Survey of Income and
Analyzing The Characteristics Of Blacks: A Comparison O f Data From SIPP And CPS
Reynolds Farley and Lisa J . Neidert Population Studies Center University of Michigan
April 1989
Presented at the 1988 meetings of the American Statistical Association, Social Statistics Section; New Orleans, Louisiana; August 22, 1988. The views expressed are the authorsf and do not necessarily reflect those of the Census Bureau.
TABLE OF CONTENTS SECTION ONE: SECTION TWO:
Introduction................................ Using SIPP Data to Study BlacksSample Size Considerations..
1
................
2
Table 1.
Counts of Observations from March, 1985 CPS and Contemporaneous SIPP, for Persons Classified by Race, Sex, Region and Age.....
................-..
..............
3
Table 2.
Counts of Observations from March, 1985 CPS and Contemporaneous SIPP for Persons Classified by Race, Sex, Educational Attainment and Age Counts of Observations from March, 1985 CPS and Contemporaneous SIPP for Persons Classified by Race, Sex, Employment Status and Age.
5
Table 3.
.............
6
Table 4.
Counts of Observations from March, 1985 CPS and Contemporaneous SIPP for Persons Classified by Race, Sex, Marital Status and Age
.................
8
SECTION THREE:
A Comparison of Marital Status, Educational Attainment and Labor Force Status..............................
9
Table 5.
Marital Status by Race, Age and Sex; from March, 1985 CPS and Contemporaneous S P . . . . . . . . . . . . IP............ Estimated Years Spent in Each of Five Marital Statuses as Individuals Age from 15 to 64, March, 1985 Current Population Survey and Corresponding Data from S P . . . . . . . . . . . . . . . IP...............
10
Figure A.
12
13
Educational Attainment.................................... Table 6. Educational Attainment by Race, Age and Sex; from March, 1985 CPS and Contemporaneous SIPP........................
14
15
Labor Force S a u . . . . . . . . . . . . . . . . . . . . tts.................... Table 7. Labor Force Status by Race, Age and Sex; from March, 1985 CPS and Contemporaneous SIPP,.......................
16
Figure B.
Estimated Years Spent in Each of Three Labor Force Statuses as Individuals Age from 15 to 64, March, 1985 Current Population Survey and Corresponding Data from SIPP..........,........................ The Determinants of Unemployment and Earnings: A Further comparison of SIPP and C S . . . . . . . . . . . . . . P.............. Analysis of Determinants of Unemployment in March, 1985: A Comparison of SIPP and the March, 1985 C S . . . . . . . . P........ Estimates of Unemployment Rates for Selected Groups; Data from March, 1985 CPS and Contemporaneous SIPP........... Regression of Monthly Earnings on Predictor Variables; Comparison of Results from March, 1985 CPS and Contemporaneous SIPP for Persons 25 to 6 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
SECTION FOUR:
18
Table 8.
19
Figure C.
21
Table 9.
22
Cnlso........................ ocuin........................
Figure D. Estimates of Monthly Earnings for Selected Groups; Data from March, 1985 CPS and Contemporaneous SIPP...........
23
24
BIBLIOGRAPHY
ANALYZING T H E CHARACTERISTICS OF BLACKS: A COlMPARISON OF
DATA FROM SIPP AND CPS
SECTION ONE: Introduction
Although data from the Survey of Income and Program Participation (SIPPI provide extensive information about the characteristics of blacks, we have yet to determine whether the sample size permits a detailed analysis of racial differences, or whether the characteristics of the black population, as estimated from SIPP, correspond to those same characteristics as measured in other demographic surveys. To explore these issues, we studied data from the fifth and sixth waves of the 1984 panel of SIPP, and compared them to data gathered in the March, 1985 Current Population Survey (CPS); that is, the Annual Demographic File. We selected those rotation group: from the two waves of SIPP which were interviewed during March of 1985, while the interviewing for the Current Population Survey was done primarily during the first week of March. This report compares SIPP and CPS data. (For information about the design of SIPP,see David, 1985). The first section provides information about the sample size for blacks from one wave of
SIPP. This will allow potential users to know whether the sample is sufficiently large to permit
the testing of hypotheses about racial differences. Using unweighted data, this report presents sample counts for blacks and whites classified by age, sex and region of residence in
March, 1985.
Many users of SIPP data will likely be interested in the marital status, labor force status and educational attainment of adults. To provide additional information about the sample sizes for the population classified by these characteristics, persons age 15 and over were categorized by race and sex into four age groups. For each group, we present information about the sample size for five marital status, three labor force and five broad educational attainment catcgorrcs. For comparative purposes, we show similar sample counts from the March, 1985 CPS. Although unweighted counts are used in this section, we realize that most investtgators will analyze
weighted data, or will attach adjusted weights to each observation so that the weighted sample counts are equal in size to the actual sample sizes, a procedure which may lead to more nearly appropriate tests of statistical significance. The second aim of this report is to determine whether the characteristics of blacks and whites, as reported in SIPP, correspond to those reported in CPS. We focus upon three important variables: marital status, educational attainment and labor force status. In this section, weighted data will be analyzed. The final component of this comparison of SIPP and CPS data seeks to determine whether demographic and socio-economic relationships are similar in the three different sources. Two types of relationships are explored. First, a model was fit which took the log-odds of unemployment for labor force participants as its dependent variable. Independent variables were age, educational attaimitent, marital status and region. These models used the weighted sample data and were fitted separately for four race-sex groups, but were restricted to respondents age 25 to 64 in March, 1985. Second, using the same groups and a similar age range, a model was fit which related a logged function of monthly earnings - for those who reported earnings - to educational attainment, age, region and place of residence. A major difference between the two data sources is that SIPP reported earnings for a-one-month period at the beginning of 1985, while CPS reported an individual's earnings for the entire year of 1984.
SECTION TWO: Using SIPP Data to Study Black. Sample Size Considerations
What is the sample size of blacks available from one panel of SIPP? In March, 1985, when the sample size in SIPP included approximately 18,000 household, information was obtained from a total of 5,156 blacks and 37,686 whites. These numbers are shown in Table 1, along with a classification by age, sex and region.
[Table 1 1
-
There are two ways to consider the sample size for blacks in SIPP. From one pcrspcctivc, the number of observations is quite large, much larger, for example, than the sample size for
I I I I I I I
fio ry
o U
=
? at ql
+, n o
ct @ -.t
-
(o F nt F
E ;" 3 o u f - :
c'{6r
*a'
g -
s..g
XI::to6{F
o' O (9 "
. o e c tSi t I' F F f i i ln |o , c0.a'.(o-:e'.:(9.c'. gqttelc{---
g g @
qo
O .oo,6r.oFoo.o' =_t_ioririr.-;;.do;Jj-
c! qt F (' t, c{
N|n
ta-
tr,
g :
()
n g o o
o tr
o o
)
o a
:l;:;;H3is*
t.:9f:qr(,loFsro s cr G p - F ",
E;R:988;R
eeoooF-51 3.S.9;hgs;, 9 9 9 o r 8 @ o l ' ' l r n Eq$FgsF$e rgHgilEES;s -.-..
.rd;a;ddJjj 3c,(r|(r,o,flr\|e{
{F(o!r9oO-e,
cooo'oFonc'r
tn
'x
& c0
s
L
ry
o o
3
scr{r6;;-d
'e
! .!qqcgri'r .d : ; ;{ 6 o i : o o .
;
an o'
o oL
|5
u
=
+,
3
or
tE
f ,rf u E F ifr l rEo?iR- E o ; 6 b, i i
r...r.i.iciiJi
EF @ o; S o FBRg$R F E n' S o, t
id.i;;-:;;
U
ru
:t
{lt
;
o
a
t t
I I I I I I I I I I
oo. t< o o
o s +t o z
]n
q J F
g
RsE3385s' q$n$*E:3' g.iq$:H;$F il:ilpFgFRfl L c'ttot??'
o
U
'oCrnaa{(rt-?e{
-
o o c o L o o E o +t c o (J o c o
a o. C' n @
f
q9olOrt.loOtot9t9 oerolo)o!-otstct crcrgroF(oln&fi
6
,,J
o
F
o ul
J
fr... F
I
s (g (9tf Qqp gt q F fi l ol ( \| O O g1 t't O Orctllurctotoa' a a r a l a a . - t a a a
a,99 ? lillF 1919 c, n 6p 6 +- ot O t\ n 61 6 ui;o)Fnl!relflq
I
A Ul J
Cl- rF
tF
QtO(\|n(oc0oorn 9q!Fsrf)-Fo)n| o)ocuFr000stai
a a a a a a a a
o c o o (U
f
Lf
{! L
o
U
a
A
(|t6;(F.-.rD
_ll|
?UD?C.f(DFFCO l:---
u
o o 3
+t J o
I F
-s r
IL
Y,
C')ofs|GtFCl-Cto tP!F (9UtfilCt?rF
lu
IL
X tt H
{ A
l.
;.
= I FglOolfif r-ottrF rniotrtottatFai e,
€s:\r?giqr-:O r-qtclfilf{r-?$f nt
lt{ E -
tr|
(')O[fit-oO(^or\(9 Fro-UitOqv-i F€o(o(ptn(l'crlr
(!,
.a >o [fr
)).r
\no o
Or
C16 +a (t r-
LtL J
l.
o
R;s;8FP3* atrr6t-F
|o l,'Fut ?F o ot G. o ro o to r tr a et tto. o. .n
H
:il[E$x833
tnseid;;_:
fiEiE!EEF5 o6br6o66in
ro tgg g n t\|G| oto S tO6 6; id ul-- a, : c. 6 . U X r o S + V ' !
g58SR38;;
err'r000oro6
.c
(J t-
=
L r.-
o
E o
g: - qr o ct or c'o FF -o 6 n'i;t s ro?--.
o.o oc o-(s o. +. c. oc
LO
3-
? !! !: tt trt osr or ct to tgq, fl - . o . g | . o ro ol 'o ioru, lou s ri"
lnF
(-:
o.
J# (J((, {r(J
A*t
gl
r F O
c o
+to
o<
LO
!!5DSg;9tt?
5 i38" 3 YY i r ; l j - ' 3v v9 9 *
oc oqt o oc o lF+ oo o ou, ? c. =x oo ov)
.oc'.,o",.o,"uro 99:.oroerr. ;383398;ct
R3gI3=98clt,
o(' ()a,
OE tr(6
#L
CL
'.-o o
o. t dto oc L6 .rt
OL
gP[$:!EE9 T.F$HHHgR: FSEE$3Pi9 frBg:HHg$3
$r or d;ddcijiJ OtCl't(!*rol
cgo .c tn(J .c lJ -r
ttJ J
o
Eiii,r"r;
a=x;;;38 Fo-c{elt?to@ro
J
qs.f?s? +fi6U
_i.s9s'\rrrn ;o:;:tt+fiUb
F
F:lll[[fir Siliiiii*
Fo;ilF
;;iijlfi*
r
ono>
!-| Qrfltlt(pl-VltF!-'Fnl t\
q v1 o ogfi L o o L
u L9on
€
F
r.r
La
L o o n L Satrto og-^AiI -rionitr;
f-r I t\ ci-(\tcl(t,."r-vt;;Int
L n O a a {EL>-LIE il,6 qo _riot.no;
F? | i-c{t?o_ Fvt?.-?.,\l t\
L m o n L toL>-Lrq oo oo J)-ono>.
| f_* Q_c{(?(!* |,_VfF?!.nl f\
U1
o o
UJ (J g
o
tn
o
t \
(! ct
?
ln (9
tJ)s
e{(9-uJ (9(o o)
-(!lCtn er (9 !o oc0
coos!f os\rF (oo €
a a
(l:o(9F C.{ rf Oti lt, f\n
a a
-o
trl
(9n
oo
coa!8(.0 o!t00rn Ortflg,
o a
3
f.-
n o r0
o n
rn(9rntr c)o)(oF
Fqt
ir
(?
FG(OO F(Ir(?*
l'f\
.F
f\
NFoo tr'!nto (D(rttr0O
a a a
tE
Oto
(F
c,
!r (9
rF
f\
()
o L o o. L
o c o
o o.
(J
lr q
n n|
I
gtnF(9 !OCl?C)
a a rt{F
s|rnor
$ltCttt (0Uf?o)
a a
\rotfir,
ir_ (F A, Ot-iFrF
a a a a
$t(olfilO
Ftl'Ctn Oooo(t) €o|Jnct
a o -
I I
f{=
rrF?h| (!f +
F(Y glr
O
r5
o
tE (l I
o. o.
FI
n
J .xo
<-E F O+r OreCO F<22
Lp C O Q.r
L
F
o o o o o
*
lt\
E t!
)r _t .gO g Lp J O (l .r <3 E F O+, o+rco F<22
o tro oo
J
o lt .o oo
h-l J(o Lp C O Q,.r
<38 tO+, o.+Jco l-<=Z
t-r C J O tl"* <=E lO+, o+rco l-<32
o otl oo h-l so
v7
UJ C.,
c
o
UI
I
not work during the month, and spent one or more weeks looking for work or on layoff, were classified as unemployed. The "not in labor force" category includes persons who did not work during the month, and did not seek employment. Note that a person who sought a job for several weeks, and then got one is classified as employed. The CPS data are also a mixture of monthly and weekly data. Employed persons are those who worked during the week beiore the interview. Unemployed persons did not work in the week prior to the interview, were able to accept employment and made some efforts to find work during the preceding month. There is a slight difference between the SIPP and CPS classificat~ons. A person who worked for three weeks before the interview, and then spent a week on layoff, would be classified as employed by our definition with SIPP data, but unemployed with the CPS data. The sample size for blacks in SIPP will sustain a study of labor force participation, employment or the rate of unemployment. However, a specific cross-sectional study of black unemployment will involve a small sample size. A total of only 325 blacks were classified as unemployed in March, 1985by SIPP. In this section, we are analyzing unweightcd data, so it is inappropriate to estimate rates of unemployment or labor force participation from the numbers In Table 3. However, the actual counts of the unemployed are quite high in SIPP relative to CPS. That is, given the difference in the number of households sampled, we might expect the sample counts in SIPP to be about onethird those of CPS. For both black and white women, the sample number of unemployed was relatively large in SIPP compared to CPS. [Table 4 1 The final table in this section classifies SIPP and CPS respondents by age, race, sex and marital status as of March, 1985. Again, we observe that the sample size for blacks is sufficient for many investigations, although there are fcw male black widowers and, for most purposes, it will probably be necessary or desirable to pool the rnamcd-spouse-absent, the divorced and tile separated populations since the numbers In these spec~fic manta1 statuses are few.
B
a
o
i
o
G g
+ !n (o
l-C)OOt|f\ F(9totg Qt .F.rC(9
u
(o
!c
I
-'
OqtF(\f nct : @?e! tF ( \ t g F fq
Q--C{O19 Oq-!'3.,i ton Fai'
rt ot ttt ot nt e) noc0s@s
q,o!9
g o ,o
F
c,)cog)a(?tn S)r-c)m(^O
FQrf,1
a a a a
F?\
(0 C0
t1Q r-
c
trl
g :
o o {0
Q
n
st
QgOO61qr(g On-orFrF q!jr) ;C-:'
(9OOt'.Fc-tl
r\(9lo(og)o =.=. tr'cl t9n
co(0sfFCOrft Oo
:c{sons
sg)of
* ,r; =Toerorr O(oFfi'F(o
€€Ffl?Ui'
l a a -
()
F
o.
(J
\r I n
c{
rf
q!o)?cO(r)-
qq9o cog)
t}ro{n
f l o t c { 1-g g e : l d. oo
} O
r9gqfoDrf
oD
c{n(9nfir(') *Q(of')(otir OOOta a a -
fl?-
o c o u, L o o. o |!
o. o.
t{
?tn $l r'
c 1 fi t
Fg:t $lr
e)ol
tt t\l
n
I
cFFo-: o)o t 600 @. a. cl
(J m Ul r.< r.< o.==302
o (.)
!f
t? 'FtF
lO
r
cr
!F
(?
o. o'
i-
!E (?
I
rf,
(9 F(1r(9€(9 GO lr, '!' OT (9 -' $l ol
(Tt tv-q?o :(Y.. Fqr tfl '(9c{ ol
o.
F
tP
tf
ClO(7rC{(\|
F
qo'1\F(\I
-(g{F
r\f s,ct
-(9$f
a(o
n|. rr ofit
F
o}F
(\l
!o
{
o tn
t
fJ
.n o.
qp Blt\oos (rl
o)!n
F(Y C) .(gCf F @F
'et cr
. F
iF
c{FFoo coq
(O
oF r r !
m
o () |! & .a
o a
+J (t +) q
F-{
o. o.
L
n
(F
rf fl I
vr n o.
FoOlto P {,J C' (!+ -e ,Fg F ct
n
.P
fn$OF€ fit o}c{19 .!_co
co O f).SlF o
Ct
fs+.?ttr@
sf sl
u
q gt\€ (9
r\ t(o?rtlfll tl,....
-e
O
F i o
FFOact f{r-rF€o tf,
?(Y?c{q tf (l!..
O
tn{.,.|..H= o. EE=oZ
(J
F(\r\(9c0 !f -(9r{ g) t\
o
fl {o€@orc0 l^()tgfl-.coO
-tO?(\l
€
E
I
I
()
o-
A O H l.r
c. ==3(f:z
o. U1 tnr-rHE o. ==302
(J
11
c) (r)tnultll
o c.r otoF(t|v (? tr0
e)
(')
o) rv t\sf (')t0c,
?{o
Ft\C')o@
O
\J
t\c:t.)vc)
e.ttl)g.tftn q\f
a
?
?D(gFCOcO lf,lrtOOrs
F F
CX
fn
-
o} n lnurco(9n
(9
cDltCFV
n co F€onn
C?q
i!:
fr:Of-u)
: o) ro
ctt\tcrer€ (F(''€F
(!F
o (') €
l'toqsq nno}fo!f
(q
P
tJ
FctO-61
gQ-
o) ootE)roct
.F F(!tClOco et ,'- o€!?allr)
!l g{:
a &
L)
q
-
FO)t\t'tg
tFrF€OF rF {!} -
o)
f? (*) c t c t F f l t f , l . rV OIF
e'olo}fitv
I I I I I I
o. o.
lrt
(o m O
iD
?F(9('o) rtO-fitO
r r $ l p i -
o)(9Ot'-t: \J (!t O(!-ron
.qr|Dl-rFrF !F
rt lt,
I
vl
rtt \r
a c.
()
I
t't nO(DFtn -c{1-!-11, O
. ( 9 F < F i r F
(p I
z ul x (J
o
I
C) 3
z l! -
(')
o
fn131FFt-
t?tC$ng1
. V < F - ?
q-
ut roor-o
F \t rF l$+ Ulrrf F rFr-
o.
FI
J
Y (J
J E I
!o
c0 cosolGl
[ ) : r F t F
o
tDotFsct
t? tc
I
rt,
In (?
a CL (J
to {o s
.
f.-OUt€ct (p:tl) -
\f F
rF .f-
P \, o
_
'\Ooon n( c?, t ' o ( o r r
f{-
oc
Fl
rt ft
I
tn
$Oof : 1n.....-.. sf o)(o . (11 (\|
FF Ol= tO
c, ' F. $f
(tr=lpgf
19
Crq?-CfO (ttr F.,,
|n
il
t
I I I I I
a c .-{
!
n oCJ
c.{qco
' (\t (lt
I c2
FOG|qf\
19
st1g?,Fg) r-t,
tod
rit
E. (ut ; C j * * ! e{
o. ot{
o) 9-(0
r1
?-lnfitO-
o
tt c{
I lt,
.lD
v,
('' o)
aO.
(oo)OO(o I \,,
(') Ot s, (g
sr
F fl € t\ !tClClp6 .@ o}(\t -q)
{9
a o
I
c o
a o. (J
o} r0OO-nr
@
q ulc.{
st o}
9
OG of Fo-
c
o. o.
l{
o
f-
{!-(9oto
a
+- c}f\fDtnc) -c?\r o lf ':3
t ?
\,
.
o|@@(9o -cr)t-O(t (1 -g-c-)
a
o)(?rrfrlOf'l
COt\(})FCt
tJ')
trl Fl
o.
o
v1 a
(J
rF
'c)q CD
(9 c) O
.
ottnc:cao) r-coctOtp
(T F?,(?
o
d F{
(J
o.
o- E=302
{ . / |t f I H x E
(J o.
o.<(o> A.n H H E E=302
FIGURE A.
Estimated Years Spent in Each of Five Harital statuses 2s Individuals Age from 15 to 64, March, i385 Current Population Survey and Corresponding Data from SIPP
TOTAL POPULATION
CPS SlPP
WHITE MEN
CPS SlPP
1
CPS
SlPP
BLACK WOMEN
CPS SiPP
I
YEARS
SINGLE
hTv
MARRIED-SPOUSE
PRESENT
MARRIED-SPOUSE ABSENT OR SEPARATED DIVORCED
pTA WIDOWED
Using the information in Table5, we can test the hypothesis that the proportion in a p e n marital status category was the same in SIPP and CPS, that is, that the difference between the two proportions was zero. For example, the pcrccnt of black men age 15 and over who were married and living with their wife was 39.6 perccnt in SIPP and 38.9 perccnt in CPS. Using weighted data to obtain these percentages, but the actual sample sizes, and making the assumption that there were no design effects in either survey allows this test. That is, we get a pooled estimate of the mean and its standard deviation by assuming that each survey is a random sample. We tested to determine if the observed difference in percent married, spouse present is significantly different from zero, and concluded that it was not.
The SIPP survey found that 36.0 percent of the black women age 15 and over had never mamed;
CPS, 36.9 percent. Once again, using the actual sample sizes and the assumption of no design
effects leads to acceptance of €hehypothesis that the proportion never-married was the same in
both samples.
Educational Attainment [Table 61 Information about the educational attainment of the adult population is reported in Table 6 which refers to black and white men and women age 25 and over. Three measures of attainment are shown: the percent who reported completing high school, the percent who said they completed college and the median attainment of the age group. A look at these measures shows that SIPP and CPS give very similar views of attainment leveb. The maximum difference
between a SIPP and CPS estimate of the percent who finished high school was three points, and
involved a group with a very small sample size - black women age 65 and over. The greatest difference between two estimates of the percent completing four or more years of college was less than three percentage points and also involved a group with a small sample size: black men who were 55 to 64 in 1985. There was only one lnstancc in which the median attainrncnt oi a group as estimated from SIPP differed by more than two-tenths oi a ycar trom the CPS estimate and that involved black women age 65 and over.
L4
I I
rr@c? crg)o 'V-
o-fin ctr..c
FI
n F
r-(g(g
F
q) _
o)o)SI
.ti. F
c
c o{
. 9 - r F
+ tn (9
o ,n U
(9 c-{ (9$ls T €cCg)
o (
o)
f!-
roco g) c3 r Lcl c!
q l-
cofto (r)g)(o
€
o ao
:
c) ?
a
15 c o
q
Fqo ? olo)c\t
- S
(c
oi
o
C-'l F CC - rfi a?
nl'tm {.9
o
O{
-C{Q ot €?\r c-{
r? r-
g)net ? (?c{tC
-(\l
r) €
()
(9
O.
r\(?(j {? (9C)lc
a-
s c,J
F
l?
o)c)6
n ol r
s
I
-(9
.l-
-
UI
ol L1 (9aoq t\ (ol'lq .(O * c{ r (\t
or u)
o
c{u'o} .(9 ?
(o C) tf, ..{O -
(\GCOT
r
.F-
or
no)F o c g)$rrn
rf?
*{tt9
i') -Orf
. rif
o
{F
rF
o
tn
@
I
|.c
C.t
Sf !F
IF
g)
o)oF
'(g
c'r
r{
OIFO
cO-fCIo)C
rir $'l
F
gt\F
c
F N
-
. ( 9 -
g)rr(9 c?
-
o)
g)
cJFtf c? '.'F -t(r
o) O -
\ .c
E
C) l{ tlt
q
e
ti? :
(r)
(g
u:
I
I
i\('l(9 JF. c{ cc(9r\ tn
F|f)@ JF(\l
n
e{ F
Orqo |.rt ttf).F
o|
(9
$r F I
C{
o(.9€ o) €!n(\
-!\r--
(o (\
('!Ef\ TD Oc-{€
'lf) .t-
o n,
!F
r-
o
F
I
C)gO j,n
o
of
(0
c?
C-t 'F
t{ q.l
E o
tn
=
o
H
l-t J
z
.aa
FfiI(\ dtt !iFr? .F Ol g}
atir
cY -
lr.l
3
a \
cn
s s
'r/:
x o
o. e o.
J F F to
i{qo
o
d
(t o U|
("
u)
I
o
I
o (')
ci' !q('rtn 'O(\f
c
11 -
rrJ F
z E () =
lit
cj@? .t\=.F
c
n
(9
(\j
I
z ur o)
i\-
maro) ln O. - n o c t
o ol
?
(9 tfl c,
I
c)
=
I
FI rl'
On cr n
f\
s f?
F
tO'''€ f?
(? *
lrJ l-
C).-.?
= =
I
rs
(g
v (J
J E
I
z trl = (f =
Y U
J
g?OO .( 9 -
r)
r{ .r -
tn|'flc)
fs?u?
o t\ r-
(o
$l
r.i
ccg)F €g\n| ,tg -
rg'
-
o
u, C{ rp
'€o{ c: ---(9
a
I
o.
()
(9Qf. ,a sa t" tJ: .CC t\
o) $l iF
ro(9(.)
F (? u? ot '€e) (t
f')
fl) -
(\
@
ot! rF
lO (?
o ('t
lf,l -00(\|
(9F(9 volrg
-(9 IF -
(9 n|
tF.
rnfln
6etq
- F F
o
(9
${
, F
I I I I I I I I I I I I I I
,{
o o
o. v (?
I
nc'n o) nln(\,
.@
c (.)
rr*
tOC) fltn|,n
-€ t\
o
ctr
(\
n
(r')
g) e)
C{
r-
co$t(g sf ri t\ c\f
-€n| F
o c)
iF
*o(9 tn
$f (l) -Fc tf
0
(if
(')(') * |')
iF
l-O*
-(D-
(\
t7-
e\| f) C-l {F C.l $, !n $f g) c)
r| \ -
c\t
F c{
? fl
& o
c o F EA .i{
+,
+J
(t
tn c{
tn q, (J
rgFs. (9 |'r(9e, .€ sl o) c,
g) (\
ccn ( ) . .o t
c)t\(g
t\ 'CO $f
o
t')
tF
C') -C) f? nCOC) '((}s| F
.F
r)
(0-O
t(9Otf
.tf! ?
r$t
!F
c{
@TOF t\ @sf@
' F P
F sr
rf,
IE tt F
g} (? $f cgo (9
C-{
?
O(D* rt C) !t ('t -F hf g) lo $OC
F
co $t r-
(\rCrtC (9 (.9t?(9 .l- -
(9 c{
C(\|ftno.(g t-
tr) $,
iF
-{o(9
o rroo
t?
(r|
!-
ro (o
t*
(o
€
c) (\|
?
'lfJ
'(3
,-{ I gt
o
& (ttH +J cn
o
U:
tl
Ff?o)
. f \ . -
o
O-cl
t
to $l
F -
s(ts
. F ( \
o{
< F
r,ln(9
' F
(9 st ot tn Uf {D? {9
. if:, r-
c000 (.o
OIO- ( O F
s,
11
G
(J
o}
(?,
o) tn n o
t? P G lo il C' (i *t
I t r f , \ , J l r q 'v
In (o o o
t"
co
o
{J
;g n{ 6 X J rx8
.7t H
q3
,Y (g
o
o
6
*t
o
Q
:n t o .,
r!
t l - a \ I J .!it lt
(! =a ugf
i,
a - t 9 , ? t v r J , \ J
tt
pG OL
!,
c(,
,G OL
!
-€
f'1 r.J
?r
Q .-' L(0 (Jf, 1l
F G
(! f(n EO
:o
:a
a a -
{t .-,
rrJ
\,
l a -
o(9
rr.
1
| -
c(j
, 1 A. t t , Y
*; o {
fr1 Fl
UO ao)
L'O O[!+Oi-.
\.,
trl I
a(} Z O
uc:
Z
. ? - - a L
r4g}
a 9
OL o(5 (J(u a6|
|La-i
_r!
n . L./ t \.,
v
b t ? lr.,
r I 'J
,.|t L/
L2O U)O*O.-i.
: o)-
E{
zlqx
E=(J 7uJ
EI()
zxbq
)
c]
i.i.,:
!;'e €'?;r
=
J
C
t't
L5Z g E)C.-OFr
E ' l -b
/1r !,
tno|
f'r
v a ? - - F z
rr \-,
L.J
E
Z bqb{
=
tru zlqle
=
1].-OrE=L. )tr! Z bele
c: o-
O :E
There is no evidence suggesting that educational attainment was estimated from SIPP is consistently greater or smaller than when estimated from CPS. We tested and confirmed the hypotheses that the percent high school graduate for blacks age 25 and over was the same from both sources for men and women. Labor Force Status Information about labor force status in March, 1985 is shown in Table 7. Persons are classified as "at work", "unemployed", or "out of the labor force". Figure B shows the estimated number of years a person would spend in each of these three labor force statuses as they aged from 15 to 64 according to the labor force activities reported in SIPP and CPS. [Table 7 and Figure B 1 The estimates derived from CPS and SIPP differ slightly. The proportion of both men and women who were at work was higher in SIPP than in CPS reflecting, perhaps, the slight difference in definition of employment used here. (For a comparison of SIPP and C E labor force concepts, see Ryscavage, 1984). The proportion who were out of the labor force was higher in CPS than in SIPP. Such differences are not substantial but they were consistent. According to CPS, if a white man experienced the labor force rates of March, 1985 he would work for 39.4 years and be out of the labor force for 8.2 as he aged from 15 to 64. According to SIPP, he would work
40.0 years and be out of the labor force one-half year less; that is, 7.7 years. For black men, the
difference between the CPS and SIPP estimates of being out of the labor force were even larger. According to SIPP estimates, a black man would spend 11.5 years out of the labor force; from
CPS, 13.6 years. When we compare the percent of black men - or black women - out of the labor
force as reported in CPS and SIPP we cannot accept the hypo thesis that the difference equal zero.
SIPP appears to estimate a significantly lower proportion of blacks "out of the labor force" than
CPS.
Quite likely these differences come about because SIPP inciudcs a larger array of questions dealing w ~ t h employment and job search actlcitlcs than docs CPS. As a result, SIPP may produce estimate of the size of the labor force which are slightly greater than those derived
TABLE 7 .
L a b o r F o r c e S t a t u s by Race,
Age a n d Sex;
f r o m March,
1 9 8 5 CPS a n d Contemporaneous SIPP
PC = P o p c ~ l a (t o n C o u l ~ t AW = A t Work UN = U n e m p l o y e d O L F = O u t o f Labor F o r c e
-. . -
.
- - - - -..-..-- . . .
lotnl Poptl 1 n t i r1r1 15-24
CPS SIPP
25-34
CPS SlPP
.
-- --- -75 +
CPS
-
--
-
35-4.1
CPS SIPP
45-54
CPS SIPP
55-64
CPS SIPP
65-74
CPS SIPP
CPS
S II'P
SIPP
-
TOTAL POPULATION
-
I
-
WHITE MEN
9.500
PC
AW
74.729 69.8 5.1 251
75.275 71.8 4.5 23.7
15,880 15.936 60.5 8.9 33.0
17,012 17.204
13. 100 13,470
9.456
9.254
9.249
6,542
6.513
3.404
3.404
cn
c . '
IJFI OLF
62.3 8.6 29.1
-
WHITE WOMEN
-
PC
AW 9.007
BLACK MEN
'76.2 5.1 ln.5 -19.1 5 4 45 5 57.4 5.3 37.3
!1.170
58.9
-
2.548 36.4 16 8 46.8
2.502 42.7 19.7 37.7 72.3 ID. 1 14.6 75.1 16.2 8.4
757 11.4 17 9
13.l.tI
UN OLF
11.4 33 0
12.2 28.8
70.1 8.0 13.1
75 7 8 2 161
- BLACK WOMEN
-
FIGURE B.
I
E s t i m a t e d Years Spent i n Each of T h r e e Labor F o r c e S t a t u s e s a s I n d i v i d u a l s Age from 15 t o 6 4 , March, 1985 C u r r e n t P o p u l a t i o n Survey and C o r r e s p o n d i n g Data from SIPP
TOTAL POPULATION
I I
I I I 1 I
I 1
CPS
SiPP
WHITE MEN
CPS SlPP
WHITE WOMEN
CPS
SlPP
BLACK MEN
CPS SIPP
BLACK WOMEN
1
CPS
I
SlPP
EMPLOYED
UNEMPLOvED
K y OUT OF
[
ABOR FORCC
from the Current Population Survey. In recent years, CPS has reported that a rather large proportion of adult black men are not participants in the labor force - 12.1 percent among black men 25 to 54 in 1987 (U.S. Bureau of Labor Statistics, 1988, Table 3). This may be an overestimate in light of these data from SIPP.
SECTION FOUR: The Determinants of Unemployment and Earnings: A Further Comparison of
SIPP and CPS
As a final step in this comparison of SIPP and CPS, we examined the determinants of
unemployment and earnings. The analysis of joblessness was restricted to black and white men and women age 25 to 64. Table 8 shows the estimated unemployment rates, that is, the percent of labor force unemployed in March, 1985, and parameters from models which have the log of the odds of unemployment as their dependent variable. The models treat unemployment as a function of a dichotomous region variable which distinguishes the South, a three category educational attainment variable, a three category age variable and a dichotomous marital status variable which distinguishes those who were mamed and lived with a spouse. [Table 81 Unemployment rates from the two sources are quite similar with the exception of that for black women where the SIPP rate was about 75 percent larger than the one from CPS, a difference which was significant at the .0l level. Looking at the fit of the log-linear models, we find that interaction terms are needed with the exception of the model based upon SIPP data for black women. The independent variables, however, were related to unemployment in the same manner for data sets. That is, unemployment rates were lower in the South, net of other variables and, for each group, they decreased with rising educational attainment. Both SIPP and CPS reported that unemployment also declined with age and that married-spouse-present men and women had lower unemployment rates than people in other marital statuses. It is not easy to compare tllc numerous cocfficicnts shown in Table 8, so Figure C presents estimated unemployment rates for ti~rce sclccted groups. The upper panel compares unemployment rates tor an extensively educated, oldcr, married-spouse-present population
I
lr
It
I I I I I I I I I I I I I I
19
x :
rFF
co. t{ .n (U a om c0
g)
E tn o, C)
sr3
q €
s C' (9 (t
C\l1ttr:
I
(ofo nn co€c OO
?otf COc{ tlqtc) nO;
(9Nq c)@(') c{f-|n flr--
c.,tc.{ 6) o of', c? c{ ("e) of 61t r\
. rFl
nl O n
GOc.|n) |-E = (.).- - a +r 'O)O n P ! o. . : a f t o $tuoNa !s.r e C.; ttl ]f, - r/ ILO L
L+r'-Qlr
o>+
o
.C
E
L
=
o .c
+t
o
I
rt (D
(G n
o. o,
F'
F
N
en
bg
rt
tn Po
rF
n
o o 0E tr| a c o
at (\l € (l)
A{lrttr
I
€0€o cocO ctf" .'-tp
OnO g$lo) c'totci 6;6
qg€ r{t.O 6;= iOnq
fl$f retr n|I? .F!-
cp 91 r1, rn e.,t O g,
.\r
oo
F'
oc
z
o
(L
=-
,C, O
a, L
o c o
tn o, o
be (') tr
I
o J
lU O
co t\
t\ o)
I
.o.c o o 3-e!.c EJ O oc()cc so(E0(1,
(r O
,F
+JL.c(o {E}ncL tu o)c{a O\f I J of o OOOEL LO+r=O Otr o ra- ta- u:l a oc?tl.r r! L OOCS+'
oE coo .g(E0(9L)
1S ! !{$ f C S
OO
lt
o
-VF c{r-b
gqlg mctd
crqlo iio;
iC{OOO
OO ::
C{$l ?!F
o rc{ O
v
+rtl . + ( c L -Fc o c'lc , :tJ (Jcoo
O L
+raE
-n|rrrrl
() ;
iF
(u o. E o
tn ru F oc F 74bq uJm
E IL
E o
noEo.cl or t! L
??.\.
O Ulr.
- - t i L L
o ut F
( F
o, o.
ha
-
(9(o
fitc'
?f.
oo
I
o o)
J
tn
f
o
L
gl
()c oo -!f,
o. ()
6
-, ul = =
ClF ul FO
f.l
Cl J Or
rF'
F.F
c{q!€ toOro
$(9C{ nOto Ittl
FO 3n lu
I
OFO ltOtt
QQO mlnqg
tt.A *si
oto ot$t
ln g) F r- 6i b .
c) s
nr
I
O I O +)Og_ LOot', o.cLJ+t r.-+, o c +tg.o +r oc o E CCOr > (!-EltO E E )O>E O? O O lL E ? tt Q. L O Q. E$C.r, E >O E J C
rr
be
I to o : Ntrrrtg t' OO ?.F rp.F ii snttg c{qt; Jdtr| rrOo ql'ro olcf * F6to O;$r OO t'lt*t nn \rtc 19 61 ,F F an q; O v
= |+)
o
oo c o c f o||. u JT' L O L ooo rt.f.c3u,o, o
-
c o
E
F t^ gl I
o
F
Oo+r lJ E ooc" (ll-r O.C O E+rr 3 f1l v, o o_ OOEE.CTT C+,O oo+ro CO O $- O +. O .O+.r $. O O O CDL OE+,-O +
cL E o c
$
o. o.
|lra
x 1
9
O f\ o
6(,
sq Cl)o clc,
FlF
(ltrO olt\
rn o n
|!
+tc .r (!
o
C{t I
tgt
fl
Ctl ?
(\
€oo|(\f iirllF 60€ :g: ,r"('
€6 $q{ c{N srrr
(? g l: i_
ol fl
(F (} v
+t
c c
E
L
=
c-
Q O f .g E LOtl+, ool o o {., >IL tf- .- O
c'lo,o o c of+r(Jt0tl,
fpfr
bq
.F
o. ()
tn
1
u)
€ o A t9
a a t a a a a a a a
(9(9 nn FF ,FrF
ggo ctcti :q!o lo6rr
ot6o *16 :e1 ( ) 6 59
o o o
+.
tl-
oo FF crclfrt
(|r o fl F c{ O r9
-O OO u o 0t6 c JCNLI: .F J 6lr, .g o .o
.FO+r
_ cuo
E
o|rtrrrtr
I
ult o >c{ o
|
o o o
+t
yrn cIn
t
IJ Icl I rF
c
G +r(!
a.o
C6rU
gt_1 ELLi-
oL
c o o
aOOEo CL=l +, p EEL E ao OOEL O ECooo
.c> O"{tt 3 O r(- .E
+,
o:x
E
*,
EAIF
rF
.cc
t*)
tI'
+t cL
a OOll
o
iF
irl
c{
o
tt e
. I rn !' 6 tu L0 +' a-, |l'+r L =O
E
E
ilo x s ) .-
0 3
trl
n+'
y or
C I | ( ) r r -O O iJ(l, n O L e O O|tlltl _ a .iq't!
";
I
r-.
v g 9f U C yp L O 3 L q QO + ' :O Q Z cou616.oL
d,
aFcc*'{it. q a a o E q .gf-.c G O P rgF t F I ;
l.|.l
o o .g t p o o LOf oc o o FLUDO o o co
n
.+Jl:
O ; .o
F
€
ot['$
a
o
O
O
+)
ur;O s-t a OotO J;=
E
firrtto | | | lnr|lQ g\lmd
o .o O .E L lt cD O
CO63.C F-€
||l O f.r
. }ut I rtt
(, g 3 O o
t5l
living outside the South. Data in the bottom panel rctcr to a younger, less cxtcnsively educated unmarried population living in the South. [Figure C] The differences between the SIPP and CPS estimates of uncmploymcnt are not consistent. For the group with little unemployment, the CPS rates are larger
-
in one case double
- the
estimates rates from SIPP. At the other extreme, whcn we look at the group with high unemployment, we find that the rates estimated from SIPP are larger than those estimated from
CPS.
Returning to the effccts parameters in Table8, we can see why this occurs:
the
relationships of educational attainment to uncmploymcnt and age to unemployment are much stronger in SIPP than in CTS. That is, the SIPP data suggest that increases in schools and in age reduce unemployment by considerably greater amounts than do the CPS data. Data about earnings are presented in Table 9. SIPP asked about earnings in the previous month, and responses to that question are used in this table. CPS included a question about earnings in the previous year - 1984 and weeks worked in that year. We determined average weekly earnings, and then multiplied by 4.35 to obtain an estimate of monthly earnings. People who reported no earnings were deleted from this analysis. The estimates of monthly earnings from the two sources are remarkably close. The biggest discrepancy is among black women. Their average monthly earnings as estimated from CPS were about I1 percent greater than those estimated from SIPP. For the other groups, the CPS estimates were one to four percent larger those based on SIPP data. [Table 91
-
Five independent variables were used to predict monthly earnings. Each person was
given a score equal to the number of years of elcmcntary and secondary schooling and another equalling their reported years of collcge education. Years of potential labor force expenence were estimated by taking a respondent's current age subtracting six and then subtracting the total was also years of schooling. The square of the years ot potential labor iorce cxpcncnce var~able included since wage rates generally do not increase lincarly with years of cxpenence throughout
FIGURE C.
Estimates of Unemployment Rate for Selected March, 1985 CPS and Contemporaneous S I ? ?
Groups;
Cata
from
PANEL A.
MARRIED-SPOUSE-?RESENT PERSON, AGE 50 7 0 64 WITH ;3 OR MORE YEARS OF EDUCATION LIVING OUTSIDE THE SOUTH
WHITE MAN
BLACK MAN
4.07.
WHITE WOMAN
BLACK WOMAN
I
I
I
i
0
1 0
20
30
40
PERCENT UNEMPLOYED
PANEL 0.
WHITE MAN
1
4.07:
MARRIED-SPOUSE-PRESENT PERSON, AGE 35 TO 49 WlTH :2 YEARS OF EDUCATION LIVING OUTSIDE THE SOUTH
6.07..
9.87.
B U C K MAN
3.67.
8.47.
WHrrE WOMAN
BLACK WOMAN
0
3.57.
7.57.
I
I
1
I
1 0
- PERCENT -
20
30
40
UNEMPLOYED
PANEL C.
NEVER-MARRIED PERSON, AGE 25 TO 3 4 LESS THAN 12 YEARS OF EDUCATION LIVING IN THE SOUTH
WHITE MAN
BLACK MAN
32.87:
I
WHITE WOMAN
6.9X
15.1%
1 12.47.
BLACK WOMAN
I
30.27.
I
I
1
0
1 0
20
30
40
PERCENT UNEMPLOYED
ESTIMATES FROM SIPP
1-1
ESTIMATES FROM C P S
22
=
o
a \
I
ot<
o
:
a
Ot-'O . .tt+ -(0 t'+
(0OO :-c''l
c?O
Fo -Q c
L
()
Qr?o O|oc.t
t\lttF
o . r f D ( l {
trF
rFrD
u:! rF
f{nc{ 6r-.,o
6
ol cl
o o)
+. tgo
Otr
c{F? e'.3+
o
q gg nBr'
n|F(!to)o€o{' t4O : a
*t.t++
o
tgtotiO
oeto;
I
LE (p* oo E fo cc
FE
o
=
q- O, O r.r V'
..-
o.
c oo
ao oo oc
f,
c')L oo ao
ol
[!
LIE
(]
F
F=c !.c': oqc l r JE L ^ O q1 >lE rg : .ctt I OO t'__ OZZ =JJ
P:q a€t; :>Q'CrrroO
o
gl O.-
Q= c9 :o c LC
op
n C. GC -
+.
c ur (l, {u150, 0 'F' (, t.l O L F a a Io. rgco o(u
U Orrtr- 5 O
O
a0
o
L+t t\lg
+t la (J(' o
('3 Jsr
t)ootooA >:o0E6r/f >LEl_.rcJc L euJ 0O -g rr.-r+.JDE 16lt I E o o L O L 6 f, q.- 6 C O CD O OnOr_tn Et)On r-ir_ F[trLx Ol(/ltOutOUtOC F O O O-8.: ur cJ> t Rx
; g g6 0-
Ig
g L O
I C r-t
q
t0# | t! Etr Our F UJ
'o c t r . ? l@ + L !q { E g O 'f , + , tt, q oo q 3 o(J _O > 1l J C L.r t! o
E 9 t, o
c
o (r L o
u_J Ld
oU 0O aa oo
o o L ou.c 6 a n
.
go +,
t{lr
} F
(!r
o
O
qe E3
FO ()
; O L n o o $ {, O o cr n o.; o '0 L X {Ul! O L X t'u, O-
r*-.r 0 L
rt-.r
co (E6 ; e o
f 6 vr
C
,C lfl O.F
oo
.o
{E
o+r
NG aa
o
A}
uto
L O L t 15 .1 q,,1
E €rr rEoO tn 3.o 6 q-#
oo F LI NT' o ou
L
A 3 (J V,
I I I I I I I I I I I I T I I I I I I
the entire life span. Because mamed-spouse-present men earn more than other men while mamed-spousepresent women earn less than other women, a dichotomous variable indicating marital status was included. Finally, since wage rates In the South remain below those in other regions, a dichotomous region variable identified this section of the nation. (For other examples of the fitting of earnings equations w t h SIPP data, sce: U.S. Bureau of the Census, 1987). Table 9 presents the means of the variables used in this analysis and shows coefficients from the regression models. Figure D reports estimated monthly earnings from CPS and SIPP for three selected groups. [Figure Dl An examination of the three panels of Figure D shows that, in every instance, monthly earnings estimates from SIPP were nearly equal to those estimated from CPS. A look at the regression coefficients in Table 9 suggests the reasons for this. The earnings returns associated with investments in education were very much the same in the two sources. The regression coefficients also indicate that the increments in earnings associated with years of labor market experience were similar.
CONCLUSION
This investigation, we believe, suggests that sample sizes in SIPP are sufficiently large to support extensive analysis of black-white differences and allows, to a more limited degree, comparisons within the diverse black population. Furthermore, we are confident that estimates of basic demographic, social and economic characteristicsin SIPP are similar to those obtained in
CPS. That is, estimates of educational attainment and marital status for blacks obtained from the
two sources were very similar although SIPP procedures lead to estimates of labor force participation which exceed those of CPS. Both data sources provide similar information about the determinants of unemployment rates and earnings. Indeed, the earnings models b a d on data from the two sourccs are remarkably s~mllar ~mplylng, once agaln, the value oi SIPP for the analysis of racial differences.
FIGURE D.
E~timatesof Monthly Earnings isr Selected Groups; Data from March, 1985 CPS and Contem~oraneous 3Z2P
PANEL A.
WHITE MAN BLACK MAN
MARRIED-SPOUSE-PRESENT PERSON, AGE 45 WITH 16 YEARS OF EDUCATION LIVING OUTSIDE THE SOUTH
j
$2679 $2664
1 $1028 WHITE WOMAN BLACK WOM4N
0 500 600 900 1200 1500
$1084 $17 1 9 1597
I
1
I
I
1
1800
2100
2400
2700
3000
MONTHLY EARNINGS
PANEL B.
WHITE MAN
MARRIED-SPOUSE-PRESENT PERSON, AGE 35 WlTH 14 YEARS OF EDUCATION LIVING OUTSIDE THE SOUTH
$1603
BLACK MAN WHITE WOMAN BLACK WOMAN
*
$155 1 $8 1 6 $85 1
1
I
I
1
I
1
I
I
I
1
0
300
600
900
1200
1500
1800
2100
2400
2700
3000
MONTHLY EARNINGS
PANEL C.
WHITE MAN BLACK MAN
NEVER-MARRIED PERSON, AGE 30 WlTH 11 YEARS OF EDUCATION LIVING IN THE SOUTH
$120 1 $1143 $1026 $870
WHITE WOMAN BLACK WOMAN
1 I
$927
$67 1 $611
I I
1
I
8
I
I
I
0
300
600
900
1200
1500
1800
2100
2400
2700
3000
MONTHLY EARNINGS
ESTIMATES FROM SIPp
1-1
ESTIMATES FROM CPS
BIBLIOGRAPHY David, Martin 1985 "Introduction: The Design and Devclopmcnt ot SIPP," lournal ot Economic and Social Measurement, Voi. 13, Nos. 3 and 4, (Dcccmbcr), Pp. 215-221. Ryscavage, Paul
1984 "SIPP and CPS Labor Force Concepts: A Comparison," Proceeding of the Social Statistics
Section, American Statistical Association, 1984. Pp. 523-528.
US. Bureau of the Census
1987 'Male-Female Differences in Work Expencncc, Occupation, and Earnrngs: 1984," Current Population Reports, Serics P-70, No. 10 (August)
U. S. Bureau of Labor Statistics
1988 Employnent and Earnings. Vol. 35, No. 1 (January).