The Jobs Evaluation: How Well are They Faring? AFDC Families with Preschool-aged Children in Atlanta at the Outset of the Jobs Evaluation


The JOBS Evaluation:

How Well Are They Faring?

AFDC Families with Preschool-Aged Children in Atlanta at the Outset of the JOBS Evaluation.

Prepared by Kristin A. Moore, Martha J. Zaslow, Mary Jo Coiro, Suzanne M. Miller of Child Trends, Inc.and Ellen B. Magenheim of Swarthmore College.

February 1996

This report was prepared for the U.S. Department of Health and Human Services (HHS), and the U.S. Department of Education under a subcontract from the Manpower Demonstration Research Corporation (MDRC). MDRC is conducting this evaluation under a contract with HHS. HHS also receives support for the evaluation from the U.S. Department of Education. Additional funding for this study has been provided by the Foundation for Child Development, the William T. Grant Foundation, and an anonymous funder. A list of other reports about the JOBS evaluation is available here.





The centerpiece of the 1988 Family Support Act (FSA) is the Job Opportunities and Basic Skills Training (JOBS) Program, which requires eligible recipients of Aid to Families with Dependent Children (AFDC) to participate in educational, job training and work experience, or job search activities, in order to reduce welfare dependency and promote self-sufficiency. Although most services offered through JOBS are aimed at meeting the needs of adults, there are numerous reasons to expect that JOBS may also affect children in families that receive AFDC.

By What Pathways Might JOBS Affect Children?

The legislative debate that led to the passage of the Family Support Act was "two-generational" in focus. That is, lawmakers recognized the implications of poverty and welfare dependency for both parents and children. However, the JOBS Program focuses primarily on the parental generation, although transitional child care and Medicaid benefits mandated under JOBS do recognize the needs of young children. Because the JOBS program is part of the government's effort to interrupt the inter-generational transmission of poverty, it is important to consider the possibility of either positive or negative effects on children. There are several mechanisms by which JOBS could affect children. These include changes in parent education or family income; changes in the home environment; changes in mothers' psychological well-being; and increased participation in child care.(1)

As a national policy, the underlying assumption of the FSA is that the needs of poor children are best addressed through providing parents with education and job training services. An important potential pathway of influence of JOBS on children is via increasing maternal education, employment skills, and eventually employment. There is ample evidence to support the view that maternal education and family income are associated with children's development.(2) Education and income gains may produce changes in children's home environments, such as the provision of more cognitively stimulating materials or activities. These qualities of the home environment are positively associated with children's development,(3) and in fact are better predictors of child outcomes than are measures of parent education or socioeconomic status. However, despite evidence that higher parental educational attainment and family income are beneficial for children, we do not know whether JOBS participation will result in sufficiently large gains in these areas to influence outcomes among children.

Implementation of the JOBS mandate among AFDC mothers may also affect children's participation in non-maternal care. Meyers has summarized evidence that participation in welfare-to-work programs is associated with an increase in the amount of child care used and a greater reliance on formal child care arrangements, such as day care centers.(4) Mothers' participation in JOBS and the provision of child care subsidies for JOBS participants may result in more children from AFDC families participating in out-of-home, formal child care arrangements. High quality, educationally oriented child care programs are associated with cognitive gains, particularly for children from low-income families. Consequently, the FSA may provide an important opportunity to enhance the development of disadvantaged children. On the other hand, if parents place their children in poor quality care in order to fulfill their JOBS participation requirements, children's development may well suffer.

Furthermore, JOBS may affect children through changes in maternal psychological well-being. For example, mothers' stress or depression levels may increase in response to the participation mandate and the need to arrange child care fairly quickly. By contrast, mothers may experience decreases in depression and increases in role satisfaction or self-esteem as a result of gains in their education or job skills, or because participation provides social interaction, a respite from child care, and a sense of future opportunity. These areas of maternal well-being have been linked in turn to aspects of the home environment and to children's development.(5)

In sum, prior research suggests that JOBS, while primarily focused on parental education and employability, may affect the lives and well-being of young children as well. Both economic and non-economic mechanisms for such effects are possible. However, it is not known whether effects, if they do occur, will be positive, negative, or a mix of both; or whether any effects that are found will be large or modest. It is not known whether effects will differ by area of child well-being or for different subgroups of the JOBS population. The JOBS Child Outcomes Study has been designed to allow a careful examination of effects on children, as well as of the mechanisms by which such effects occur.

The JOBS Evaluation

The FSA legislation recommended a random assignment evaluation of the JOBS program to test its effectiveness, and this evaluation is currently being conducted by the Manpower Demonstration Research Corporation (MDRC). The impacts portion of the JOBS Evaluation involves random assignment of more than 55,000 JOBS eligibles to either a control group or one or two program groups, in seven sites around the country. The impact study is designed to examine the effects of various JOBS approaches on individuals' employment status, earnings levels, receipt and amount of AFDC payments, income levels, and educational attainment, using two types of experimental designs. The design of the impact study, and rationale for choosing each of the seven sites, are described fully in "The JOBS Evaluation: Early Lessons from Seven Sites".(6) Because JOBS departs from earlier welfare-to-work programs by mandating the participation of parents whose children are as young as three years of age, a special substudy of these parents and children, called the Child Outcomes Study, is being conducted within the larger JOBS Evaluation to examine outcomes for young children (see Figure 1).

The JOBS Child Outcomes Study Design

The JOBS Child Outcomes Study, part of the larger JOBS Evaluation, has been designed to examine both the effects of JOBS on children and the mechanisms that explain any effects that are found. Data for the Child Outcomes Study are being collected for approximately 3,000 mothers and children in three sites: Fulton County, Georgia; Riverside County, California; and Kent County, Michigan. The Child Outcomes sample includes all eligible families with a youngest child aged three to five who are enrolled in the JOBS evaluation in these three sites. Analyses of the impacts of the JOBS program for children will rely on follow-up data collected in these three sites from mothers and children two years after random assignment,(7) and from schools approximately four years after random assignment.(8)

The Descriptive Study Within the JOBS Child Outcomes Study

The current report provides a descriptive account of the Child Outcomes sample in one of these sites -- Fulton County, Georgia -- near the start of the evaluation. In the Fulton County site, the JOBS Evaluation is designed to measure the effectiveness of two alternative approaches to welfare-to-work programs: a human capital development approach, which emphasizes education and training activities, and a labor force attachment approach, which emphasizes quick entry into the job market through job search strategies. AFDC applicants or recipients in Fulton County who were subject to the JOBS mandate were randomly assigned to one of these two program groups, or to a control group. Those in the control group, while eligible for AFDC benefits, were not required to participate in any JOBS activities.(9)

This descriptive account of mothers and young children in the Fulton County site close to the start of the JOBS Evaluation will be referred to as the Descriptive Study, and the sample for this study as the Descriptive sample. For all participants in the JOBS Evaluation, including those in the Descriptive Study, we have baseline data, collected just prior to random assignment to either a program or control group. Baseline data include characteristics of the mothers and families at the time of random assignment, as well as a limited set of questions concerning maternal attitudes and subjective well-being. In addition, for the participants in the Descriptive Study, we also have data from a survey collected in respondents' homes on average three months after random assignment. This Descriptive survey included interviews with the mothers, assessments of the children, and direct observations of the home environment.

Seven hundred and ninety respondents from the JOBS Child Outcomes Study in Fulton County participated in the Descriptive survey. All are mothers whose youngest child was between the ages of three and five at the time of random assignment in the JOBS Evaluation, and all of these mothers were 20 years of age or older when they were assigned to a group within the JOBS Evaluation. Ninety-six percent are African American. Although none of the mothers were teenagers at the time of the Descriptive Study, 40 percent were 19 or younger at the birth of their oldest child living in the household. The present report refers to the child of between three and five years as the "focal" child, or the child whose circumstances and development were focused upon in the study. If the mother had two children between the ages of three and five, one was chosen randomly to be the focal child.

Key Questions and Selected Findings From the Report

The purposes of this report are to describe the lives and circumstances of this sample of AFDC families with preschool-aged children in Fulton County, Georgia and to inform policy makers about the mothers' goals and the development of their children. In addition, the study provides a context within which we will examine later impacts of the JOBS program on children. Below we summarize key findings from the report.

What is the community context of families in the Descriptive Study?

Fulton County, Georgia, includes most of the city of Atlanta, as well as suburban and rural areas. Compared to both the United States as a whole and U.S. metropolitan areas, Fulton County has higher rates of overall poverty, child poverty, and mother-headed households. Fulton County was selected as a site for the JOBS Evaluation because it represents a southern, urban site with a welfare population that is relatively disadvantaged compared to other sites (Hamilton and Brock, 1994).

Mothers in the Descriptive Study were asked to describe their neighborhoods. At the time of random assignment, about two-thirds of the sample reported that they lived in public (39 percent) or subsidized (29 percent) housing.(10) At the time of the Descriptive survey, about half of the sample (55 percent) reported that "very few" of the other mothers in their neighborhoods worked regularly at paid jobs. Four in 10 mothers described their neighborhoods as a "not too good" or an "awful" place to raise children, and about two in 10 mothers described their neighborhoods as an "excellent" or a "very good" place to raise children.

How job-ready are mothers in the Descriptive sample in terms of fertility plans, education, reading and math literacy, labor force experience, attitudes regarding work and welfare, and psychological well-being?

Mothers in the Descriptive sample varied substantially in terms of their apparent preparedness to pursue JOBS activities and employment. The majority of the mothers have had some previous experience in the labor force, although much of that experience was in low-paying, low-wage jobs. Two-thirds of the women are high school graduates or have a GED, suggesting that they are at a point where they could benefit from job training or further education, or could take an entry level job.

Despite the fact that most of the mothers in the Descriptive sample have a high school diploma or a GED, more than half of the mothers have low levels of basic reading and math literacy. Fifty-three percent of the sample have low levels of basic reading literacy, based on their scores on the Test of Applied Literacy Skills document literacy scale. Even among women with high school diplomas, 46 percent scored in this range. Sixty-two percent of the sample scored below seventh grade levels on the Greater Avenues for Independence (GAIN) Appraisal Math test. Only 14 percent of the women with high school diplomas scored in the highest level on the GAIN Appraisal Math test, indicating functioning at least at a high school entry level in basic reading and math. While the test scores point to low levels of reading and math literacy for many mothers in the Descriptive sample, we note that it is possible that having a high school diploma might be more important in acquiring a job than one's tested literacy level.

Data from the Descriptive Study contradict the stereotype that welfare mothers tend to have many children. Most mothers in the Descriptive sample have few children. Sixty-five percent had only one or two birth children living in the household at the time of the Descriptive survey, and only 13 percent had four or more birth children. Seventy-two percent of the households in the Descriptive Study consist only of the respondent and her child(ren). The total household size was small, with nearly three-quarters of the households composed of four or fewer people.

A woman's current fertility status and childbearing plans are important determinants of the likelihood that she will participate successfully in education and/or employment activities.(11) Women who want to have additional children may be a group particularly likely to drop out of JOBS activities due to pregnancy, whereas women who have already had all of the children they plan to have may participate more actively in JOBS. Most of the women in the Descriptive Study expressed a desire to limit their family size, with 96 percent neither being pregnant nor wanting to become pregnant. The majority reported using effective contraception or sterilization to avoid unwanted pregnancies (see Figure 2). Sixty-six percent of the women reported that they were not trying to become pregnant and were using a very reliable birth control method, such as the Pill, IUD, Depo Provera, or sterilization; the majority of these had a tubal ligation. On the other hand, 30 percent of the mothers responded that they were not trying to become pregnant, but were either using an unreliable method of birth control or were not using any birth control. As the women were not asked questions about their sexual activity, it is not clear whether contraceptive non-users are at risk of pregnancy or whether they are not sexually active.

Although there were variations in ratings, most of the respondents expressed positive attitudes toward employment, negative attitudes toward welfare, and a sense that they could locate child care if they become employed (see Figure 3).

In terms of educational attainment, attitudes about welfare and employment, and fertility status, many mothers in the Descriptive Study appear to be in a good position to participate in and benefit from JOBS. However, other characteristics of the mothers may impede their participation. A substantial proportion of mothers in the Descriptive Study (42 percent) reported depressive symptoms high enough to be considered in the clinically depressed range. Other studies using the same measure of depressive symptomatology in community-wide samples have found much lower rates of depressive symptoms, ranging from 9 to 20 percent. Further, most women in the Descriptive Study have experienced difficult life circumstances, including problems with housing, or having a relative or close friend in jail. Smaller groups of women also reported health-related barriers to employment or substance-use problems.

What assistance do the children's fathers provide to the mothers? Who other than the father provides emotional, childrearing, and economic support to these mothers, and to what extent?

Contact between the focal children and their biological fathers was limited. Only 16 percent of the mothers in the Descriptive sample had ever been married to the focal child's father. Further, only 2 percent of the children's biological fathers lived in the same household at the time of the Descriptive survey.

Mothers report that only one-fifth of the children with non-residential fathers had seen their fathers at least once a week in the year prior to the Descriptive survey. Mothers reported that only 10 percent of the fathers living outside of the household had "often" bought clothes, toys, or presents for the focal children; about 10 percent had "often" served as a babysitter for the focal children in the past year; and about 4 percent had "often" bought groceries in the past year. Mothers in the Descriptive Study did not often report the family of their child's father as an alternative source of support. Indeed, sixty-two percent of the mothers reported that over the past year the family of the child's father had done none of the following: bought clothes, toys, or presents, babysat, or cared for the child overnight.

Few of the mothers in the Descriptive sample reported the establishment of legal paternity for the focal child, and few reported formal child support agreements. Only 13 percent of the women who had never been married to the focal child's father reported having gone to a court or child support office to establish paternity, and 2 percent of the never-married women had had the biological father sign the birth certificate.

Among the women who did not reside with the child's biological father, 30 percent had ever had child support payments agreed to or awarded to them. Fifty percent of these child support arrangements that had been established were court-ordered, and about half (46 percent) arranged through a voluntary written agreement. Formal child support arrangements did not appear to assure payments. Among mothers in the Descriptive sample who had child support awards, 78 percent reported receiving no money from the father in the year prior to the interview. Among the mothers in the sample without a formal child support agreement, 88 percent reported that they had not received money on a regular basis directly from the father. Only 9 percent of these women reported that they had legal proceedings to establish paternity "in process" or have established paternity.

Mothers expressed great dissatisfaction with the emotional and financial assistance they were receiving from the children's fathers, and yet acknowledged that the fathers might not have been in a position to provide further economic assistance. Fifty-two percent of the mothers in this sample said that they were very dissatisfied with the amount of love and caring that their child's father has shown for the child, and an additional 10 percent were somewhat dissatisfied. Two-thirds of the mothers (66 percent) were similarly very dissatisfied with the amount of money and help that the father had been providing for raising the child. However, less than half of the sample, 41 percent, felt that the father could pay more for child support than he did, or could pay something if he currently paid nothing.

Despite the reported lack of involvement of their children's fathers, many mothers had other persons to turn to for emotional and instrumental support. Most mothers had frequent contact with members of their own families. For instance, 63 percent of the respondents who did not live in the same household as their own mothers saw their mothers once a week or more. About 33 percent of the Descriptive sample reported that their mothers helped to take care of their children "quite a bit" or "a lot." Only 10 percent of the mothers said that they had no one "who would listen to them, reassure them, or show them that they care." Most respondents did not feel overburdened by having other people ask them for their support. Many women had friends or relatives to turn to for economic or childrearing assistance as well. More than half of the respondents felt that it was true most or all of the time that they had someone who would lend them money in case of an emergency. However, mothers perceived instrumental support (e.g., economic assistance and help with childrearing) from these other sources to be less available than emotional support.

Mothers reporting low levels of social support were more likely to live in public or subsidized housing, to report high levels of depressive symptomatology, a limited sense of control over events in their lives, and more barriers to employment. Mothers with low levels of social support also had lower educational attainment and literacy scores.

How are the focal children in the Descriptive Study faring in terms of their cognitive development, school readiness, socioemotional development, and health at this early point in the JOBS Evaluation? Are there subgroups of children who are at greater risk in terms of their developmental status?

Children's developmental status was measured across several domains in order to provide a descriptive picture of child well-being. Direct assessments of cognitive development were obtained, one focusing on receptive vocabulary and one on school readiness. In addition, mothers reported on their perceptions of their children's socioemotional development and health status. At this early point in the JOBS Evaluation, children in the Descriptive sample appear to be faring poorly on assessments of their receptive vocabulary and school readiness, but not their health or social maturity as reported by their mothers.

The measure of receptive vocabulary used was the Peabody Picture Vocabulary Test-Revised (PPVT-R). This measure is highly correlated with measures of both intelligence and school achievement and is a predictor of IQ for African American as well as white children. Yet concerns about cultural bias have been raised regarding this measure, particularly the possibility that it underestimates the cognitive ability of minority children. As a result, we present comparative data from a national survey for African American children only.

Children in the Descriptive Study had a mean score of 70. By contrast, African American children from welfare families in a national sample had a mean score of 76 on this measure, and those from non-poor families had a mean score of 80 (see Figure 4). Thus, children in the Descriptive Sample had lower scores particularly than their non-poor peers in the national sample.

On average, mothers describe their children as showing fairly high levels of social maturity on the Personal Maturity Scale. Although maternal reports of personal maturity do not indicate a problem in this area, it must be noted that assessments of the child from a different source, such as a teacher, might result in a differing conclusion.

More than three out of four children were rated by their mothers as currently in excellent or very good health. Approximately half (49 percent) of the children were described by their mothers as in "excellent" health, and a further 29 percent were described as in "very good" health.

When the ratings of the child's health were combined with a maternal rating concerning the presence of conditions that limited the child's activities, 47 percent of the children in the Descriptive sample were described as in excellent health with no limiting condition. This proportion is lower than the proportion in a national sample of non-poor children. This generally positive portrayal of the children's health is in keeping with the fact that serious health problems in the child were a basis of mothers' exemption from JOBS.

Close to the start of the JOBS Evaluation, those children in the Descriptive sample showing the least optimal development are those whose mothers have the least education, and the lowest reading and math literacy skills, whose mothers feel the least control over events in their lives, and whose mothers perceive the most barriers to employment. In addition, boys in the Descriptive sample show less optimal development than girls on all four measures.

Finally, when mothers were asked to consider all of their children (including the focal child), a substantial minority (8 percent) reported that they had a child with an illness or handicap that demanded a great deal of attention or interfered with the mother's ability to work.

How supportive and stimulating are the children's home environments?

Findings from the Descriptive sample are in accord with previous reports that children living in poverty receive less cognitive stimulation and emotional support in their home environments than non-poor children. At the same time there is evidence of variability in the home environments of the families in the sample. This variability is related to family background characteristics, especially the extent of economic deprivation. In addition, the focal children's developmental status at the time of the Descriptive survey is significantly associated with the cognitive stimulation and emotional support they receive at home.

The Home Observation for Measurement of the Environment (Short Form), or HOME-SF, was used in the Descriptive Study to measure the emotional support and cognitive stimulation available in the home environment. Other analyses looking at the HOME-SF within the National Longitudinal Survey of Youth - Child Supplement indicate that this measure is closely related to several different indices of family poverty, and further, that the HOME-SF is sensitive to small increments in family income, particularly when looking at the home environments of children born into poverty.(12) Finally, the full HOME Scale,(13) from which the HOME-SF is adapted, has been found to be related to measures of child cognitive development and IQ, developmental delay, and poor school performance,(14) all important outcomes in the Descriptive sample. Families in the Descriptive sample showed, on average, similar levels of emotional support and cognitive stimulation to AFDC families with three- to five-year-olds in a national sample. However, scores were lower in the Descriptive sample and the national sample of AFDC families than in non-poor families in the national sample, both in terms of cognitive stimulation and emotional support in the home.

Within the Descriptive sample, scores on the two subscales of the HOME-SF were lower for families with specific characteristics. Mothers who had not received a high school diploma or GED, families receiving welfare for two or more years, families living in public housing, families with three or more children, mothers scoring in the lowest groups on reading and math literacy, and mothers with less of a sense of control over events in their lives, scored lower on both the cognitive stimulation and emotional support subscales of this measure of the home environment, net of control variables.

After controlling for the influence of child age, gender, and research group, the total score and cognitive and socioemotional subscales of the HOME-SF were all significant predictors of children's scores on the Descriptive Study's measures of receptive vocabulary, school readiness and children's maturity. In addition, higher scores on the cognitive stimulation subscale and the total HOME-SF scale were associated with children receiving a positive health rating from their mothers.

Are there changes in use of child care for the focal children in the Descriptive Study in the early months of the JOBS Evaluation?

Previous evaluations of welfare-to-work programs indicate that maternal program participation is associated with increased use of child care for young children.(15) In keeping with these earlier findings, there was a substantial increase in the proportion of Descriptive sample children in child care in the two program groups very shortly after enrollment in the JOBS Evaluation. Two months prior to random assignment, 44 percent of the three- and four-year-olds in the human capital development group were participating regularly in some form of child care, but two months after random assignment the figure was 72 percent. In the labor force attachment group, 48 percent of three- and four-year-olds were participating in child care two months prior to random assignment, but 83 percent were receiving some regular child care two months after random assignment. Over the same time period, use of child care in the control group increased only from 43 to 49 percent (an increase that probably reflects increasing child age and transitions to employment among control group mothers.)

Differential increases in the use of child care in the program groups relative to the control group occurred both for formal and informal care(16) settings, but as in previous studies of welfare-to-work programs,(17) we find a particularly marked increase in the use of formal child care settings following enrollment in JOBS.

The greater use of regular child care at the time of the Descriptive survey for the two program groups appears to be a reflection of their greater participation in employment and educational activities, not a differential propensity to use child care. There was a strong relationship between maternal participation in educational and/or employment activities following random assignment and the use of regular child care for the child. This relationship held in both the program and control groups. By the time of the Descriptive survey (on average three months after random assignment), program group mothers were substantially more likely to be participating in educational or employment activities than were control group mothers.

There was a statistically significant difference between research groups in the primary form of child care used by families with three- and four-year-olds at the time of the Descriptive survey. The most frequently reported primary care arrangement for children in the control group was care by the mother (used by 53 percent of control group families with three- and four-year- olds).(18) By contrast, care in a formal child care setting was the most frequently noted primary care arrangement for children in either program group (used by 53 percent of human capital development group children and 54 percent of labor force attachment group children).

Federal recommendations exist for formal child care settings in the form of the 1980 Federal Interagency Daycare Requirements (FIDCR). The 1980 FIDCR were never implemented as national regulations, yet researchers frequently refer to the FIDCR recommendations as a benchmark against which to measure the quality of center care. For children between three and five years of age, the FIDCR recommendations are for group sizes of 16 or smaller, and for staff-to-child ratios of 1:8 or better. The requirements for group size and ratio in the state of Georgia depart substantially from the FIDCR recommendations, allowing group sizes of up to 36 and ratios of up to 1:18 for four-year-olds. Among three- and four-year-old children in the Descriptive Study whose primary arrangement was a formal one, and for whom data on both group size and ratio were available, 34 percent were in settings that met both of these FIDCR recommendations; 17 percent were in settings that met one of the recommendations; and 49 percent were in settings that met neither recommendation.

Sixty-seven percent of mothers with three- and four-year-olds in some form of regular child care at the time of the Descriptive Study reported that someone else paid some or all of the cost of the primary care arrangement. The most common source of assistance, according to the mothers, was the welfare office. Among those mothers whose child had a regular child care arrangement, mothers in the program groups were more likely than those in the control group to receive assistance for child care from the welfare office. Sixty-seven percent of those in the human capital development group, 64 percent in the labor force attachment group, and 47 percent of the control group reported getting help from the welfare office.(19)

Only a minority (21 percent) of the Descriptive sample mothers reported paying anything towards the cost of the primary child care arrangement. Among those mothers with three- and four-year-old children who paid something for care, 74 percent reported paying $0.50 or less per hour.(20) Considering payments toward the cost of child care for all children in the household, mothers in our sample reported paying $19.11 per week on average. We note, however, that this figure does not take into account either the number of children in the household in care or number of hours in care.

Does mothers' psychological well-being, approximately three months after random assignment to the JOBS Evaluation, vary by baseline characteristics? How does the well-being of children differ by baseline characteristics?

Although all AFDC mothers are economically disadvantaged, as a group they vary substantially on several important characteristics that may be related to maternal and child well-being. For example, some families have been on welfare longer than others, and some have less education and lower literacy skills than others. Can we identify factors such as these, documented at the time of random assignment, that are associated with differences in the well-being of the mothers and children at the time of the Descriptive Study?

Measures of maternal and child well-being at the time of the Descriptive Study were examined in light of the following characteristics documented at baseline: maternal education, family size, duration of welfare receipt, residence in public or subsidized housing, reading and math literacy, depression, locus of control (sense of control over events in one's life), sense of social support, and perception of barriers to employment. Mother and child well-being at the time of the Descriptive Study varied significantly with regard to these baseline characteristics. The associations are profiled variable by variable in the full report.

It is noteworthy that in many instances, however, the same baseline characteristics that were associated with well-being among the mothers at the time of the Descriptive Study were also found to be related to their children's well-being. In particular, low maternal education, long-term welfare dependency, residence in public housing, low maternal reading and math literacy test scores, and poor maternal psychological well-being at baseline were all associated with lower scores on measures of the developmental status of the children, measures of the home environment, and measures of maternal circumstances at the time of the Descriptive Study.

Baseline characteristics can thus be used to identify meaningful subgroups of families who appear to be faring more and less well close to the start of participation in the JOBS Program. It will be important to track the development of mothers and children in these differing subgroups throughout the course of the JOBS Child Outcomes Study, asking whether participation and program impacts also differ.

How do multiple risk factors combine to affect children's well-being? Is the presence of protective factors associated with child well-being?

The analyses briefly summarized above consider whether the well-being of mothers and children at the time of the Descriptive Study differ for baseline subgroups considered one at a time, for example according to maternal education at baseline, or according to maternal depressive symptomatology at baseline. In reality, individual children will have differing profiles in terms of the number of baseline variables that place them at risk developmentally. Previous research suggests that the number of risk factors to which a child is exposed is an important predictor of development.(21)

Risk Factors and Children's Development

To explore the relationship between number of risk factors and children's well-being, we developed a cumulative risk index formed from the set of subgroup measures assessed at baseline prior to random assignment. These subgroups include maternal educational attainment and literacy; family size; welfare duration; maternal psychological well-being; and barriers to employment. Scores on the risk index range from 0 to 10 with a mean of 4.6 risk factors. The children divided nearly evenly into three groups according to the number of risk factors: Zero to three, four to five, and six to ten, indicating the presence in the Descriptive sample of children with few, some, and many risk factors.

Analyses indicate a strong association between the accumulation of maternal and family risk factors and the well-being of children in the Descriptive sample. Overall, 29 percent of the Descriptive Study children scored at or above the median for a national sample of African American preschool-aged children on the Peabody Picture Vocabulary Test-Revised.(22) However, the proportion of children with scores above the median was heavily concentrated among low-risk families, with 39 percent of children with zero to three risks scoring above this cutoff, compared to 17 percent among children with six to ten risks.

Scores on a measure of school readiness, the Preschool Inventory, show a similar pattern. Because national norms are not available for the Preschool Inventory, we have established a cut-point for this sample that identifies those children in the Descriptive sample whose scores are in the top quartile of the Descriptive survey distribution. Thirty-four percent of the children from low-risk environments scored in the top quartile, compared with 30 percent of children whose family environments posed four to five risks, and just 16 percent of those in very high-risk families (those with six to ten risk factors).

Children from low-risk family environments were also substantially more likely to be described favorably in terms of scores on the Personal Maturity Scale, while children from multiple-risk backgrounds were much less often described so positively. In addition, an increased number of risk factors is associated with a lower likelihood of being rated in excellent health with no disabilities. Specifically, 57 percent of children with zero to three risks received a positive health rating, compared to only 37 percent of those with six or more risks.

We also find a strong relationship between the number of risk factors and the emotional support and cognitive stimulation provided to the child as measured by the short form of the HOME Scale. Approximately a third of the Descriptive survey children in families with zero to three risks enjoyed home environments that were above a designated cutoff in terms of cognitive stimulation and emotional support, while only 12 percent of children in families with six to 10 risk factors experienced similarly supportive homes.

Protective Factors and Children's Development

Although increased risk is associated with poorer child outcomes overall, we see in these analyses that the presence of risk by no means guarantees that a child will exhibit adverse outcomes. Based on a typology of protective factors developed by,(23) we have used the measures of the Descriptive Study data to identify protective factors in each of the following categories: child characteristics, warmth and cohesion in the family, and an external support system. While our risk factors are all derived from baseline data, the protective factors are all based on data collected as part of the Descriptive survey. As for the risk factors, we have computed a summary index of protective factors. This ranges from zero to nine with a mean of 4.5 protective factors. We again group children into three groups according to the number of protective factors: zero to three, four to five, and six to nine.

To parallel the analyses looking at risk factors, we examined whether the number of protective factors was related to the proportion of children scoring above the cutoffs we defined on the same four measures of children's developmental status (the Peabody Picture Vocabulary Test, the Preschool Inventory, the Personal Maturity Scale, and rating of health). Results consistently indicate that, as the number of protective factors increases, a greater proportion of children score above the positive cutoff we delineated for each of the outcome measures. For example, the proportion of children scoring in the upper quartile on the Preschool Inventory increases from 15 percent among children with zero to three protective factors, to 36 percent among children with six or more protective factors. Similarly, the proportion of children in excellent health with no disabilities increases from 41 to 55 percent, as the number of protective factors increases.

Association Between Risk and Protective Factors

Does child well-being reflect the conjoint presence of risk and protective factors for the children in the Descriptive Study? To address this question, we grouped children according to their level of risk, and then within each risk group examined the proportion of children with favorable developmental status according to the number at each level of protective factors. We used the same categories of risk and protective factors described above, yielding a total of nine groups of children, ranging from those with few risk and few protective factors, to those with high levels of both.

Figure 5 shows that for the Peabody Picture Vocabulary Test, higher numbers of protective factors are associated with more optimal outcomes at each level of risk, while at the same time children at greater risk exhibit poorer outcomes overall. A similar pattern was observed for scores on the Preschool Inventory.

Figure 6 illustrates that the pattern is less clear when we consider children's socioemotional development. For children with fewer than six risks, more protective factors are generally associated with higher scores on the Personal Maturity Scale. However, for children with six or more risk factors, the presence of protective factors does not improve children's well-being. Regardless of the number of protective factors, only 14 to 17 percent of children with high levels of risk were rated by their mothers as having high levels of personal maturity.

These analyses illustrate that even within a sample of children who are all at risk by virtue of living in poverty, those with multiple risk factors are exhibiting less optimal development. Thus, the risks experienced by the mothers in the first generation are clearly translated into diminished opportunities for the children in the next generation. At the same time, a number of protective factors were found to be associated with more positive development for the children. For measures of cognitive development, protective factors offset the influences of risk factors. However, for our measure of socioemotional development, protective factors do not offset the influences of risk factors at the highest level of risk.

What Are the Implications of These Findings?

The JOBS program was designed to affect parents directly by providing services aimed at ending long-term welfare dependency. Nevertheless, indirect effects on children are also possible, if the JOBS program affects parental education, income, mother's psychological well-being, childrearing practices, or child care arrangements. The purpose of the current report is to explore the circumstances of eligible families at the outset of the program, rather than whether and how JOBS has impacted children. What have we learned?

A clear theme is that the mothers in the Fulton Descriptive sample are in many ways highly disadvantaged. On average, their reading and math literacy skills are low. Although they enjoy social support from family and friends, they report minimal economic or non-economic assistance from the fathers of their children. In addition, they have high rates of depressive symptoms and they experience numerous difficulties in the course of everyday life. At the same time, however, we note that most of the mothers in the sample had completed high school or a GED, most had positive attitudes about maternal employment, and most had taken steps to limit their childbearing.

Similarly, the three-to five-year-old children are also clearly disadvantaged at the outset of the JOBS program. As rated by their mothers, the children's maturity does not represent a problem; however, the children's receptive vocabulary is substantially below the mean for a national sample of children; and many of the children appear to lack the skills and knowledge that would make them ready to enter school. While a large majority of mothers in the sample described their children as in excellent or very good health, these ratings are somewhat less favorable than those reported in a national sample of non-poor children. Given that these children are already faring poorly in some respects, it seems entirely appropriate that policy makers, program providers, and the public consider whether and/or how the JOBS program may affect children.

A second recurring theme of the analyses is the heterogeneity of the population of welfare mothers eligible for JOBS. For example, some mothers hold positive attitudes about becoming employed, while a minority feel that mothers with young children should

not work. Some mothers have received AFDC for a much longer period of time than others. A substantial proportion of women have high levels of depressive symptoms, but many others do not. Most use reliable methods of contraception or have been sterilized, but a minority of mothers are at risk of an unplanned pregnancy. Because of this variation, it is likely that maternal participation in, and reactions to, JOBS activities will vary. Those mothers who are eager to work, know where they can obtain child care, and have recent employment experience seem more likely to respond favorably to the JOBS mandate. Other mothers face substantial obstacles to participation, such as low literacy levels, little support from family and friends, and negative attitudes about employment. It will be important to determine how both groups respond to the JOBS mandate.

Had the mothers proven to be more uniform in their work attitudes, goals, psychological well-being, skills, and the social support they receive, the JOBS mandate might have more uniform implications for children. However, early results indicating substantial subgroup variation suggest that the JOBS program is likely to elicit varied responses from both mothers and children. Hence, subgroup differences should be a critical component of further analyses. In particular, multiple risk families stand out as a group whose children are especially disadvantaged. On a more positive note, we were also able to identify a set of protective factors, greater numbers of which were associated with more positive child development. The mutual influence of risk and protective factors present at the start of the JOBS program may be an important determinant of both participation in, and impacts of the program.

Finally, the data suggest that the JOBS mandate is translating into initial changes in the lives of many AFDC mothers and their children. The effects of these apparent early changes will combine with any later program impacts on maternal education, earnings, and self-sufficiency. Thus, early data suggest that the JOBS mandate has the potential to affect the lives of two generations, and provide strong reason to track the well-being of both generations over time.





1) Wilson, J. B., and Ellwood, D. T. (1993). Welfare to work through the eyes of children: The impact on children of parental movement from AFDC to employment. Cambridge, MA: Malcolm Weiner Center for Public Policy, John F. Kennedy School of Government.

Zaslow, M. J., Moore, K. A., Morrison, D. R. and Coiro, M. J. (1995). The Family Support Act and children: Potential pathways of influence. Children and Youth Services Review, 17, 231-249.

2) Desai, S., Chase-Lansdale, P. L., and Michael, R. T. (1989). Mother or market: Effects of maternal employment on intellectual abilities of four-year-old children. Demography, 26, 545-561.

Duncan, G., Brooks-Gunn, J., and Klebanov, P. K. (1994). Economic deprivation and early childhood development. Child Development, 65, 296-318.

Hauser, R. M., and Mossell, P. A. (1985). Fraternal resemblance in educational attainment and occupational status. American Journal of Sociology, 91, 650-673.

3) Bradley, R. H., Whiteside, L., Mundfrom, D. J., Casey, P. H., Kelleher, K. J., and Pope, S. K. (1994). Early indications of resilience and their relation to experiences in the home environments of low birthweight, premature children living in poverty. Child Development, 65, 346-360.

4) Meyers, M. K. (1993). Child care in JOBS employment and training program: What difference does quality make? Journal of Marriage and the Family, 55, 767-783.

5) Downey, G. and Coyne, J. C. (1990). Children of depressed parents: An integrative review. Psychological Bulletin, 108 , 50-76

6) Hamilton, G., and Brock, T. (1994). The JOBS evaluation: Early lessons from seven sites. Washington, DC: U. S. Government Printing Office.

7) Anticipated sample sizes for the Child Outcomes Study Two-Year Follow-up survey are approximately 1,125 families in Fulton and in Riverside and approximately 750 in Kent.

8) The Department of Health and Human Services will be funding a four and one-half year follow-up of the JOBS-mandatory population. Further information about outcomes for children will be obtained at that time.

9) Respondents in the control group are not eligible for JOBS services, but are eligible for all other employment and training services in the community, and they can on their own obtain access to child care funded by the JOBS program.

10) Public housing project is operated by the local government to provide housing for low-income people. Receiving a rent subsidy, participating in a housing program like Section 8, or living in a building renovated by the government is not defined as living in a public housing project.

11) Long, S. (1990). Children and welfare: Patterns of multiple program participation. Washington, DC: Urban Institute Press.

Moore, K. A., Myers, D. E., Morrison, D. R., Nord, C. W., Brown, B. V., and Edmonston, B. (1993). Age at first childbirth and later poverty. Journal of Research on Adolescence, 3, 393-422.

12) Garrett, P., Ng'andu, N., and Ferron, J. (1994). Poverty experiences of young children and the quality of their home environments. Child Development, 65, 331-345.

Moore, K. A., Morrison, D. R., Zaslow, M. J., and Glei, D. A. (1994). Ebbing and flowing, learning and growing: Family economic resources and children's development. Paper presented at the Workshop on Welfare and Child Development sponsored by the Board on Children and Families and the National Institute of Child Health and Human Development's Family and Child Well-Being Network.

13) Caldwell, B. M. and R. H. Bradley. (1984). Home observation for measurement of the environment. Revised edition. Little Rock, AR: University of Arkansas.

14) Bradley, R. H., Caldwell, B. M., Rock, S. L., Barnard, K. E., Gray, C., Hammond, M. A., Mitchell, S., Siegel, L., Ramey, C., Gottfried, A. W., and Johnson, D. L. (1989). Home environment and cognitive development in the first 3 years of life: A collaborative study involving six sites and three ethnic groups in North America. Developmental Psychology, 25, 217-235.

Elardo, R. D., Bradley, R. H., Caldwell, B. M. (1975). The relation of an infant's home environment to mental test performance from six to thirty-six months. Child Development, 46, 71-76.

Gottfried, A. W. (1984). Home Environment and Early Cognitive Development. Orlando, FL: Academic.

15) Kisker, E., and Silverberg, M. (1991). Child care utilization by disadvantaged teenage mothers. Journal of Social Issues, 47, 159-177.

Meyers, (1993).

Quint, J. C., Polit, D. F., Bos. H. and Cave, G. (1994). New Chance: Interim findings on a comprehensive program for disadvantaged young mothers and their children. New York: Manpower Demonstration Research Corporation.

Riccio, J., Friedlander, D., and Freedman, S. (1994). GAIN: Benefits, costs, and three-year impacts of a welfare-to-work program. New York: Manpower Demonstration Research Corporation.

16) Formal care includes care in child care centers, preschools, nursery schools, Head Start, kindergarten, and before-and-after school programs. Informal care includes care by a relative or non-relative babysitter.

17) Kisker and Silverberg, (1991); Meyers, (1993); Quint et al., (1994); Riccio et al., (1994).

18) The children's primary care arrangement is the arrangement that they were in for the most hours each week. This can include sole maternal care.

19) These differences were statistically significant; Chi square (2) = 8.29, p < .05.

20) We note that this figure does not take into account variation in cost per hour according to number of hours in care.

21) Rutter, M. (1979). Protective factors in children's responses to stress and disadvantage. In M. W. Kent and J. Rolf (Eds. ), Primary prevention of psychopathology, Vol. III: Social Competence in Children. (pp. 49-74). Hanover, NH: University Press of New England.

Sameroff, A. J., Seifer, R., Barocas, R., Zax, M., and Greenspan, S. (1987). Intelligence quotient scores of 4-year-old children: Social-environmental risk factors. Pediatrics, 79, 343-350.

22) We used a cutoff based on the median score for African American children because of concerns that the Peabody Picture Vocabulary Test, like many other tests of achievement, may be racially biased (but comparable results were obtained using a standard cutoff).

23) Garmezy, N. (1985). Stress-resistant children: The search for protective factors. In J. E. Stevenson (Ed. ), Recent Research in developmental psychology (pp. 213-233). Oxford: Pergamon Press.


Publications from the JOBS Evaluation

[Before 1995]

From Welfare to Work. Judith M. Gueron and Edward Pauly. 1991. New York: Russell Sage Foundation.

The Saturation Work Initiative Model in San Diego: A Five-Year Follow-up Study. Daniel Friedlander and Gayle Hamilton. 1993. New York: Manpower Demonstration Research. Corporation.

The JOBS Evaluation: Early Lessons from Seven Sites. Gayle Hamilton and Thomas Brock, 1994. Washington, D.C.: U.S. Department of Health and Human Services and U.S. Department of Education.

Five Years After: The Long-Term Effects of Welfare-to-Work Programs. Daniel Friedlander and Gary Burtless. 1995. New York: Russell Sage Foundation.

The JOBS Evaluation: Adult Education for People on AFDC - A Synthesis of Research. U.S. Department of Education and U.S. Department of Health and Human Services. 1995. Washington, D.C.: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation.

The JOBS Evaluation: Early Findings on Program Impacts in Three Sites. U.S. Department of Health and Human Services and U.S. Department of Education. 1995. Washington, D.C.: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation.

The JOBS Evaluation: How Well Are They Faring? AFDC Families with Preschool-Aged Children in Atlanta at the Outset of the JOBS Evaluation. U.S. Department of Health and Human Services and U.S. Department of Education. 1995. Washington, D.C.: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation.

The JOBS Evaluation: Monthly Participation Rates in Three Sites and Factors Affecting Participation Levels in Welfare-to-Work Programs. U.S. Department of Health and Human Services and U.S. Department of Education. 1995. Washington, D.C.: U.S. Department of Health and Human Services, Office of the Assistant Secretary of Planning and Evaluation.