Full text of DOI:10.1111/j.1741-3737.2000.00808.x

JUDITH R. SMITH Fordham University
JEANNE
B
ROOKS-GUNN, PAMELA K. KLEBANOV, AND KYUNGHEE LEE
Teachers College, Columbia University
Welfare and Work: Complementary Strategies for
Low-Income Women?
We examine the effects of mothers’ strategies of
combining employment and welfare receipt during
the first 3 years of their child’s life on the child’s
cognitive development, behavior problems, and
home learning environment at ages 5 to 6. We
compare the child outcomes of those mothers who
were continuously employed and received no wel-
fare with (a) those who worked some or all of the
Act (PRWORA) of 1996 and the Family Support
Act of 1988, aim to move welfare-dependent
mothers with young children into the paid labor
force. These policies express the societal expec-
tation that welfare-dependent women with young
children must engage in paid or unpaid employ-
ment outside the home in order to be eligible for
income assistance. Yet emerging research dem-
onstrates that, even prior to the new welfare re-
form legislation, many welfare recipients were us-
ing paid employment as part of their strategy to
support themselves and their children, despite
very low wages and added child-care costs (Edin
3
years and also received public assistance and
(b) those who were totally dependent on public
assistance. We studied children in single-parent
families (N ϭ 1271) living below 200% of the
poverty threshold using data from the National
Longitudinal Survey of Youth—Child Supplement.
No negative association was found on most child
outcomes with a mother’s employment whether or
not it was combined with public assistance. How-
ever, mothers’ not working at all and receiving
financial support solely from AFDC was associ-
ated with negative child outcomes. We discuss the
implications of these findings for the possible ef-
fects of the new welfare laws on families and
young children.
&
Lein, 1997; Edin & Jencks, 1992; Harris, 1992,
1
996; Pavetti, 1992, 1995; Spalter-Roth, Burr,
Hartmann, & Shaw, 1995). Marianne (1998) re-
fers to the various strategies for survival used by
welfare-dependent women as ‘‘income packag-
ing.’’ What has not yet been studied is the effect
on child well-being of mothers’ combining wel-
fare receipt and employment. Will this strategy
benefit or harm young children? It is possible that
additional income from maternal employment will
have positive effects on the child’s well-being, but
the mother’s working might also be associated
with indirect negative effects on the parent-child
interaction, which in turn might be associated with
negative child outcomes.
Recent welfare reform legislation, including the
Personal Responsibility and Work Opportunity
Graduate School of Social Services, Fordham University,
1
13 W. 60th Street, New York, NY 10023 (jsmith@
fordham.edu)
Research on welfare-to-work programs, which
target poor mothers with young children, has fo-
cused on the woman as provider, rather than the
Key Words: poverty, welfare, welfare-to-work programs,
work.
8
08
Journal of Marriage and the Family 62 (August 2000): 808–821

Complementary Strategies: Employment and Welfare
809
woman as parent. Most evaluations of welfare re-
form programs have measured success by the
changes in parents’ economic status (Long, Wood,
Haskins, 1985; Schwartz, 1983). Recent studies
with larger samples have not substantiated these
claims (NICHD Child Care Study Group, 1998).
Other studies have focused on child cognitive out-
comes related to early maternal employment. Us-
ing secondary analysis on large samples, negative
associations have been found between a child’s
verbal ability score at ages 4 and 5 and a mother’s
employment hours during the child’s first year of
life, but not between maternal employment and
cognitive scores after this first year (Baydar &
Brooks-Gunn, 1991; Blau & Grossberg, 1990;
Parcel & Menaghan, 1994; Smith, Brooks-Gunn,
& Jackson, 1997; but see also Harvey, 1999)
&
Kopp, 1994; Maynard, 1995; Wood, Bloom,
Fellarth, & Long, 1995). Less attention has been
paid to the possible influence on children of the
employment of women in welfare-to-work pro-
grams (Collins & Aber, 1996; Haskins, 1985; Wil-
son, Ellwood, & Brooks-Gunn, 1995; Zaslow,
Moore, Morrison, & Coiro, 1995). The association
of welfare-to-work programs with the parenting of
the mother, as well as the organization of the
home and the mother’s mental health, have been
missing in most reports (see Aber, Brooks-Gunn,
&
Maynard, 1995; Zaslow & Emig, 1997, as ex-
ceptions). Welfare-to-work evaluations focus on
the number of families who stop receiving public
assistance or the number of hours these former
recipients engage in paid employment. An untest-
ed implicit assumption made by some policy an-
alysts and by the public at large is that limiting
public assistance income benefits and moving
welfare-dependent women to employment will
have positive effects on the women themselves
and on their children.
SOCIAL STRESS RESEARCH
Research has investigated how social stressors re-
lated to employment affect child outcomes
through family interaction patterns (Menaghan,
Jowaleski-Jones, & Mott, 1997). Findings dem-
onstrate that jobs that encourage autonomy and
self-direction influence a mother’s intellectual
flexibility and have positive associations with
mother–child interactions at home (Menaghan &
Parcel, 1991; Miller, Schooler, Kohn, & Miller,
1979; Parcel & Menaghan, 1994). Low-income
working mothers are the least likely group to hold
jobs that offer opportunities for autonomy or self-
direction. Research on parental employment done
with fathers has found that there is a negative as-
sociation between a father’s loss of employment
and child outcomes. The negative child outcomes
are associated with the fathers’ increased hostile
behavior toward their male children after a job
loss (Conger, Ge, Elder, Lorenz, & Simons, 1994;
Elder, 1974). McLoyd (1990) also found that the
path from income insecurity and child outcomes
was through the parent–child interaction.
There are four approaches to studying the as-
socation between maternal employment and child
well-being: investigating (a) the impact of moth-
ers’ absence from the home because of employ-
ment; (b) the effect of stressful family interaction
patterns; (c) the changes in family income secu-
rity; and (d) the effects of multiple roles on ma-
ternal mental health.
MOTHER AS HOME BASE
The impact of maternal employment on child
well-being was initially studied by developmental
psychologists in terms of the possible effects on
the child’s well-being of the mother’s employ-
ment-related absences. Attachment theory has
been used by many to examine whether the
mother’s employment outside the home is asso-
ciated with deleterious child outcomes that can be
ascribed to the mother’s absence during the criti-
cal periods of early child development (Ains-
worth, 1964; Barglow, Vaughn, & Molitor, 1987;
Belsky & Eggebeen, 1991; Clarke-Stewart, 1989).
Some early studies, using small samples of pri-
marily middle-class children, have shown that tod-
dlers who experienced out-of-home care were less
likely to be securely attached and, in a few stud-
ies, were more aggressive with peers and less
compliant with adult demands (Belsky, 1988;
MATERNAL
E
MPLOYMENT AND
FAMILY INCOME
SECURITY
A few recent studies have found negative effects
of welfare receipt and nonemployment on child
well-being; negative effects have been explained
primarily through low family income (Zill,
Moore, Smith, Sief, & Coiro, 1995). There is also
a growing body of research that documents the
negative effects of poverty on young children
(Brooks-Gunn & Duncan, 1997; Duncan &
Brooks-Gunn, 1997; Duncan, Yeung, Brooks-
Gunn, & Smith, 1998). Several researchers have
investigated the labor-force trajectories of welfare-

8
10
Journal of Marriage and the Family
dependent women who seek employment (May-
nard, 1995; Pavetti, 1995; Zill et al., 1995). Many
women find jobs and stop receiving welfare, yet
do not develop a secure attachment to the labor
force (being employed over the course of an entire
year) and recycle back onto public assistance for
a period of time (Ellwood & Bane, 1994; Harris,
tween child well-being and six maternal self-suf-
ficiency strategies involving welfare or employ-
ment in a sample of low-income families with
children from birth to 3 years. We ask the ques-
tion: How is a mother’s employment pattern,
whether combined with public assistance or not,
associated with a child’s cognitive and achieve-
ment abilities, with the child’s emotional adjust-
ment, and with the cognitive and emotional en-
vironment the parent provides at home? Based on
current research on women who are not poor, we
hypothesize that a mother’s employment per se
will not be detrimental to young children, whether
or not it is coupled with welfare receipt. We ex-
pect that the effects of combining welfare and em-
ployment will be similar to the effects of maternal
employment in general—that is, we do not expect
to see direct negative effects of this complemen-
tary strategy on child well-being. We also hypoth-
esize that increases in family income will have
positive effects on child well-being and will offset
some of the negative effects of maternal employ-
ment on child well-being, if some negative effects
are found to be associated with employment. We
also expect that welfare receipt and lack of em-
ployment during the child’s early years will have
negative associations with child outcomes, with
this effect being due in part to the loss of income
experienced by the family. Finally, we do not ex-
pect that a mother’s limited intellectual abilities,
as measured by the Armed Forces Qualifying Test
(AFQT), will fully explain the negative effects on
child well-being of the mother’s lack of employ-
ment and her receipt of welfare benefits.
1
992, 1996; Pavetti, 1995). Most welfare-depen-
dent women who combine or substitute employ-
ment for public assistance enter the low-wage
market, which does not provide an annual income
sufficient to lift a mother and her child out of pov-
erty. In addition, a working mother incurs addi-
tional expenses, including child care, transporta-
tion, and wardrobe costs. Finally, most low-wage
jobs do not offer health insurance.
MULTIPLE
R
OLES AND
MATERNAL MENTAL
HEALTH
Conflicting views exist regarding the effects of
combining work with family roles. Some re-
searchers have found that role strain emerges
when women combine family responsibilities and
outside employment; others have argued that mul-
tiple roles enhance a woman’s well-being by of-
fering her new opportunities for increased status,
privileges, and self-esteem (Barnett & Marshall,
1
992; Baruch & Barnett, 1987). Marshall and Bar-
nett (1991) found that the strain of combining
work and family was greater among working-class
women because of their limited resources for eas-
ing the workload of the double shift. A mother’s
level of satisfaction with her role has been the
most extensively studied mediator between mater-
nal employment and child development. Many
studies have confirmed that a mother’s satisfaction
with her role, whether she is employed or not, has
positive indirect effects on her children through
the mother–child interaction (Baruch & Barnett,
METHOD
Design
We conduct analyses using the National Longitu-
dinal Survey of Youth—Child Supplement
(NLSY—CS), which includes information on the
mother’s employment pattern or welfare pattern
during the child’s first 3 years of life, as well as
assessment data on the child’s cognitive abilities,
behavior problems, and Home Learning environ-
ment at age 5. We examine five maternal welfare
or welfare–employment patterns during the first 3
years of their child’s life and we compare these
via regressions to the patterns for mothers who
were employed for each of the first 3 years of their
child’s life and did not receive public assistance.
We realize that if differences are found, these are
subject to selection bias—that is, unmeasured
1
987; Hock, 1980; Hoffman, 1989; Spitze, 1988).
Poor women with young children may use sev-
eral different patterns for combining employment
and welfare receipt. Some people work and si-
multaneously receive public assistance (this can
be a legal option if the income is reported to the
welfare department and adjustments are made to
the benefit); some people cycle between the two;
and some remain primarily reliant on public as-
sistance without any paid employment. Human
capital differences have been found among the
women in each of these groups (Spalter-Roth et
al., 1995).
In this paper, we examine the association be-

Complementary Strategies: Employment and Welfare
811
T
ABLE 1. SAMPLE HARACTERISTICS: MOTHERS AND
C
characteristics that differentiate the mothers in one
welfare transition group from those in another.
T
HEIR OUNG CHILDREN: NLSY–CS (N ϭ 1271)
Y
Therefore, our models include a number of pre-
existing child, maternal, and family characteristics
to decrease the probability that unmeasured char-
acteristics are accounting for any group differ-
ences.
Variable
Range
Mean
SD
Male
Black
Low birth weight
Mother’s education
Teen mother at birth
AFQT
Income-to-needs (3 years)
Hours per week (year 1)
Hours per week (year 2)
Hours per week (year 3)
Home learning score
Home warmth score
PPVT–R
0–1
0–1
0–1
1–20
0–1
1–99
0–1.9
1–45
1–57
1–65
29–129
37–132
40–136
0.49
0.60
0.13
11.4
0.18
22.6
0.89
14
14
18
92.9
92.0
82.9
—
—
—
1.8
—
20.4
0.48
9
11
13
16.8
16.6
17.2
12.7
13.4
6.5
a
Data and Sample
The National Longitudinal Survey of Youth—
Child Supplement (NLSY—CS) is a panel study
of 12,000 young men and women begun in 1979.
The Child Supplement of the NLSY includes de-
tailed longitudinal demographic information on
the families, as well as cognitive and socioemo-
tional assessments of the young children admin-
istered biannually. The original NLSY sample,
drawn in 1979, oversampled poor and minority
youths, providing samples of disadvantaged fam-
ilies large enough for within-group comparative
analyses (Baker & Mott, 1989; Chase-Lansdale,
Mott, Brooks-Gunn, & Phillips, 1991). Data on
income, employment, and welfare receipt are
available for each mother for every year of the
child’s life. The sample consists of Black and
White children who were living in families with
incomes below 200% of the poverty threshold
when averaged over the first 3 years of the child’s
life. In addition, all the children in the sample
were part of a single-parent family for some of
their first 3 years of life, and all were 5 or 6 years
of age in 1986, 1988, 1990, or 1992. Hispanic
children were excluded from analyses, because
key child outcome variables were not adminis-
tered in Spanish. The result was missing data in
some cases and English assessments for bilingual,
but primarily Spanish-speaking, children in other
cases (Baydar & Brooks-Gunn, 1991). The sample
consists of 1,271 children. Employment and wel-
fare receipt are measured when the child is in the
first, second, and third year of life. Parenting and
child outcomes are examined with the child is 5–
PIAT reading
PIAT math
Behavior problems
65–135 103
65–135
0–32
95.5
11.4
Note: Low birth weight Ͻ2,500 grams (yes or no).
Note: 1986–1992 cohorts used of children ages 5 and 6
years of age.
a
Mother’s education is highest school grade attended
(continuous variable).
high-school graduates, and 17% had attended col-
lege. The average income-to-needs of the families
over the first 3 years of the child’s life was .90.
Sixty-five percent were living below the official
poverty line, 20% were living with incomes be-
tween 100% and 150% of the poverty line, and
15% were living between 150% and 200% of the
poverty line. Table 1 presents means and standard
deviations on the outcome variables and explan-
atory variables.
Measures
Welfare and employment groups. We examined
mothers’ patterns of employment during the first
3 years of their child’s life. A woman’s use of
employment, whether coupled with public assis-
tance income or not, was coded and classified her
as belonging to one of 6 possible welfare–em-
ployment groups. Women reported on their em-
ployment hours each quarter of their child’s life.
They also reported on whether they received in-
come from public assistance (and how much) for
each month of each year. For this study, the
NLSY—CS analyses defined employment as
working for an average of at least one hour per
week during the calendar year. Work hours per
week were computed by the mean of the total
number of hours the mother reported working in
the last 2 quarters of each year (Baydar & Brooks-
Gunn, 1991).
6
years of age. Sixty percent of the children are
Black, 18% had a mother who was a teenager
when they were born, 14% of the mothers
changed their marital status over the child’s first
5
years of life (some married, others separated or
divorced; the sample excludes all families in
which the mothers were consistently married for
the first 3 years of the child’s life). Mothers had
on average 11.5 years of education by the time
their children were 3 years of age. Forty-one per-
cent did not graduate from high school, 42% were

8
12
Journal of Marriage and the Family
Child characteristics. The child’s gender, race,
birthweight, and age in months at assessment are
included in all analyses.
acy (Brooks-Gunn, Guo & Furstenberg, 1993;
Guo, Brooks-Gunn, & Harris, 1996) and, along
with scores of other standardized tests, with sev-
eral family and environmental variables. We use
the standardized score, with a mean of 100 and a
standard deviation of 15 (M ϭ 82.9, SD 17.2).
Children who took the PPVT-R prior to 1990 with
standardized scores below 40 were given a score
of 40 (n ϭ 6; Baker & Mott, 1989). These scores
were kept in these analyses. Analyses were done
without these cases and the findings were parallel.
When the children were 5 to 6 years old, the
Peabody Individual Achievement Test (PIAT) was
given (Dunn & Markwardt, 1970). The PIAT is a
wide-range measure of academic achievement for
children aged 5 and over and is among the most
widely used brief assessments of academic
achievement. It has high test–retest reliability and
concurrent validity. The PIAT Mathematics as-
sessment consists of 84 multiple-choice items of
increasing difficulty, beginning with early skills
(recognizing numerals) and progressing to ad-
vanced concepts in geometry and trigonometry.
The child looks at each problem and then chooses
an answer by pointing to or naming one of four
options. The PIAT Reading Recognition assess-
ment measures word recognition and pronuncia-
tion ability, which are considered essential com-
ponents of reading achievement. The PIAT
Reading Recognition contains 84 items, which in-
crease in difficulty. Skills assessed include match-
ing letters, naming names, and reading single
words aloud (Baker, Keck, Mott, & Quinlan,
1993). We use the standardized score on both as-
sessments, each of which has a mean of 100 and
a standard deviation of 15 (In PIAT Reading M
ϭ 103.0, SD ϭ 12.7); In PIAT Math M ϭ 95.5,
SD ϭ 13.4).
Family and maternal characteristics. The mother’s
education is the highest grade she achieved up to
the year of the child assessment. The variable for
family structure compares families that were con-
sistently single parent (female-headed) with those
in which the mother was married for part of the
child’s first 3 years of life (families in which the
mother was consistently married were omitted).
The mother’s age at birth of the child (teenage
mother or not) is also included. To examine the
effect of income, we used the family’s total in-
come from wages and public assistance for each
year of the child’s life. Our sample is limited to
families whose incomes averaged less than 200%
or below of the ‘‘official’’ U.S. poverty threshold
over the child’s first 3 years of life. The poverty
threshold is based on a set of income thresholds
that were developed in the 1960s and are adjusted
for family size and the changes in the cost of liv-
ing, using the Consumer Price Index. We com-
puted a family income-to-needs ratio by dividing
each household’s income by its corresponding
poverty threshold. Then, income to needs was av-
eraged over the first 3 years of the child’s life
(
Duncan, Brooks-Gunn, & Klebanov, 1994). We
include average family income-to-needs ratios as
a control variable in our analyses. Welfare receipt
was measured by the respondents’ reports of
whether they had received public assistance (wel-
fare) in the past year, when interviewed when
their children were 1, 2, and 3 years of age. (Fam-
ily income was not adjusted to include food
stamps.) We used the mothers’ percentile scores
on the Armed Forces Qualifying Test (AFQT) to
examine possible selection effects based on moth-
ers’ aptitude. Although the AFQT is not an intel-
ligence test, per se, it is highly correlated with IQ
and is a widely used measure of adult aptitude.
Child outcomes: emotional adjustment. The Be-
havior Problems Index (BPI) is a maternal report
of her child’s behavior problems. In the NLSY—
CS, the BPI was administered when the children
were 5 to 6 years of age and consists of 28 items
(1986) or 32 items (1988–1992) drawn from sev-
eral measures (Achenbach & Edelbrock, 1981;
Peterson & Zill, 1986). Questions related to
school attendance were used only for those en-
rolled in a school-based program. The BPI is
scored on a 3-point scale—1 ϭ often true, 2 ϭ
sometimes true, and 3 ϭ not true. Examples of
questions are: ‘‘Child has difficulty getting along
with others,’’ ‘‘Child has sudden changes of
mood,’’ ‘‘Child cheats or lies.’’ The mean total
Child outcomes: cognitive assessments. Cognitive
assessments include a receptive verbal ability test
and achievement tests. The Peabody Picture Vo-
cabulary Test–Revised (PPVT-R) was given to the
children at ages 3 to 4 (or at ages 5 to 6, if for
some reason they did not take the assessment at
the younger age). The PPVT-R provides an esti-
mate of the child’s receptive English vocabulary
and verbal ability (Dunn & Dunn, 1981). It is not
an intelligence test. Scores on the PPVT-R are as-
sociated with later school performance and liter-

Complementary Strategies: Employment and Welfare
813
score for the sample at ages 5 to 6 is 11.4 (SD ϭ
.8).
structure.) The outcomes of interest include (a) the
child’s cognitive test scores, (b) behavior prob-
lems score and the Home Learning environment,
and (c) the emotional warmth in the home.
1
Parenting behavior. The Home Observation of the
Measurement of the Environment (HOME) is a
widely used measure of the provision of learning
experiences and maternal warmth. It is studied via
observation of the home environment, an inter-
view with the mother, and observation of the
mother and child in interaction (Bradley et al.,
Coding is as follows: child’s gender (0 ϭ fe-
male; 1 ϭ male); birth weight (1 ϭ Ͻ2,500 g; 0
ϭ Ͼ2,500 g), age of child in months when child
took the PPVT-R assessment, mother’s education
(continuous variable); age at birth of child (1 ϭ
Ͻ18; 0 ϭ Ͼ18); maternal race or ethnicity (0 ϭ
non-Black, non-Hispanic; 1 ϭ Black); marital sta-
tus (0 ϭ consistently single parent; 1 ϭ married
for part of the child’s first 3 years). Family in-
come is a continuous variable computed as the
average income-to-needs ratio of the family over
the first 3 years of the child’s life. For the Behav-
ior Problem Index, we recoded all questions so
that higher scores described more behavior prob-
lems. We use all 32 questions for those children
who were in a formal school program; we use 28
questions for those not in school. If more than
25% of items were missing, the case was dropped.
The measurement of the HOME cognitive and
emotional scale are both continuous variables.
1
989; Caldwell & Bradley, 1984). The HOME is
highly associated with a variety of child out-
comes, often being a stronger predictor of cogni-
tive and school academic readiness than maternal
education (in regressions entering both variables;
Bradley et al., 1989; Brooks-Gunn, Klebanov, &
Liaw, 1994). We use a short version of HOME
(
Baker & Mott, 1989; 26 items—15 are maternal
report and 11 are interviewer ratings; potential
range is 0 to 26). Two scales are derived from the
standardized scores generated by Baker and Mott
(
1989; Chase-Lansdale, Gordon, Brooks-Gunn, &
Klebanov, 1997)—a cognitive stimulation scale
M ϭ 92.9; SD ϭ 16.8) and an emotional support
(
scale (M ϭ 92.0; SD ϭ 16.6).
RESULTS
Analytic plan. We use multiple linear regressions
(
ordinary least square [OLS]) to examine the as-
Intercorrelations
sociation between a mother’s pattern of combining
employment and welfare during the first 3 years
of her child’s life and various child outcomes and
parenting behavior at 5 to 6 years. Child charac-
teristics (age, gender, birth weight) and family
characteristics (mother’s education, age at birth,
race/ethnicity, and marital status) are all included
in the regressions to control for possible preexist-
ing group differences.
In the first set of analyses, five of the six con-
trast-coded welfare–employment groups are en-
tered into the analyses. Those who were employed
during the first 3 years and never received public
assistance are the omitted contrast group. In these
analyses, Model 1 includes gender, age, birth
weight, mother’s education, race or ethnicity, and
age at birth to reduce the possibility of preexisting
family characteristics accounting for welfare and
employment group differences. Model 2 addition-
ally controls for family income over the first 3
years of the child’s life. Model 3 additionally con-
trols for the mother’s score on the Armed Forces
Qualifying Test and for family structure. (We orig-
inally ran family structure as a separate model, but
this yielded no significant findings. Therefore, we
present the findings for AFQT along with family
Mother’s employment and poverty status was neg-
atively correlated and statistically significant dur-
ing each of the first 3 years of a child’s life (Ϫ.24
in year 1, Ϫ.28 in year 2, and Ϫ.33 in year 3).
Total family income was positively correlated
with employment (.31 for employment all 3 years
and no welfare; .13 for some employment and no
welfare). Child outcomes such as PIAT Reading,
PIAT Mathematics, and PPVT-R were correlated
with each other, as expected (rs between .39 and
.48, p Ͻ .001). None of these variables is used as
an independent variable entered into the same
equation with any other variable.
Description of Welfare and Employment Groups
Twenty-four percent of the mothers were em-
ployed for all 3 of the child’s first 3 years; 39%
had a pattern of some employment over the 3
years; and 37% were never employed. Of those
who were employed for each of the first 3 years
of the child’s life, 65% did not receive any public
assistance, 28% received public assistance for 1
or 2 of the 3 years, and 8% received public assis-
tance for the entire 3 years. Of those who worked

8
14
Journal of Marriage and the Family
for 1 or 2 of the first 3 years, 38% received no
public assistance; 41% received it for some period
of time; and 21% received it for all 3 years. Of
those who had no employment during the first 3
years of their child’s life, 30% received no public
assistance; 36% received assistance during some
of the 3 years; and 35% received public assistance
for all of the 3 years.
Table 2 presents the demographics of the wel-
fare and employment groups. Group differences
were found. Those mothers who were employed
during all of the first 3 years of their child’s life
and received no public assistance had, on average,
mother’s absence of employment coupled with
welfare receipt. Children whose mothers were un-
employed during their first 3 years of life and also
received public assistance for part or all of this
time had PPVT-R scores that were on average 4.4
points lower (or one-quarter of a standard devia-
tion) than those children whose mothers did not
receive public assistance and were employed the
entire 3 years. On the PIAT Reading scores, again,
children whose mothers were unemployed for the
first 3 years of life and received public assistance
had scores that were on average 3.7 points lower
(SD ϭ .13) than the contrast group of children of
employed mothers. On the PIAT Math test, a neg-
ative association was found between lack of em-
ployment and welfare receipt (p ϭ .07).
Model 2 shows that family income accounts for
much of the negative influence of nonemployment
coupled with the receipt of public assistance. That
is, once income is controlled, the negative associ-
ation between nonemployment and receipt of pub-
lic assistance on the child’s cognitive score is high-
ly reduced. On the PPVT-R, the association is no
longer statistically significant, and the beta decreas-
es by one third from Model 1 to Model 2 (Ϫ.11
without income, compared to Ϫ.08 with income
controlled). On the PIAT Reading test, controlling
for income decreases the beta also by approxi-
mately 33% (Ϫ.12 without income, compared to
Ϫ.09 with income), but the association is statisti-
cally significant at the .05 level. On the PIAT Math
score, the association of nonemployment and wel-
fare receipt is not significant once income is con-
trolled.
1
2.3 years of completed schooling; 17% had been
married at some point; 18% were teenage mothers
at the birth of their child; and the average income
to needs was 1.4. In contrast, those mothers who
were unemployed for the first 3 years of their
child’s life and received public assistance during
this time had on average 10.7 years of completed
education; 10% had been married for part of the
3
years; 38% were teenagers at the birth of their
child; and the average income-to-needs ratio dur-
ing the first 3 years of their child’s life was .70.
Effects of Employment and Welfare on Child
Outcomes
Cognitive assessments. Differences were found
for the associations between the child’s cognitive
test score and the various employment and welfare
groups (Table 3). Model 1 presents the findings,
controlling for child’s gender, birth weight, age, as
well as the mother’s education, age at birth of
child, and race or ethnicity. Model 2 also controls
for average family income-to-needs over the first
In Model 3, controlling for the mother’s AFQT
score, there is no change in the associations between
child outcomes and mother’s employment strate-
gy, adjusting for mother’s intellectual ability.
3
years of life. Model 3 also controls for family
structure and the mother’s AFQT score. In the
text, we present the unstandardized coefficients,
which represent predicted differences between the
welfare–employment group and the omitted com-
parison group—those mothers who were contin-
uously employed. Standardized and unstandard-
ized beta coefficients, standard errors, and p
values are included in Tables 3, 4 and 5.
Behavior problems. Behavior problems were as-
sociated with a mother’s lack of employment (Ta-
ble 4). Lack of employment, with or without wel-
fare receipt, was associated with elevated behavior
problems. The effect is statistically significant for
those mothers who were unemployed and not re-
ceiving public assistance (p ϭ .03) and margin-
ally significant for those who were unemployed
and receiving public assistance (p ϭ .08). Con-
trolling for income as well as for mother’s AFQT
score did not change the association between in-
creased behavior problems and lack of employ-
ment coupled with lack of welfare. Yet, once in-
come was controlled for, the association between
lack of employment and elevated behavior prob-
In Model 1, without controlling for income, no
negative associations (p Ͻ .05 significance) were
found between the child’s cognitive test score, as
measured by the PIAT or PPVT-R assessments,
and maternal employment patterns, whether or not
combined with public assistance (as compared to
the omitted category of mothers who were contin-
uously employed). Yet, statistically significant (p
Ͻ .05) negative associations were found for a

Complementary Strategies: Employment and Welfare
815
TABLE 2. MEANS AND
S
TANDARD
DEVIATIONS OF
PATTERNS OF MOTHERS’ EMPLOYMENT AND WELFARE RECEIPT
Work All
Years;
No Welfare Some Welfare
Work All
3 Years;
Some
Employment;
No Welfare
Some
Employment;
Some Welfare
No
No
3
Employment;
No Welfare
Employment;
Some Welfare
Characteristics
Mother’s
education
Black
12.3
(1.7)
.55
12.1
(1.7)
.72***
11.6***
(1.9)
.49***
11.5***
(1.6)
.64***
10.8***
(1.7)
.40***
10.7***
(1.7)
.70***
a
(
.36)
.17
b
Ͻ18 at birth
.08
.17
.17
.24*
.21**
.17
(.38)
.10
Divorced, married
or separated
.22*
.15**
.15***
.09ϳ
c
(.30)
Family income
average (3 years)
Hours mother
work (year 1)
Hours mother
work (year 2)
Hours mother
work (year 3)
N
1.4
(.64)
24
(13.7)
34
(13.7)
33
(12.3)
197
1.0***
(.57)
1.1***
(.60)
7***
(10.9)
10***
(11.5)
13***
(13.6)
189
.92***
(.49)
6***
(9.0)
8***
(10.8)
12***
(13.7)
305
.98***
(.65)
.71***
(.46)
15***
(11.2)
24***
(12.5)
28***
(14.0)
108
139
333
Note: Means, standard deviations, and significance level of contrast groups (Bonferonni posthoc comparisons used with
continuous variables; Pearson R used with dichotomous variables: welfare groups with the contrast group of work 3 years
and no welfare).
a
b
Dichotomous variable; (omitted category is ‘‘non-black’’). Dichotomous variable; (omitted category is ‘‘nonteen at
c
birth’’). Dichotomous variable; (omitted category is ‘‘continuously one parent family’’).
*
p Ͻ .05. **p Ͻ .01. ***p Ͻ .001.
lems was no longer significant for those who were
unemployed as well as receiving welfare (p ϭ
HOME warmth score. In Model 1, no associations
between employment–welfare patterns and home
warmth were found. Controlling for income, mar-
ginally significant positive effects of some em-
ployment and lack of public assistance were found
on the warmth expressed by the mother toward
the child (p ϭ .09). Controlling for the mother’s
AFQT score did not alter the findings (p ϭ .10).
.
12).
Effects of Employment and Welfare on Home
Environment
HOME cognitive score. Table 5, Column 1, shows
a trend for a negative association between a
mother’s pattern of some employment combined
with receipt of public assistance and the cognitive
home learning environment (p Ͻ .10). These chil-
dren had home learning environments that were
DISCUSSION
Many welfare-dependent women use the strategy
of combining employment with welfare receipt.
Several income–support strategies may be used by
low-income women to compensate for low wages
within the job market and public assistance that is
not sufficient to cover monthly expenses. Edin and
Lein’s research (1997) demonstrates that only
11% of a sample of welfare recipients (drawn
from three different states) came within $50.00 of
covering their monthly expenses with their wel-
fare benefits; 79% were unable to meet their ex-
penses without augmenting their income from
welfare benefits with employment or informal fi-
nancial help from family or male partners. On the
other hand, Edin and Lein’s study also shows that
low-income women who received no public assis-
tance and attempted to support their families sole-
2
.6 points lower (one sixth of a standard devia-
tion) than those of the continuously employed
group and no welfare receipt. Model 2 shows that
controlling for family income offset the marginal
negative effect of employment coupled with pub-
lic assistance.
Negative effects were found for lack of em-
ployment and welfare on the home learning en-
vironment (p Ͻ .001). The effect is still found
when entering income into the equation (p Ͻ .05).
However, the effect size is lowered by almost 50%
(
beta changed from Ϫ16 to Ϫ.09) once income is
explained. Controlling for the mother’s AFQT did
not change the effect size of this association, but
the p value dropped from .05 to .06.

8
16
Journal of Marriage and the Family

Complementary Strategies: Employment and Welfare
817
T
ABLE 4. OLS REGRESSION
A
NALYSES. CHILD
MPLOYMENT AND
ECEIPT
B
EHAVIOR
ployed for some or all of their child’s first 3 years
of life, that there were no negative associations of
employment with several child outcomes. We did
find a negative effect on the cognitive stimulation
in the home to be associated with a mother’s pat-
tern of some employment over the first 3 years
coupled with receipt of public assistance (com-
pared to being employed for the first 3 years with-
out any public assistance income), but this effect
was no longer significant once we controlled for
family income. These findings indicate that em-
ployment undertaken during the first 3 years of a
child’s life, within a sample of low-income moth-
ers, has no negative effect on a child’s cognitive
test scores or behavior problems at age 5 to 6.
Third, we found that not working at all and
receiving financial support solely from AFDC was
associated with negative child outcomes—com-
pared to being employed for all or some of the
first 3 years of the child’s life. Children in this
group had, on average, lower scores on the PPVT-
R test, lower PIAT Reading and Math achieve-
ment tests, higher behavior problems, and lower
scores on the cognitive stimulation offered to
them in their homes. To understand how contin-
uous welfare support and no additional income
from maternal employment might be detrimental
to children, we looked at the family income situ-
ations of this group of mothers and children. The
mothers who were solely dependent on public as-
sistance had family incomes that were way below
the official poverty threshold (.70 income-to-
needs ratio), whereas those mothers who were em-
ployed for each of the first 3 years had income-
to-needs ratios that were twice as high (1.4
income-to-needs ratio). An income-to-needs ratio
of .70 is considered to represent deep poverty.
Duncan and Brooks-Gunn (1997) have reported
that living in deep poverty exerts the strongest
negative effect on young children as contrasted
with school-age children and adolescents. Conse-
quently, we examined whether the welfare only
effect would still be found once we entered in-
come into the regressions. Once income was in-
cluded, many associations with child outcomes
were reduced, and some became insignificant.
Significant effects remained, controlling for in-
come, on the PIAT Reading score and the child’s
home learning environment (but the negative ef-
fect was reduced by almost half once income was
controlled). Yet, because there remain statistically
significant and marginally significant effects of
unemployment and welfare receipt on child well-
being once income and maternal ability were con-
PROBLEMS: EFFECTS OF
E
WELFARE
R
Employment and
Welfare Group
Model 1
Model 2
Model 3
a
Work 3 years,
some welfare
1.17
(.85)
1.11
(.87)
[.05]
1.15
(.86)
[.05]
b
c
[
.05]
1.07
(.67)
.07]
Some employment,
some welfare
.98
(.69)
[.07]
.98
(.69)
[.07]
[
Some employment,
no welfare
Ϫ.05
(.75)
Ϫ.08
(.76)
[Ϫ.00]
.07
(.75)
[Ϫ.00]
[Ϫ.00]
No employment,
no welfare
1.74*
(.81)
1.69*
(.83)
[.08]
1.62*
(.83)
[.07]
[.08]
No employment,
some or all wel-
fare
1.74ϳ
(.81)
[.08]
1.10
(.72)
[.07]
1.04
(.72)
[Ϫ.07]
N
891
891
.08
891
.08
2
Adj. R
.08
Note: Model 1 controls for gender, age, low birth weight,
mother’s education, race or ethnicity, and age at birth or
child. Model 2 also controls for total average family income
(income-to-needs) over first 3 years of life. Model 3 also
controls for marital status and mother’s AFQT score.
a
b
Unstandardized coefficients.
standardized coefficients].
p ϳϽ .10. Ͻ .05. **p Ͻ .01. ***p Ͻ .001.
(standard errors).
c
[
*
ly through low-wage jobs (less than $7.00 an
hour) faced worse financial problems than the
welfare-reliant mothers. Many of the employed
mothers did not receive food stamps, and all had
the additional expenses of transportation and child
care.
In this paper we have examined the effects of
mothers’ various strategies for family income sup-
port and welfare receipt on child well-being in a
sample of 1,271 low-income children living at or
below 200% of the official poverty threshold. We
had several findings. First, we found, as expected,
that the mothers who used the complimentary
strategy of working and receiving public assis-
tance (either for part or all of their child’s first 3
years of life) had more income than those mothers
who were solely reliant on public assistance,
which supports our premise that this strategy does
increase income. Yoshikawa (1998) also found
that the strategy of combining employment with
welfare receipt was associated with higher wages
for women when their children were 6 years of
age.
Next, we found, among mothers who were em-

8
18
Journal of Marriage and the Family
TABLE 5. OLS REGRESSION
A
NALYSES. PARENTAL
O
UTCOMES OF THE
HOME
LEARNING
E
NVIRONMENT AND
EMOTIONAL
WARMTH: EFFECTS OF
E
MPLOYMENT/WELFARE
R
ECEIPT PATTERNS
Model 1
Model 2
Model 3
Employment and
Welfare Group
Home
Cognitive
Home
Warmth
Home
Cognitive
Home
Warmth
Home
Cognitive
Home
Warmth
a
Work 3 years,
some welfare
Ϫ2.28
Ϫ.29
(2.1)
[Ϫ.00]
Ϫ0.65
(2.06)
[Ϫ.01]
.95
(2.09)
[.03]
Ϫ0.85
(2.06)
[Ϫ.01]
.47
(2.06)
[.00]
b
(2.05)
Ϫ.04]
c
[
Some employment,
some welfare
Ϫ2.58ϳ
(1.6)
Ϫ1.51
(1.6)
Ϫ0.53
(1.6)
Ϫ.07
(1.6)
Ϫ0.63
(1.6)
Ϫ.28
(1.6)
[Ϫ.07]
[Ϫ.04]
[Ϫ.01]
[Ϫ.00]
[Ϫ.02]
[Ϫ.00]
Some employment,
no welfare
1.02
(1.8)
2.29
(1.8)
[.05]
2.19
(1.8)
[.05]
3.1ϳ
(1.8)
[.07]
2.13
(1.8)
[.04]
3.02ϳ
(1.8)
[.06]
[.02]
No employment,
no welfare
Ϫ2.48
(1.9)
0.46
(2.0)
[.01]
Ϫ.91
(1.9)
[Ϫ.02]
1.66
(2.06)
[.03]
Ϫ.60
(1.9)
[Ϫ.01]
2.16
(2.03)
[.04]
[Ϫ.05]
No employment,
all welfare
Ϫ5.91***
(1.6)
Ϫ2.38
(1.7)
Ϫ3.28**
(1.7)
Ϫ0.47
(1.7)
Ϫ3.10ϳ
(1.7)
Ϫ.13
(1.7)
[Ϫ.16]
[Ϫ.06]
[Ϫ.09]
[Ϫ.01]
[Ϫ.08]
[Ϫ.00]
1
056
1048
1056
.14
1048
1056
.14
1048
.12
2
Adj. R
.13
.11
.12
Note: Model 1 controls for gender, age, low birth weight, mother’s education, race or ethnicity, age at birth or child.
Model 2 also controls for total average family income (income-to-needs) over first 3 years of life. Model 3 also controls
for change in marital status and AFQT.
a
b
c
Unstandardized coefficients. (Standard errors). [Standardized coefficients].
p ϳϽ .10. Ͻ .05. **p Ͻ .01. ***p Ͻ .001.
*
trolled, the question remains as to what accounts
for the effect. Could it be motivation or the
amount of cognitive stimulation in the home? We
did find that when analyzing the effects of un-
employment and welfare receipt on the cognitive
stimulation available in the home, the negative ef-
fects were reduced by half with controls for in-
come and maternal ability. Finally, we cannot rule
out possible selection effects—that is, that these
mothers might differ on unmeasured characteris-
tics. However, because entering income into the
equation significantly lowered the effects of wel-
fare receipt, this does suggest that income (and
associated unmeasured characteristics possibly as-
sociated with income) might account in part for
the effects that are operating.
Finally, we examined whether the negative as-
sociations of welfare receipt and unemployment
on child well-being would be explained by the
mother’s intellectual capacity as measured by the
AFQT score in the NLSY data. The mothers in
this sample, on average, had relatively low AFQT
scores (mean percentile was 23rd percentile).
Nevertheless, controlling for the mother’s ability
score did not explain the negative associations be-
tween reduced child well-being and mother’s pat-
tern of employment or unemployment.
In considering the policy implications of these
findings, it is important to remember that the
women in this study were not under any manda-
tory work requirements. The data for the study
were collected on cohorts of women who were
receiving public assistance prior to the employ-
ment mandates of either the Family Support Act
of 1988 or the PRWORA of 1996. Under the new
law, women have reduced choices regarding com-
bining public assistance and employment. Those
low-income women who cannot find satisfactory
child care or who prefer to remain at home with
their infants or toddlers and who rely solely on
public assistance are likely to have their income
benefits cut and will only be able to use this pat-
tern for a lifetime limit of 5 years (2 years in some
states).
The PRWORA is a major shift from previous
welfare policy in terms of the demands made on
mothers with young children to enter the labor
force. The new welfare regulations might influ-
ence the outcomes of welfare-dependent women
and their children in two ways. First, we might
expect to find fewer mothers in the welfare-only
group—as enrollment in an employment-related
program is now considered prerequisite to receipt
of benefits. Each state is allowed, however, to give

Complementary Strategies: Employment and Welfare
819
exemptions to approximately 20% of their case-
load regarding work requirements. The group of
women who will be given exemptions and who
will comprise the new welfare-only group are ex-
pected to be those women who have mental or
physical health problems or very low literacy
skills or who are victims of domestic violence. We
would therefore expect even more pronounced
differences between this group and the other
groups. Such differences might not be accounted
for completely by variations in income (Duncan
ment (other developmental studies are subject to
this limitation, as well). Finally, because our study
is based in the context of welfare in the mid- to
late 1980s, it does not capture the effects of the
PWROWA. However, it provides a basis for com-
parison as new studies enter the field early in the
2000s.
In summary, mothers who rely exclusively on
public assistance have very low incomes and often
work to replace or supplement their welfare ben-
efits. The pattern of seeking employment by low-
income, welfare-dependent women was practiced
by many mothers prior to recent welfare reforms.
Our analyses found no negative effects on chil-
dren’s test scores at ages 5 to 6 of mothers’ mixing
work and welfare. Once we controlled for income,
no negative effects on the child’s home learning
environment were found. Combining employment
with public assistance is associated with higher
family incomes, as expected. Total reliance on
welfare receipt is associated with lower test scores
and less stimulating home learning environments,
due in large part to income.
&
Brooks-Gunn, 1998). Second, we might expect
the mothers in the welfare and work categories to
be more heterogeneous—some will be receiving
public assistance and ‘‘working off’’ their benefit
(
for no pay); some will be receiving reduced ben-
efits because of noncompliance and working ‘‘off
the books’’; others will forgo public assistance
and seek employment because of mandatory re-
quirements; others will seek employment volun-
tarily. Variations in income will depend on which
strategy a woman selects and on the wage she is
able to earn. Current legislation allows welfare-
dependent women who begin employment (within
an income ceiling) to keep their Medicaid cover-
age and food stamps for the first year as they par-
ticipate in welfare-to-work programs. In addition,
Child Health Insurance Plus (CHIP) has been ex-
tended to low-income (nonwelfare) families.
Larger income differences are expected for the
newly employed group, once they lose food
stamps, housing assistance, and medical coverage
NOTE
We want to thank the National Institute for Child Health
and Human Development Research Network on Child
and Family Well-Being and the MacArthur Network on
the Family and the Economy for their support.
An earlier version of this paper was presented at the
workshop on Welfare and Child Development, jointly
sponsored by the Board on Children, Youth and Fami-
lies of the National Academy of Sciences and the Na-
tional Institute of Child Health and Human Develop-
ment in Washington, D.C.
(
see Edin & Lein, 1997). Finally, issues such as
scheduling of work and child care and transpor-
tation costs might loom large for poor and wel-
fare-dependent women entering the labor force. If
welfare-to-work programs or employers offer little
flexibility, child care might become an even great-
er issue than current research suggests (see Kisker
REFERENCES
Aber, J. L., Brooks-Gunn, J., & Maynard, R. (1995).
Effects of welfare reform on teenage parents and their
children. The Future of Children, 5(2), 53–71.
Achenbach, T., & Edelbrock, C. S. (1981). Behavioral
problems and competencies reported by parents of
normal and disturbed children four through sixteen.
Monographs of the Society for Research in Child De-
velopment, 46 (Serial No. 188).
Ainsworth, M. (1964). Patterns of attachment behavior
shown by the infant in interaction with his mother.
Merrill-Palmer Quarterly of Behavior and Develop-
ment, 10, 51–58.
Baker, P., Keck, C., Mott, F., & Quinlan, S. V. (1993).
NLSY handbook: A guide to the 1986–1990 National
Longitudinal Survey of Youth Child Data (rev. ed).
Columbus, OH: Ohio State University, Center for Hu-
man Resource Research.
Baker, P., & Mott, F. (1989). NLSY handbook 1989: A
guide and resource document for the National Lon-
gitudinal Survey of Youth 1986 Child Data. Colum-
&
Moss, 1997).
These analyses provide important information
on an area that has been studied too little—the
effect of combining welfare receipt and employ-
ment on child well-being. Yet our analyses have
several limitations. First, our sample is not truly
nationally representative, as the NLSY—CS does
not yet include children of older mothers. At the
same time, the sample does have relatively large
samples of low-income mothers, in contrast to
most developmental studies (Chase-Lansdale et
al., 1991). Second, we do not have information on
the monthly sequencing of employment and wel-
fare for those in the group that combined or se-
quenced between public assistance and employ-

8
20
Journal of Marriage and the Family
bus, OH: Ohio State University, Center for Human
Resource Research.
Barglow, P., Vaughn, B., & Molitor, A. (1987). Effects
of maternal absence due to employment on the qual-
ity of infant-mother attachment in a low-risk sample.
Child Development, 58, 945–954.
Klebanov, P. (1997). Neighborhood and family influ-
ences on the intellectual and behavioral comptence of
preschool and early school-age children. In J. Brooks-
Gunn, G. Duncan, & L. Aber (Eds.), Neighborhood
poverty: Context and consequences for children (Vol.
1) (pp. 79–118) New York: Russell Sage.
Barnett, R. C., & Marshall, N. C. (1992). Worker and
mother roles, spillover effects, and psychological dis-
tress. Women and Health, 18, 9–14.
Baruch, G. K., & Barnett, R. C. (1987). Role quality
and psychological well-being. In F. Crosby (Ed.),
Spouse, parent, worker (pp. 63–84). New Haven, CT:
Yale University Press.
Baydar, N., & Brooks-Gunn, J. (1991). Effects of ma-
ternal employment and child-care arrangements in in-
fancy on preschoolers’ cognitive and behavioral out-
comes: Evidence from the children of the NLSY.
Developmental Psychology, 27, 918–931.
Chase-Lansdale, P. L., Mott, F., Brooks-Gunn, J., &
Phillips, D. (1991). Children of the NLSY: A unique
research opportunity. Developmental Psychology, 27,
899–910.
Clarke-Stewart, K. A. (1989). Day care: Maligned or
malignant? American Psychologist, 44, 266–273.
Collins, A. & Aber, L. (1996). State welfare evalua-
tions: Will they increase our understanding of the
impact or welfare reform on children? New York:
Columbia University School of Public Health, Na-
tional Center for Children in Poverty.
Conger, R. D., Ge, X., Elder, G. H., Jr., Lorenz, F. O.,
Baydar, N., Brooks-Gunn, J., & Furstenberg, F., Jr.
&
Simons, R. L. (1994). Economic stress, coercive
(
1993). Early warning signs of functional illiteracy:
family processes, and developmental problems of ad-
olescents. Child Development 65, 541–561.
Duncan, G. J., & Brooks-Gunn, J. (Eds.). (1997). Con-
sequences of growing up poor. New York: Russell
Sage.
Predictors in childhood and adolescence. Child De-
velopment, 64, 815–829.
Belsky, J. (1988). The ‘‘effects’’ of infant day care re-
considered. Early Childhood Research Quarterly, 3,
2
27–235.
Duncan, G. J., Brooks-Gunn, J., & Klebanov, P. K.
Belsky, J., & Eggebeen, D. (1991). Early and extensive
maternal employment–child care and 4–6-year-olds’
socioemotional development: Children of the Nation-
al Longitudinal Study of Youth. Journal of Marriage
and the Family, 53, 1083–1099.
Blank, S., & Blum, B. (1997). A brief history of work
expectations for welfare mothers. The Future of Chil-
dren, 7(1), 28–38.
Blau, F., & Grossberg, A. (1990). Maternal labor supply
and children’s cognitive development. (Working Pa-
per No. 3536). Washington, DC: National Bureau of
Economic Research.
Bradley, R. H., Caldwell, B. M., Rock, S. L., Ramey,
C. T., Barnard, K. E., Gray, C., Hammond, M. A.,
Mitchell, S., Gottfried, A. W., Siegel, L., & Johnson,
D. L. (1989). Home environment and cognitive de-
velopment in the first 3 years of life: A collaborative
study including six sites and three ethnic groups in
North America. Developmental Psychology, 25, 217–
(1994). Economic deprivation and early-childhood
development. Child Development, 65, 296–318.
Duncan, G. J., Yeung, W., Brooks-Gunn, J., & Smith,
J. (1998). How much does childhood poverty affect
the life chances of children? American Sociological
Review, 63, 406–423.
Dunn, L. M., & Dunn, L. M. (1981). Peabody Picture
Vocabulary Test–Revised. Circle Pines, MN: Ameri-
can Guidance Service.
Dunn, L. M., & Markwardt, F. C., Jr. (1970). Peabody
Individual Achievement Test Manual. Circle Pines,
MN: American Guidance Service.
Edin, K. (1995). The myths of dependence and self-
sufficiency: Women, welfare, and low-wage work.
Focus, 17(2), 5–10.
Edin, K., & Jencks, S. (1992). Welfare. In C. Jencks
(
Ed.), Rethinking Social Policy. Cambridge, MA:
Harvard University Press.
Edin, K., & Lein, L. (1997). Making ends meet: How
single mothers survive welfare and low-wage work.
New York: Russell Sage.
Elder, G. H. (1974). Children of the Great Depression:
Social change in life experience. Chicago: University
of Chicago Press.
Ellwood, D., & Bane, M. J. (1994). Understanding wel-
fare dynamics. In M. J. Bane & D. Ellwood (Eds.),
Welfare realities: From rhetoric to reform (pp. 67–
2
35.
Brooks-Gunn, J., & Duncan, G. (1997). The effects of
poverty on children. The Future of Children, 7(2), 55–
7
1.
Brooks-Gunn, J., Guo, G., & Furstenberg, F. F., Jr.
1993). Who drops out of and who continues beyond
(
high school?: A 20-year follow-up of black urban
youth. Journal of Research on Adolescence, 3, 271–
2
94.
1
23). Cambridge, MA: Harvard University Press.
Brooks-Gunn, J., Klebanov, P. K., & Liaw, F. (1994).
The learning, physical and emotional environment of
the home in the context of poverty: The Infant Health
and Development Program Children and Youth Ser-
vices Review, 17, 251–276.
Caldwell, B. M., & Bradley, R. H. (1984). Home Ob-
servation for Measurement of the Environment
Guo, G., Brooks-Gunn, J., & Harris, K. M. (1996). Pa-
rental labor-force attachment and grade retention
among urban Black children. Sociology of Education,
6
9, 217–236.
Harris, K. (1992). Work and welfare among single
mothers in poverty. American Journal of Sociology,
99, 317–352.
(HOME). Little Rock, AR: University of Arkansas at
Little Rock, 1984.
Harris, K. (1996). Life after welfare: Women, work and
repeat dependency. American Sociological Review,
61, 407–426.
Harvey, E. (1999). Short-term and long-term effects of
early parental employment on children of the Nation-
Chase-Lansdale, P. L., & Brooks-Gunn, J. (1995). Es-
cape from poverty: What makes a difference for chil-
dren? New York: Cambridge University Press.
Chase-Lansdale, L., Gordon, R., Brooks-Gunn, J., &

Complementary Strategies: Employment and Welfare
821
al Longitudinal Survey of Youth. Developmental Psy-
chology, 35, 445–459.
Haskins, R. (1985). Public school aggression among
children with varying day-care experience. Child De-
velopment, 56, 689–703.
Hock, E. (1980). Working and nonworking mothers and
their infants: A comparative study of maternal care-
giving characteristics and infant social behavior. Mer-
rill Palmer Quarterly, 26, 79–101.
Hoffman, L. W. (1989). Effects of maternal employment
in the two-parent family. American Psychologist, 44,
family social capital, and early childhood outcomes.
American Journal of Sociology, 99, 972–1009.
Pavetti, L. (1992). Dynamics of welfare and work: Ex-
ploring the process by which young women work
their way off welfare. Paper presented at the Applied
Public Policy and Management Association Annual
Meeting, Boulder, CO.
Pavetti, L. (1995). Who is affected by time limits? In I.
V. Sawhill (Ed.) Welfare reform: An analysis of the
issues (pp. 31–34). Washington, DC: Urban Institute
Press.
2
83–292.
Peterson, J. L., & Zill, N. (1986). Marital disruption,
parent-child relationships and behavioral problems in
children. Journal of Marriage and the Family, 48,
224–258.
Jackson, A. P. (1993). Black single working mothers in
poverty: Preferences for employment, well-being, and
perceptions of preschool-age children. Social Work,
3
8, 26–34.
Schwartz, P. (1983). Length of day-care attendance and
attachment behavior in eighteen-month-old infants.
Child Development, 54, 1073–1078.
Kisker, E., & Moss, C. (1997). Arranging child care.
The Future of Children, 7, 99–109.
Long, D., Wood, R. G., & Kopp, H. (1994). LEAP: The
educational effects of LEAP and enhanced services
in Cleveland, Ohio’s Learning, Earning and Parent-
ing Program for teenage parents on welfare (Report
No. 94–34851). New York: Manpower Demonstra-
tion Research Corporation.
Marianne, K. (1998). Income packaging as a survival
strategy for welfare mothers. Affilia, 13, 211–232.
Marshall, N., & Barnett, R. (1992). Race, class and mul-
tiple role strains and gains among women employed
in the service sector. Women and Health, 17(4), 1–
Smith, J., Brooks-Gunn, J., & Jackson, A. (1997). Pa-
rental employment and children. In R. Hauser & B.
Brown (Eds.), Indicators of child well-being (pp.
279–308). New York: Russell Sage.
Spalter-Roth, R., Burr, B., Hartmann, H., & Shaw, L.
(1995). Welfare that works: The working lives of
AFDC recipients. Washington DC: Institute for Wom-
en’s Policy Research.
Spitze, G. (1988). Women’s employment and family re-
lations: A review. Journal of Marriage and the Fam-
ily, 50, 595–618.
Vaughn, B., Gove, F., & Egeland, B. (1980). The rela-
tionship between out-of-home care and the quality of
infant-mother attachment in an economically disad-
vantaged population. Child Development, 51, 1203–
1
9.
Maynard, R. (1995). Teenage childbearing and welfare
reform: Lessons from a decade of demonstration and
evaluation research. Children and Youth Services Re-
view, 17, 309–332.
McLanahan, S., & Sandefur, G. (1994). Growing up
with a single parent: What hurts, what helps? Cam-
bridge, MA: Harvard University Press.
McLoyd, V. C. (1990). The impact of economic hard-
ship on black families and development. Child De-
velopment, 61, 311–346.
Menaghan, E., Jowaleski-Jones, & Mott, F. (1997). The
intergenerational costs of parental social stressors:
Academic and social difficulties in early adolescence
for children of young mothers. Journal of Health and
Social Behavior, 38, 72–86.
1
214.
Wilson, J. B., Ellwood, D. T., & Brooks-Gunn, J.
1995). Welfare to work through the eyes of children:
(
The impact of parenting on movement from AFDC
to employment. In P. L. Chase-Lansdale & J. Brooks-
Gunn (Eds.), Escape from poverty: What makes a dif-
ference for children? (pp. 241–271). New York:
Cambridge University Press.
Wood, R. G., Bloom, D., Fellarth, V., & Long, D.
(1995). Encouraging school enrollment and atten-
dance among teenage parents on welfare: Early im-
pacts of Ohio’s LEAP program. Children and Youth
Services Review, 17, 277–308.
Yoshikawa, H. (1998). Welfare dynamics, support ser-
vices, mothers’ earnings, and child cognitive devel-
opment: Implications for contemporary welfare re-
form. Child Development, 70(3), 779–801.
Zaslow, M. J., & Emig, C. (1997). When low-income
mothers go to work: Implications for children. The
Future of Children, 7(1), 116–133.
Zaslow, M. J., Moore, K. A., Morrison, D. R., & Coiro,
M. J. (1995). The Family Support Act and children:
Potential pathways of influence. Children and Youth
Services Review, 17, 231–250.
Zill, N., Moore, K., Smith, E. W., Stief, T., & Coiro, M.
J. (1995). The life circumstances and development of
children in welfare families. In P. L. Chase-Lansdale
& J. Brooks-Gunn (Eds.), Escape from Poverty: What
makes a difference for children? (pp. 241–271). New
York: Cambridge University Press.
Menaghan, E., & Parcel, T. (1991). Determining chil-
dren’s home environments: The impact of maternal
characteristics and current occupational and family
conditions. Journal of Marriage and the Family, 53,
4
17–431.
Miller, J., Schooler, C., Kohn, M., & Miller, K. (1979).
Women and work: The psychological effects of oc-
cupational conditions. American Journal of Sociolo-
gy, 85, 66–94.
Moore, K., & Driscoll, A. (1997). Low-wage maternal
employment and outcomes for children: A study. The
Future of Children, 7(1), 122–127.
NICHD Early Child Care Research Network (1998).
Early child care and self-control, compliance, and
problem behavior at twenty-four and thirty-six
months. Child Development, 69, 1145–1170.
Parcel, T., & Menaghan, E. (1994). Early parental work,