               Research Issues in
               Comparing Holiday
                    Crashes

• Holidays often fall on different days of the week
• Drinking and driving patterns change monthly/
  seasonally
• Holidays and special occasions last different
  numbers of days
• Holidays, esp. single day holidays can have few DUI
  crashes
              Findings of Previous
             Studies of the Impact
             of Holidays on Crashes


• Arnold and Cerrelli (1987) examined the effect of
  holidays on all traffic fatalities (not DUI). Found:
   – decrease in fatalities on Christmas (-11%)
   – increases in fatalities on Memorial Day (32%), July
     4th (37%), Labor Day (35%), Thanksgiving (19%),
     and New Year's (64%)
             Findings of Previous
            Studies of the Impact
                of GDL on DUI

Limitations of Arnold and Cerrelli’s study:
• Did not examine DUI crashes, only total fatalities
• Statistical significance established a priori as
  “deviations >10%” from normalized fatality levels”
• Not analyzed by hour of the day
                 Findings of Previous
                Studies of the Impact
                    of GDL on DUI

Redelmeier (2003) examined crashes, including DUI
  crashes on Super Bowl Sunday (1975-2001)
• Determined that, in the hours after the Super Bowl, there
  was a 41% increase in fatalities/injuries nationally.
  "Alcohol-positive" fatalities increased by about 80%
  during those hours, A-R injuries increased slightly less.
• Limitation: Did not systematically compare Super Bowl
  crash mortality to that of other holidays/special occasions
 Method

• Data source: California Statewide Integrated
     Traffic Records System (SWITRS)
• Time period: 1994-2003
• Crash types
  – Had been drinking (HBD) fatal and injury (F/I)
    crashes
  – HBD F/I crashes as a proportion of all F/I crashes
  – “Daytime”: crashes: 4am-5pm; “nighttime”
    crashes: 5pm – 4am
 Method

• Evaluation approach
  – Fourteen holidays and special occasion
    periods (1-4 days long) are examined
  – Control variables for day of the week and month
    of the year are added to log-linear models (for total
    number of crashes) and to logistic regression models
    (for the proportion of crashes that are AR)
  – Estimators can be directly translated into odds ratios
    of over-representation since they vary directly as a
    function of ex
             Table 1: Estimators of Holiday Alcohol-Involved Fatal
                    and Injury Crash Over-representation:
                   California, 1994-2002, Nighttime crashes
                                                                                                               # Years Over-
Holiday         1994      1995     1996     1997     1998      1999    2000     2001      2002     Over all Represented
New Year's       -.14     -.19+    .79***   .96***   .56***    .32**    .13     .78***   1.33***    .37 ***         6
Super Bowl      .39***    .23+     .41***   .48***    .28*     .36**   .37**     .16     .45***     .34 ***         8
Presidents       .06      .20*     .26***    .02      .01      .16+     .09     -.07      .12       .10 ***         3
St. Patrick's   -.80***    .22     -.24+    .37*     -.50*     .38*    -.10     -.05      -.14      -.13 *          2
Easter           -.09      .07     -.14      .26     -.31*     -.05     .18      .02     -.63***    -.36 ***        0
Cinco de mayo    -.04     -.02      .03      .06      .18      -.08     .01      .08      .08       .08 +           0
Memorial        .25***    .28***   .16*     .17*      .12      -.04    .24**     .00      .11       .15 ***         5
Independence    .35***     .14     .50***   .50***   -.23*     .03     .60***   .83***    .04       .14 ***         5
Labor Day       .33***     .02     .12+     .17*      .02      .18*    -.06     .15*      .19*      .05 +           6
Columbus         .18*      .04     -.07     -.13      .17*     -.07     .05     -.13      -.13      -.06 *          2
Halloween        .13       .00      .10     -.04     -.44***   .06      .00     -.02      -.06      -.14 **         0
Veterans         .20+      .10     -.00     -.09      .08      .33**    .14      .11      .18       .19 ***         2
Thanksgiving     -.03     -.14+     .06     .17*      .05      .01      .02     -.23*     .00       -.01            1
Christmas        .06      .23*     .65***    .05     .44***    .23*     .00     .47***   .73***     .23 ***         6
          Table 2: Estimators of Over-representation of Holiday
             Alcohol-Involved Fatal and Injury Crashes as a
          Proportion of All Fatal and Injury Crashes, California,
                       1994-2002, Nighttime crashes
                                                                                                                # Years Over-
Holiday         1994     1995      1996      1997     1998     1999     2000      2001      2002     Over all Represented
New Year's       .16      .01     1.04***   1.16***   .91***   .49***   .46***   1.24***   1.78***    .69 ***        7
Super Bowl      .37**    .44**    .45***     .30*     .36*     .25+      .17      .19      .51***     .34 ***        7
Presidents       .01      .03      .22**     .02      -.05     .20*      .11      .03       .01       .07 *          2
St. Patrick's    .15     .37*      .22       .40*     .40+     .53**    -.15      -.13      -.27      .11 *          4
Easter           .04     -.25+     .24+      .29       .25     -.01     .36*      -.02     -.30+      .05            2
Cinco de mayo    .01     -.13      .06       .06      -.17     .44**    -.00      -.02      .03       .02            1
Memorial        .26***   .20*      .08       .17*     .19*     -.03     .24**     .17+      .21*      .18 ***        7
Independence    .13**    .72***   .46***    .47***    .27*     -.02     .71***   .85***    .57***     .44 ***        8
Labor Day       .35***   .18+      .12       .24**     .14     .30***   -.00      .17*      .18*      .18 ***        6
Columbus         .06     -.06      -.01      -.14     .18*     -.05      .08      -.01      -.13      .00            1
Halloween        .07     -.14      -.16      -.19     -.12     -.08     -.06      -.26      -.19      -.12 **        0
Veterans         .09      .08      .05       -.04      .02     .42**     .06      -.03      .24+      .11 **         2
Thanksgiving    .30***   .17*      .19*      .01       .05     .29**    .20*      -.06      .13       .14 ***        5
Christmas       -.18+    .56***   1.10***   .85***    .87***   .58***    .19     .83***    1.25***    .58 ***        7
          Table 3: Estimators of Holiday Alcohol-Involved Fatal
                 and Injury Crash Over-representation,
                 California, 1994-2002, Daytime crashes
                                                                                                             # Years Over-
Holiday         1994     1995    1996     1997     1998      1999     2000     2001      2002     Over all Represented
New Year's      .54***   .33*    .70***   .90***   .95***    .88***    .16     .39+     1.62***    .60 ***        8
Super Bowl      -.13     .05      .13     -.08      .15       .03     .2424     .11      -.28      .05 +          0
Presidents       .16     .34*     .07     -.12      .12      -.04      .09      .02      .18       .10            1
St. Patrick's   -.10     -.52    -.26     -.14     -.2.02*   -.98+    -.16     -.09     -.54+      -.26 *         0
Easter           .23     -.18    -.08     -.45      -.14     -.03     -.05     .36+      .16       -.21 *         1
Cinco de mayo    .27     .17      .21     -.54*     -.22     -.43     -.14     .46**     .17       .10            1
Memorial         .14     .25+     .06     -.06      .01      -.09      .06      .02      .35*      .09 +          2
Independence    -.05     -.03    .67***   .51**     .12      -.12     .42*      .14      -.10      .13 *          3
Labor Day        .06     .35*     .12     .31*      .23+      .17     -.17      .13      .18       .09 +          3
Columbus        .25+     .19     -.05     -.22      -.09      .10     -.33+    -.15      -.03      -.07           1
Halloween       -.03     -11      .01      .25      -.02      .08      .09      .10      -.14      -.00           0
Veterans        .53***   -.05     .17      .29      .27      -.46     -.10      .06      -.02      .15 *          1
Thanksgiving     .16     .20      .12     -.10      -.12     -.38*     .06     -.01      -.00      .01            0
Christmas       .52***   .49**   .66***   .63**    .69***    .34+     .54***   .74***   .95***     .59 ***        9
          Table 4: Estimators of Over-representation of Holiday
             Alcohol-Involved Fatal and Injury Crashes as a
               Proportion of All Fatal and Injury Crashes,
                 California, 1994-2002, Daytime crashes
                                                                                                                  # Years Over-
Holiday         1994     1995      1996      1997      1998      1999     2000      2001      2002     Over all Represented
New Year's      .94***   .60***   .92***    1.06***   1.29***   1.15***   .50**    .75***    1.96***    .91 ***        9
Super Bowl      -.20      .23      .09       -.33      .20       -.13      .06      .11       -.24      .01            0
Presidents      0.19     .24+      .06       -.11      .09       .03       .14      .16       .12       .11 *          1
St. Patrick's   -.03     -.34      -.25      -.15     -1.93+     -.87     -.18      -.16      -.67*     -.33 **        0
Easter          .37*     -.47      -.14      -.42      .02       -.00      .11      .30       .12       .02            1
Cinco de mayo   0.34      .19      .12       -.59*    -.56+      -.38     -.05      .39*      .14       -.00           1
Memorial         .19     .25+      .08       -.07      .11       -.03      .14      .07      .50***     .01            1
Independence    -.10      .10      .60**     .49*      .21       -.20     .50*      .14       -.15      .19 **         3
Labor Day        .10     .32*      .15       .41**     .36*      .29+     -.08      .16       .20       .20 ***        4
Columbus        .24+      .18      -.07      -.19      -.03      .08      -.27      -.07      .00       .01            1
Halloween       -.00     -.20      -.23      .18       -.06      -.01      .05      -.05      -.25      -.03           0
Veterans        .49*     -.02      .28       .03       .26       -.39     -.15      -.02      .07       .09            1
Thanksgiving    .14*     .34*      .27+      -.19      -.06      -.13      .25      .00       .14       .11 *          3
Christmas       .44**    .85***   1.06***   .84***    1.10***   .76***    .87***   1.13***   1.42***    .87 ***        9
 Summary of                             Results
• Nighttime AR F/I crashes concentrate most notably during:
   – winter holidays: New Year's, O-R=44%; Christmas, O-R=26%;
   – summer holidays: Memorial Day, O-R=16%; Independence Day,
     O-R=15%; Labor Day, O-R=5%
   – during/following the Super Bowl, O-R=40%

• Nighttime AR F/I crashes as a proportion of all F/I crashes
  show results similar to those above, except:
   – O-R is notably more pronounced on three holidays: New Year's, O-R
     from 44% to 100%; Independence Day, O-R from 15% to 55%;
     Christmas, O-R from 26% to 78%
   – St. Patrick's Day O-R rises from -12% to +12%
   – Labor Day O-R rises from 5% to 19%
   – highly protective effect of Easter, O-R= -70%, declines to a non-
     significant +5%
  Summary of                             Results
• Daytime AR F/I crashes concentrate most notably during:
   – Winter holidays: New Year's, O-R=82%; Christmas, O-R=80%;
   – Summer holidays: Memorial Day, O-R=9%; Independence Day,
     O-R=14%; Labor Day, O-R=9%
   – During/following the Super Bowl, O-R=5%

• Daytime AR F/I crashes as a proportion of all F/I crashes
  show results similar to those above, except:
   – O-R is notably more pronounced on four holidays:
      • Winter: New Year's, O-R from 82% to 148%; Christmas, O-R from
        80% to 139%
      • Summer: Independence Day, O-R from 14% to 21%; Labor Day
        O-R from 9% to 22%
   – protective effect of St. Pat's Day O-R drops further, from -23% to -28%
   – protective effect of Easter, O-R= -19%, declines to a non-signif. +2%
Summary of                           Results

· Despite concerns about drinking and driving on holidays
  like Halloween and Cinco de Mayo, the days pose little
  problem. St. Patrick’s Day has only a slightly elevated
  problem. (+10% increase in AR F/I crashes as a proportion of all F/I
  crashes)
· Most holidays and special occasions show similar increases
  in drinking and driving during daytime and nighttime
  hours. Two holidays, however, differ notably from this
  pattern: Super Bowl Sunday and St. Patrick's Day. In both
  cases, it is clear that excess drinking is a nighttime, not
  daytime, phenomenon.
                    Conclusions
• The longstanding concern with excessive drinking and
  driving during the winter holidays is largely well
  founded. Christmas and New Years are among the three highest
  DUI holidays of the year. Thanksgiving, however, should not be
  included in this group.

• Recent interest in drinking and driving during summer
  holidays is also well founded, although some of the
  rise in DUI rates for summer holidays is likely
  attributable to the general rise in summer drinking and
  driving. When summer DUI levels are controlled, summer holiday
  drinking and driving, while high, is not quite as excessive as winter
  holiday levels.
                Conclusions

• The Super Bowl is of particular concern.
  Nighttime F/I drinking and driving crashes on
  Super Bowl Sunday are not just at the second
  highest level among all holidays, but are an even
  more predictable phenomenon year after year.
  – In eight of the nine years analyzed, AR fatal and injury
    crashes rose significantly on Super Bowl Sunday night.

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:1
posted:8/27/2011
language:English
pages:18