Welfare reform policies are sometimes referred to as "two generational" because not only are the lives of the parents changed by government welfare-to-work programs, but the lives of the children are changed as well. At the most basic level, children's time use patterns and child care patterns are likely to change. Changes in parental education and/or family income -- the target of welfare reform efforts -- also have potential implications for children. Moreover, other changes in parent-child relationships and family interaction patterns may come about as a result of mandated participation in welfare-to-work activities. The question increasingly asked by policy makers is whether the varied changes set in motion by welfare reform policies affect the development of the children in the families of welfare recipients.
This report is one of a series focusing on a set of welfare-to-work strategies implemented under the Job Opportunities and Basic Skills Training (JOBS) Program(1) as a part of the 1988 Family Support Act. The National Evaluation of Welfare-to-Work Strategies (NEWWS) (formerly known as the JOBS Evaluation) is a longitudinal evaluation of JOBS Program approaches, being carried out in seven sites across the country: Atlanta, Georgia; Grand Rapids, Michigan; Riverside, California; Columbus, Ohio; Detroit, Michigan; Oklahoma City and surrounding counties in Oklahoma; and Portland, Oregon. The Family Support Act established basic requirements for JOBS programs nationwide -- including the mandatory nature of welfare clients' participation -- but it allowed states considerable flexibility on matters such as the exact sequence and content of services (Hamilton, et al., 1997).
In the present report, we focus on a special study carried out with a subset of families participating in the National Evaluation of Welfare-to-Work Strategies: the Child Outcomes Study. This special study provides an in-depth examination of the development and well-being of children who were preschool-age when their mothers enrolled in the evaluation. It also examines in some detail the family context of these young children. We focus both on describing the well-being of the children and families in the absence of a JOBS program (that is, presenting outcomes for children and families in the control groups in the study sites), and on the impacts of the JOBS programs (that is, statistically significant differences on measures of child, adult, and family outcomes for program and control group families in the study sites).(2)
A companion report presents findings regarding economic impacts on families and impacts on a limited number of brief measures of children's well-being for all the children in the families in all seven research sites, two years after enrollment in the evaluation (Freedman, S., Friedlander, D., Hamilton, G., Rock, J., Mitchell, M., Nudelman, J., Schweder, A., and Storto, L., 2000). A third report (Hamilton, with Freedman and McGroder, 2000) synthesizes results from this latter report and the present report, integrating the brief findings relating to children in the full evaluation sample (consisting of families with children of any age) and the in-depth examination of findings relating to children for the Child Outcomes Study sample (consisting only of families with preschool-age children at baseline).
The Child Outcomes Study
The Child Outcomes Study is being carried out in three of the seven research sites of the NEWWS (Atlanta, Georgia; Grand Rapids, Michigan; and Riverside, California). It includes families in which there was at least one child between the ages of three and five at enrollment -- the families newly required to participate in welfare-to-work activities by the Family Support Act. In the Child Outcomes Study sites, the NEWWS is examining the impacts of two types of program approaches which represent very different theories of how government might best help recipients of public assistance move toward economic self-sufficiency (Hamilton et al., 1997). The "labor force attachment" (LFA) approach encouraged a rapid transition into the labor force, whereas the "human capital development" (HCD) approach followed a longer term strategy of investing in recipients' basic education, with the aim of increasing their qualification for higher wage jobs.
Eligible individuals were randomly assigned to one of these programs or they were placed in a control group. Both the HCD and LFA programs were composed of a "package" of components, including: (1) mandated participation in education, training, and/or employment activities, (2) messages about the importance of such activities for eventual employment, (3) enhanced case management services to direct and monitor clients' progress, and (4) services to facilitate employment. In each program approach, case managers could request or impose sanctions for mothers not complying with the participation mandate. Actual participation rates varied for families in the larger evaluation: Only half or fewer of mandatory enrollees in Riverside's two programs participated in any work preparation activity for at least a day since random assignment, compared to about two-thirds of mandatory enrollees in Grand Rapids' programs, and between 61 percent and 74 percent of mandatory enrollees in Atlanta's HCD and LFA programs, respectively. Among those participating, the average length of stay in activities ranged from 3 months (in Riverside's LFA program) to 9.4 months (in Atlanta's HCD program).(3)
Mothers in the control groups, however, were not subject to mandated participation in JOBS welfare-to-work programs and, thus, received no self-sufficiency messages, no enhanced case management, no JOBS services, and were not mandated to participate in any activities (but could seek out available education and training in their communities). They continued to be eligible for Aid to Families with Dependent Children (AFDC) and, because the Family Support Act established an entitlement to child care for all welfare recipients in "approved" job preparation activities, mothers in the control groups (as well as those in each program group) were guaranteed child care.(4)
Enrollment into the Child Outcomes Study occurred between September 1991 and January 1994.(5) Two years after enrollment, the families were visited by carefully trained interviewers. Mothers were interviewed about work and family topics and about their children, with more detailed information obtained about one specific child, the "focal child," who was between the ages of about three and five years at the time of random assignment. In addition, developmental assessments of focal children's cognitive school readiness were conducted. The developmental status of the children was assessed across three domains: cognitive development and academic achievement, behavioral and emotional adjustment, and physical health and safety.
There is only a small body of previous research that examines the effects of welfare-to-work programs on children (see summary in Zaslow, Tout, Smith and Moore, 1998). Given the limited previous research, rather than beginning with a specific hypothesis about the direction of impacts for children (that is, with an expectation that assignment to a JOBS program would be associated with favorable, or alternately with unfavorable impacts on children), this study was undertaken with the aim of exploring a range of possibilities: (1) that there would be no discernable impacts on children; (2) that mothers' assignment to a JOBS program would result in favorable impacts on children; (3) that mothers' assignment to a JOBS program would result in unfavorable impacts on children; (4) that there would be impacts on children only in specific subgroups of families. The study was designed with sufficient power to reasonably assess whether there were harmful impacts on children, an important hypothesis to examine when the new provisions of the Family Support Act were being implemented. The design also has sufficient power to assess whether favorable impacts occurred or whether different impacts occurred for different subgroups. Unfavorable impacts on children would undermine the policy goal of enhancing economic self-sufficiency across the generations in families receiving welfare.
JOBS programs did not include components addressed specifically at improving the development and well-being of young children. For example, they did not include early childhood education components, or screening of children for health or developmental problems. Accordingly, in order to affect children, JOBS must have affected outcomes of importance to children's development. Outcomes considered are those directly targeted by the programs (such as employment, earnings, and overall family income) as well as derivatives of these targeted outcomes (such as health insurance and child care use), and those not directly targeted by the programs but which, nevertheless, could have been affected by them (such as maternal psychological well-being, and parenting behavior).
Our conceptual framework for the ways in which children's developmental outcomes may be affected by welfare-to-work programs is presented in Figure 1.1. This conceptual framework illustrates that:
- assignment to a JOBS program (Box A), could affect
- mothers' exposure to, and experiences in, the JOBS Program as the program is actually implemented (Box B) and, through program exposure, affect
- outcomes directly targeted by the JOBS Program (such as maternal educational attainment, employment, and economic resources) as well as derivatives of these targeted outcomes (such as health insurance and child care use) (Box C), as well as affect
- additional outcomes not directly targeted by JOBS programs (such as maternal psychological well-being and parenting) (Box D) and,
- through these changes, affect child outcomes (Box E).
(Chapter 8 provides a detailed discussion of this conceptual framework, including rationales for the specific outcomes considered in this study and a description of measures employed to operationalize these outcomes.)
To provide a context for this study, a number of key questions are addressed in this introductory chapter.
Key Questions Addressed in Chapter 1
- Why are outcomes for children considered in a program aimed primarily at parents?
- What child outcomes might be affected by welfare-to-work approaches under JOBS?
- Through what pathways might welfare-to-work approaches under JOBS affect children?
- Why are findings for welfare-to-work approaches under JOBS still relevant, given the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996?
Why Are Outcomes for Children Considered in a Program Aimed Primarily at Parents?
Children's development is widely understood to be influenced by multiple contexts -- including family, neighborhood, child care, and school. These contexts, in turn, are influenced by larger societal and structural circumstances, such as economic conditions, culture, and public policy. This perspective is described as the ecological approach (Bronfenbrenner, 1979; 1986) and represents the starting point for much of the current research on children, including this study. This perspective clearly posits that social programs aimed primarily at parents may also affect their children, by changing their experiences both outside of the home (for example, as children's care arrangements are altered) and/or within the home (for example, if the program influences parent-child interactions).
Evaluations of programs that have successfully altered outcomes for children at risk for problems in development tend to adopt an ecological approach, that is, they "see the child in the context of the family, and the family in the context of its surroundings" (Schorr, 1991, p. 267). Further, programs directed at parents are increasingly being examined for their effects on children in the families. Thus, in the Child Outcomes Study, the circumstances and development of the children living in welfare families are considered in light of their mothers' assignment to a JOBS labor force attachment or human capital development welfare-to-work program.
A large number of evaluations have examined the implications of job training and welfare-to-work programs on economic outcomes for adults. Established in 1967, the Work Incentive (WIN) program provided education, job training, and social services to AFDC recipients on a primarily voluntary basis, evolving into a mandatory job search program for mothers with children six and older (Hamilton and Brock, 1994). A modification of WIN, California's federally-funded Saturation Work Incentive Model (SWIM) demonstration, focused more heavily on education and training and less on immediate job search, compared to other WIN models. Evaluations of selected WIN programs and SWIM indicated that, compared to control group clients, program group clients obtained jobs more quickly and witnessed small increases in earnings and a reduction in AFDC benefits, with earnings increasing over a five-year period (Friedlander and Burtless, 1995). Moreover, the most persistent long-term earnings impacts appeared to be concentrated in programs offering education and training and not just job search activities (Friedlander and Burtless, 1995).
These and other studies suggest that small to moderate effects on employment are typical, and that reductions in public transfers are common (Gueron and Pauly, 1991). However, large impacts on total family income are less likely, often because gains in earnings are offset by reductions in AFDC benefits (Aber, Brooks-Gunn, and Maynard, 1995; Friedlander and Burtless, 1995). Consequently, clients may move off welfare and join the ranks of the working poor (Chase-Lansdale and Vinovskis, 1995). Recent findings from three sites of the National Evaluation of Welfare-to-Work Strategies indicate that clients in the labor force attachment program in each site experienced gains in employment and earnings averaging more than $1,000 at the two-year follow-up point, clients in the human capital development program in two of the three sites also experienced gains (albeit, smaller) in employment and earnings, and clients in both programs in each site were less likely to be receiving welfare (Hamilton, Brock, Farrell, Friedlander, and Harknett, 1997). It appears that both the labor force attachment and human capital development strategies of the mandatory JOBS Program were more effective than the earlier (often voluntary) welfare-to-work approaches, at least in the short run. Extended follow-up is important to understand the efficacy of both JOBS approaches -- especially the human capital development strategy -- for improving the long-term economic circumstances of families.
A small but growing evaluation literature also examines the implications of economic security programs for the development of children, as well as for adult economic and non-economic outcomes. These studies provide evidence that interventions for disadvantaged families can have effects on multiple family members. One early example is a study of the Negative Income Tax Experiment, which provided a guaranteed minimum income to a sample of low-income families in several communities and was associated with an increase in children's reading achievement (Maynard and Murnane, 1979; Murnane, Maynard, and Ohls, 1981).
While as yet only a handful of evaluation studies among populations receiving welfare have considered possible implications for children, a very substantial body of basic research studies has documented the linkages between family characteristics and processes, and the well-being and development of children. Studies indicate, for example, that family structure is strongly related to children's development (McLanahan and Sandefur, 1994), that family size and child spacing are strongly related to educational outcomes for children (Blake, 1989), that maternal child rearing practices predict children's developmental outcomes (Maccoby, 1999; Maccoby and Martin, 1983; Bornstein, 1995), that fathers play an important if not completely understood role in the development of their children (Federal Interagency Forum on Child and Family Statistics, 1998; Lamb, 1997; Marsiglio and Day, 1997; Mosley and Thomson, 1995), that maternal psychological well-being affects mother-child interaction (Belsky, 1981; 1984; Coiro, 1997) and thereby children's development (McGroder, 2000; McLoyd, 1990), and that poverty and economic insecurity undermine children's development (Duncan and Brooks-Gunn, 1997). Thus, JOBS may affect children's development indirectly, to the extent that JOBS influences these family characteristics and processes.
Not only do programs for mothers receiving public assistance have the potential to affect the well-being of children, but programs aimed at children also have been found to have effects on parents. A review of programs that provide educationally-oriented intervention services for young children suggests that such programs can result in higher rates of maternal employment and more stable employment (Benasich, Brooks-Gunn, and Clewell, 1992). Thus, the child care benefits provided by welfare-to-work approaches under JOBS may not only affect children directly, they may also affect mothers' participation in JOBS activities and, through such participation, child outcomes.
The recognition that interventions can affect both children and parents has led to the implementation of programs that are explicitly two-generational in focus (Smith and Zaslow, 1995). Such programs "pursue the dual goals of economic self-sufficiency for families and healthy development of children" (Smith, Blank, and Collins, 1992, p. 2). Smith and Zaslow suggest that two-generational programs have the potential to improve children's well-being in the long term by focusing not just on one area of family functioning but on the multiple challenges and needs of at-risk families.
Several recent two-generational programs -- Even Start, Project Redirection, New Chance, and the Teenage Parent Demonstration -- provide evidence that such interventions can have effects on the experiences of both mothers and children. For example, the Even Start family literacy program, which provides early childhood education, parenting education, and adult education to disadvantaged families, had a positive effect on the presence of reading materials in the home (St. Pierre, Swartz, Murray, Deck and Nickel, 1993).
Project Redirection, a demonstration initiated by Manpower Demonstration Research Corporation (MDRC) that offered comprehensive services on a voluntary basis to economically disadvantaged mothers aged 17 and younger, was found to have positive effects on mothers' warmth and acceptance toward their children and on language stimulation in the home, as well as modest, but significant, positive effects on the cognitive and socioemotional development of participants' children five years after the program began (Polit, Quint, and Riccio, 1988). Children in the Project Redirection group were also more likely to have attended Head Start. These program impacts for children were more powerful than the long-term economic impacts found for the mothers who participated in Project Redirection.
New Chance was a comprehensive, voluntary demonstration program for young women on AFDC who had given birth to their first child at age 19 or younger and who had dropped out of school. Mothers attended classes on GED preparation, parenting, life skills, family planning, and occupational skills. Children of enrollees received child care while their mothers participated in the program and pediatric health services, as needed. The evaluation found impacts on the educational activities and attainment of mothers in the experimental group at both the 18- and 42-month follow-ups, compared to those in a control group who did not have access to the New Chance Program (but who did have access to other services and programs in their communities). Not surprisingly, given the child care component of the New Chance Program, children of experimental group mothers were more likely to have been in non-maternal care, especially center-based care, at the 18-month follow-up (Quint, Polit, Bos, and Cave, 1994), though there were no longer differences in child care participation at the 42-month follow-up (Quint, Bos, and Polit, 1997). Program children's home environments were slightly more emotionally supportive, and their mothers reported somewhat less punitive child rearing attitudes at the 18-month follow-up (Quint et al., 1994). Results from analyses of observational and survey data collected approximately 21 months after random assignment for a subset of families within the New Chance Evaluation also point to small reductions in the incidence of harsh parenting behavior and increases in the quality of maternal book reading to children. In addition, mothers in the experimental group reported spending more time in parenting chores, reported greater warmth in their relationships with their children, and were found to have a more supportive home environment than mothers in the control group (Zaslow and Eldred, 1998).
However, by the 42-month follow-up, the only positive impacts on children's home environment occurred for the subgroup of mothers who had shown lower levels of depressive symptoms at baseline. Also, by the 42-month point, there were negative program impacts on maternal report measures of the children's social behavior, though subgroup impact analyses revealed these to be significant only for specific subgroups of families, for example, those families in which mothers had shown higher levels of depression at baseline (Quint, Bos, and Polit, 1997). These findings suggest that multiple sources of influence on children (such as the positive program impacts on parenting but negative impacts on maternal psychological well-being) can accumulate and offset each other (Zaslow and Eldred, 1998).
Evaluators hypothesize that New Chance may have raised expectations of program mothers beyond what the program was able to do to alter their life circumstances, resulting in "dashed hopes," manifested in more depressive symptoms and dissatisfaction with economic circumstances compared to control group mothers (Quint, Bos, and Polit, 1997). In sum, the New Chance Evaluation illustrates the importance of examining shorter and longer-term program impacts on mothers and children; the possibility that program impacts are concentrated in certain subgroups; the potential for two-generational programs to have both positive and negative effects on mothers and the family environment, leading to either positive or, in the case of New Chance, negative impacts on some child outcomes; and the difficulty in making a positive impact on mothers and children in families experiencing numerous life stressors.
In 1986, the federal government funded the Teenage Parent Demonstration, an experiment targeted to those most likely to become long-term welfare recipients: teenage parents. Anticipating many features of the Job Opportunities and Basic Skills Training (JOBS) Program established in the Family Support Act of 1988, experimental group mothers were mandated to participate in educational and/or employment activities, as well as in workshops on personal skills, problem-solving, and parenting, and they faced financial sanctions for non-participation. At the same time, they were provided child care, help with transportation, and informal counseling by case managers in order to support the transition from welfare to school and work.
Evaluation results indicate that, after two years, program mothers were significantly more likely to be involved in self-sufficiency activities and had higher monthly earnings and lower AFDC benefits compared to a control group (Aber, Brooks-Gunn, and Maynard, 1995). An embedded observational study of a subsample of evaluation mothers and their three- to five-year-old children revealed no impacts on mothers' positive or harsh parenting and no impacts on their children's enthusiasm, persistence, or anxiety (as rated by observers); sociability or mental health problems (as reported by mothers); or verbal ability (as assessed by the Peabody Picture Vocabulary Test-Revised, PPVT-R). However, non-experimental analyses examining parenting and child outcomes for mothers at different levels of program participation revealed that mothers who did not participate in any self-sufficiency activity displayed more negative emotion toward their children and used more harsh control, and their children displayed less persistence and enthusiasm, compared to mothers who participated in the program at moderate or high levels (Aber et al., 1995). Of course, mothers who chose to participate to a greater degree may differ on pre-existing characteristics from mothers who "self-select" into non-participation; hence, these results cannot be unequivocally attributed to the Teenage Parent Demonstration.
Overall, these findings suggest that programs aimed at either parents or children can also affect the experiences of the other generation. In particular, the available evidence indicates that the experiences of both mothers and children, in and out of the home, can be changed. The ultimate implications of varied program approaches for the developmental trajectories of children are not yet fully understood. Examination of such two-generation effects represents the explicit purpose of the NEWWS Child Outcomes Study. Thus, although welfare-to-work approaches under JOBS are primarily directed at parents, we must consider the potential effects on both the parental and child generations.
What Child Outcomes Might Be Affected By Welfare-to-Work Approaches Under JOBS?
Children's development and well-being have been categorized in a variety of different ways (Moore, Evans, Brooks-Gunn and Roth, 1998; Zill and Coiro, 1992); however, there is a consensus that no one dimension of child well-being adequately describes the well-being of a child. Indeed, the report "Trends in the Well-being of America's Children and Youth: 1998" (Office of the Assistant Secretary for Planning and Evaluation, 1998) reports on more than eighty indicators of child and family well-being, and the report "America's Children: Key National Indicators of Well-being (Forum on Child and Family Statistics, 1997) identifies 26 measures of child well-being.
In this study, three broad aspects (or "domains") of development were identified for study:
- cognitive functioning and academic achievement;
- behavioral and emotional adjustment (including both problem and positive behavior); and
- physical health and safety.
Within each of these three domains, we have selected multiple measures of children's development and well-being. There are two primary reasons for examining an array of measures. First, welfare-to-work approaches under JOBS have the potential to have far-reaching effects on the lives of families, affecting not only parental work, parental education, family income, and child care, but also, time use patterns, parental psychological well-being, and both the amount and quality of parent-child interaction. Hence, the potential implications for the well-being of children are necessarily broad as well. In addition, since welfare-to-work approaches under JOBS represented a new approach to encouraging self-sufficiency among welfare families, and given the minimal amount of two-generation research on welfare programs, there is a dearth of evidence to suggest that effects are only to be expected in one or two domains of child well-being. Finally, because "established" child development measures and assessments are rarely developed or normed on non-normative (e.g., low-income, racially/ethnically diverse) populations, it is important to examine multiple measures within each domain of development to provide a more complete picture of impacts among children in low-income families.
Through What Pathways Might JOBS Affect Children?
There are multiple pathways by which JOBS welfare-to-work approaches could affect children (Zaslow, Moore, Morrison, and Coiro, 1995). JOBS may affect children by bringing about changes on outcomes explicitly targeted by the program (such as maternal education and employment and earnings), as well as derivatives of these outcomes (such as health insurance and child care use), and/or by bringing about changes in outcomes not targeted by the program (such as maternal psychological well-being and parenting behavior).
For example, with respect to outcomes targeted by JOBS programs, under the human capital development program approach, mothers are directed toward activities to enhance their basic educational skills. Improving their own reading and math skills may enable mothers to be more supportive of their children as the children enter school and begin learning to read and do arithmetic. In addition, mothers who are attending classes and doing homework provide a role model for their children that emphasizes the value of learning and study. This may lead the children to focus more on school. And to the extent that improved skills translate into better employment opportunities in the long run, these mothers may also experience increases in income, which can benefit children as described above.
If, as intended, mothers assigned to both human capital development and labor force attachment programs do experience program impacts in employment and earnings, families may be better off financially. Improvements in a family's economic situation may lead the family to move to a better neighborhood, invest in books and stimulating toys and experiences, and purchase better quality food and child care. In addition, higher income may improve maternal psychological well-being, leading her to be more emotionally supportive of her child (Goodman and Brumley, 1990; McLoyd, 1990).
Child care represents another pathway by which JOBS might affect children, given the increased time children may be expected to spend in non-maternal care as their mothers participate in educational and work-preparation activities or in actual employment. However, the effects of child care on children depend on the quality and consistency of care more than on the simple fact that a child attends care (Hayes, Palmer, and Zaslow, 1990; Smith, 1998). Hence, the effects on children of entering care are likely to be complex.
The implications of mandated participation in JOBS for mothers' access to health insurance and to health care services represents another pathway through which children may be affected. JOBS programs target health insurance and access to health care services in that children are covered by Medicaid while receiving AFDC and, potentially, for up to 12 months while their mothers transition off welfare. Yet few entry-level jobs for which these mothers would qualify offer health benefits. After the transitional 12 month period, it is possible that mothers who have transitioned off of welfare would lack health insurance for themselves and their children. Children's health may suffer if mothers delay or do not even seek treatment or well-child care due to lack of health insurance.
While JOBS programs did not seek to bring about changes in maternal psychological well-being and parenting behavior, these aspects of family life could nevertheless be affected by a mothers' assignment to JOBS. Given the evidence on the close linkages between maternal psychological well-being and parenting behavior with child outcomes (Belsky, 1984; McGroder, 2000; McLoyd, 1990), these aspects of family life also constitute pathways through which mandated participation in JOBS may affect children (McGroder, 2000; Moore, Zaslow, Coiro, Miller, and Magenheim, 1995; Zaslow, Moore, Morrison, and Coiro, 1995).
Contrasting hypotheses have been articulated for how assignment to a JOBS program could affect maternal psychological well-being and parenting behavior (Zaslow, Dion, and Sargent, 1998; Zaslow et al., 2000). A "stress hypothesis" posits that the requirement to participate in educational or employment activities requires a reorganization of family life and adaptation to new care situations and routines by all family members. Mothers with limited education and/or employment experiences might find welfare-to-work activities frustrating and difficult, at least initially. If the changes in routine and daily activities engender stress, this could have unfavorable implications for parenting behavior (for example, resulting in greater harshness in mother-child interactions). A "stimulation hypothesis," by contrast, posits that maternal involvement in educational or employment activities outside of the home could serve as sources of stimulation to mothers, which in turn could be reflected in the level of cognitive stimulation that they provide to their children. If mothers improve their educational attainment and gain job skills, they could experience enhanced self-esteem and competence. Over time, improved family economic circumstances could decrease stress and improve maternal supportiveness in interactions with children.
Why Are Findings Still Relevant, Given the Passage of the "Personal Responsibility and Work Opportunity Reconciliation Act of 1996"?
With passage of the new welfare law, the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996 (P.L. 104-193), the present welfare policy environment differs from that under the Family Support Act in a number of important respects (see Zaslow, Tout, Smith and Moore, 1998). For example, the 1996 law ends the entitlement to cash assistance (the AFDC program) and, instead, requires that recipients be employed within two years of applying for temporary cash assistance (now called, Temporary Assistance for Needy Families, or TANF). A certain percentage of the caseload is expected to work or be in work-related activities, with the percentage increasing from 25 percent of all families in 1997 to 50 percent in 2002. States face penalties if they do not meet these participation goals. The work requirement can be fulfilled through employment (either subsidized or unsubsidized), on-the-job training, or community service. Vocational training can address the work requirement for up to 12 months, while job search activities can fulfill the requirement for up to 6 weeks. Participation in school can fulfill the work requirement only for recipients under age 19.
Under the new law, there is a total lifetime limit of five years for the receipt of assistance through federal funds. Whereas the Family Support Act required mothers of children three and over to participate in work-related activities (with a state option to lower the age to 12 months), the new law lowered the age for all states to 12 months (or younger, if a state chooses). While under the Family Support Act, mothers were exempt from participation in welfare-to-work activities if they were needed in the home to care for a sick or incapacitated family member (including a child), there is no such explicit exemption under PRWORA. While PRWORA strengthens the requirements to work, at the same time there are also greater supports for work, with almost all states now treating earnings and assets more generously, and significantly more federal funds available for child care.
Both the welfare-to-work approaches under JOBS and the new welfare law mandated participation in welfare-to-work activities, and both sought clients' employment as the ultimate goal. However, the time-limited assistance of the new welfare law is intended to serve as an added incentive for clients to seek employment quickly and to engage in employment-preparation activities only inasmuch as they better prepare clients for immediate employment. Moreover, states face major financial penalties if they do not meet mandated participation rates, which is likely to result in a program focus on employment and limited work-related activities, rather than on longer-term education and training.
If policy makers are seeking to build a body of knowledge about how a range of welfare policies affect families and children -- in terms of both economic impacts as well as impacts on family functioning and child development -- then the JOBS labor force attachment and human capital strategies remain distinctive approaches that will provide an important benchmark against which to compare new policies and programs. Moreover, the HCD and LFA strategies remain viable approaches under TANF for many states seeking ways to foster the transition from welfare to work. New welfare-to-work approaches will need to be assessed not only in terms of initial transitions into employment and reductions in welfare caseloads, but also in terms of the stability of employment over time, longer-term impacts on family income, the likelihood of resuming the receipt of public assistance, and outcomes for children and families. If outcomes fall short of current expectations (e.g., if the employment that recipients move into lacks stability, if many resume the receipt of assistance or reach time limits, if total family income still places families below the poverty line), then it is possible that policy makers will revisit the approaches of the JOBS Program, particularly if the National Evaluation of Welfare-to-Work Strategies provides support for the view that one or both of the JOBS approaches lead to positive outcomes over time. Just as the economic record of the labor force attachment and human capital development approaches implemented under the JOBS Program will remain important in assessing the economic impacts of the new policy, so too will the record of impacts on family functioning and child outcomes remain important in understanding the non-economic impacts of the new law.
Main Hypotheses of This Two-Year Follow-Up Child Outcomes Report
This study considers four broad hypotheses with respect to the impacts of JOBS welfare-to-work programs on children:
The first hypothesis is the null hypothesis, that is, that there are no impacts on the children of welfare recipients mandated to participate in JOBS. Three scenarios might result in null findings. First, it may be that no impacts occur on such variables as maternal education, receipt of welfare and other social and health services, earnings, family income, or on maternal subjective well-being or parenting, or use of child care. In other words, welfare-to-work approaches under JOBS may not affect those targeted and non-targeted outcomes or environments that would be expected to affect children's outcomes; hence, impacts on children are unlikely. A second scenario is that, while impacts on mothers and/or the child's environment occur, they are not of sufficient magnitude to affect children's development. A third scenario is that mothers experience program impacts, perhaps even of sufficient magnitude to affect children, but these impacts are in offsetting directions that "cancel out," leading to no overall impacts on children. For example, mothers may experience significant increases in family income which may on its own bode well for child outcomes, but mothers may at the same time experience increased stress as they balance work and family responsibilities, leading to no net effect on children.
The second main hypothesis is that there are favorable impacts of welfare-to-work approaches under JOBS on children. For this hypothesis to be supported, the processes set in motion by mothers' mandated participation in JOBS -- for example, impacts on economic and/or psychological well-being, parenting, or child care use -- must yield positive effects on at least some of the child outcomes studied. It should be noted that there may be positive impacts on some child outcomes even if there are neutral or negative impacts on some maternal outcomes. For example, mothers may not experience increases in economic resources, but their participation in work preparation activities may enhance their well-being and translate into more positive parenting and child outcomes.
The third main hypothesis is that there are unfavorable impacts of welfare-to-work approaches under JOBS on children of enrollees. Negative child impacts could occur in the context of negative, neutral, or positive impacts on mothers. Children could experience unfavorable impacts, for example, if despite participation in JOBS and a transition to employment, the family experiences no net improvement in economic circumstances, and the child is in alternative care that is unstable or of lower quality.
The final main hypothesis addressed in this report is that impacts will occur only or especially among certain subgroups of children. Children considered at higher risk for poorer developmental outcomes based on selected family characteristics at the outset of the evaluation may experience negative impacts if the mandate serves to be an additional stressor in these families' lives, and/or the mothers are not able to mobilize to participate successfully in mandated activities. Yet favorable impacts for children from higher-risk families might also be anticipated if JOBS programs effectively address the risk factors in these families (such as low educational attainment and limited work experience). In a parallel manner, both favorable and unfavorable impacts for children might be predicted for children from lower-risk families. Children in lower-risk families would stand to benefit if mothers in these families show particularly favorable economic impacts. Yet if mothers in lower-risk families participate in JOBS and secure employment at particularly high rates, but more often find themselves in undesirable work situations than control group mothers (who would not be sanctioned for forgoing an undesirable job opportunity), the lower economic security of such employment and/or the stress of such employment could have unfavorable effects on the home environment and on children.
The study was designed with sufficient power to reasonably assess whether there were harmful impacts on children, an important hypothesis to examine when the new provisions of the Family Support Act were being implemented. The design also has sufficient power to assess whether favorable impacts occurred or whether different impacts occurred for different subgroups.
Outline of Report
Each of the subsequent chapters of this report addresses an underlying question.
Chapter 2: Methods: How Did We Study Impacts on Children? This chapter outlines the methods employed in the Child Outcomes Study. We describe the design of the larger National Evaluation of Welfare-to-Work Strategies and the embedded Child Outcomes Study, as well as describe the procedures of the two-year survey that was the source of information on two-year child outcomes, family processes, and mothers' economic circumstances. We describe the measures we use in the two-year survey to assess the children's developmental outcomes. Finally, we provide an overview of the strategy for data analysis.
Chapter 3: What Were the Sites Like? What Were the Programs Like in Each Site? To be included as a site in the NEWWS, programs had to have previous experience running a strong welfare-to-work program. In addition, Atlanta (Fulton County), Georgia; Grand Rapids (Kent County), Michigan; and Riverside (Riverside County), California were selected as sites for the Child Outcomes Study because: (1) they were each implementing both human capital development and labor force attachment programs, (2) they were diverse in terms of race/ethnicity, and (3) they represented different regions of the country. These sites also differ in terms of how the programs were actually implemented, including the sequence and emphasis of activities, program messages, and the propensity to sanction noncompliant enrollees. Chapter 3 describes these three sites and their JOBS programs in some detail in order to provide a context for interpreting any impact findings for each program approach within each site.(6)
Chapter 4: Sample Description: Who Are the Families in Our Sample? Chapter 4 provides a description of the families in the Child Outcomes Study at baseline. These analyses indicate the heterogeneity of these families; there is range in the sample in terms of characteristics such as educational attainment, barriers to employment, and maternal psychological well-being. Results from this chapter suggest that families in this study are likely to vary in their readiness to participate successfully in JOBS welfare-to-work programs and to secure and maintain subsequent employment -- and child outcomes may differ as a result.
Chapter 5: How Are the Children Faring at Two Years? Descriptive Analyses of Control Group Children's Well-Being and Development. Chapter 5 describes the cognitive functioning, behavioral and emotional adjustment, and physical health status of children in the control group two years after random assignment. Whereas Chapter 6 focuses on program impacts, Chapter 5 provides the context for impact findings, indicating what children's outcomes would have been absent exposure to welfare-to-work approaches under JOBS (since only control group children are examined). Despite the fact that sample families are still generally economically disadvantaged at the two-year follow-up, there is considerable heterogeneity in the developmental status of these children. However, compared to national samples, many children in the Child Outcomes Study sample can be considered at risk on measures of cognitive school readiness, behavioral problems, and physical safety.
Chapter 6: Were There Aggregate Impacts of JOBS Welfare-to-Work Programs on Children's Developmental Outcomes at Two Years? In Chapter 6, we examine the impacts of welfare-to-work approaches under JOBS on sample children's cognitive functioning, behavioral/emotional adjustment, and physical health and safety outcomes. Results are examined separately by site and by program to allow any site-specific and/or program-specific effects to be isolated and identified. Results are "aggregate" in that they are presented for all children within each site and program approach. Impacts of JOBS welfare-to-work programs for specific subgroups of children are considered in Chapter 7.
Chapter 7: How Did JOBS Affect the Developmental Outcomes of Children in Particular Subgroups of Families? Because of the heterogeneity of the families in the NEWWS, it may be that impacts of JOBS programs on enrollees' children occur especially, or only, among certain subgroups of children. Even if there are no aggregate impacts on certain child outcomes, it is possible that impacts occurred in select higher-risk or lower-risk subgroups. Chapter 7 presents impacts on children's cognitive functioning, behavioral and emotional adjustment, and physical health and safety for key subgroups.
Chapter 8: How Are the Families Faring at Two Years? Descriptive Analyses of Key Aspects of Families' Lives. There are a variety of mechanisms through which JOBS programs may have affected children of enrollees. Chapter 8 begins with theoretical rationales for why each proposed intervening mechanism under consideration may be affected by JOBS and why, in turn, each may affect children's outcomes. The chapter then presents descriptive statistics on each intervening mechanism for the control groups at the two-year follow-up to illustrate the range of maternal, family, and contextual characteristics in this sample. The chapter concludes with a summary and discussion of the relative disadvantage of these sample families, aside from any impacts of the JOBS Program.
Chapter 9: Did the JOBS Welfare-to-Work Programs Affect Families in Ways That Could Be Important to Children? Considering the aspects of families' lives described in Chapter 8, this chapter examines which of these were affected by the JOBS programs in each of the three sites. Findings are presented both in the aggregate and for key subgroups of families, as defined in Chapter 7.
Chapter 10: To What Extent Do Program Impacts on Targeted and Non-Targeted Adult and Family Outcomes Help to Explain Selected Impacts of JOBS on Children? This chapter focuses on the extent to which changes in family life help to explain the impacts of JOBS on children. Mediational analyses are reported for a subset of the child impacts, chosen because they illustrated the overall pattern of child impact findings in the different domains of development. These non-experimental analyses will help in the formulation of specific hypotheses that can be examined in the future with further longitudinal data from the five-year follow-up in the Child Outcomes Study.
Chapter 11: What Have We Learned About the Early Impacts of JOBS Welfare-to-Work Programs on Children and Families? Summary of Findings and Discussion of Implications. Finally, in Chapter 11, we summarize key findings regarding the nature of the risks faced by control group children and families in the Child Outcomes Study, and findings pertaining to the early impacts of JOBS welfare-to-work programs on children and families. Implications of these findings for understanding the well-being of children in impoverished families in general are considered. We discuss how these findings and related studies may be informative regarding possible impacts of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 on families and children. We also note the limitations of the present study. Looking to the future, we identify important next steps that can build on the findings of this report.
1.В It should be noted that the welfare-to-work strategies implemented under the JOBS Program do not reflect the changes in welfare policy under the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996. That is, this is not an evaluation of welfare-to-work approaches adopted as part of the new Temporary Assistance for Needy Families program (TANF, which replaced AFDC under PRWORA). As we will note in further detail in this and other chapters, however, an examination of impacts on children of a set of programs operated under JOBS can help us anticipate whether and how programs implemented under PRWORA may be affecting children.
2.В Throughout this report, we use the term child outcome to refer to the measures of child well-being and development. We reserve the term child impact to describe statistically significant differences on a child outcome measure between children in program group families and children in control group families.
3.В Program mothers may not have remained "JOBS mandatory" during the entire two-year follow-up period if, for example, they secured employment or left welfare. Indeed, mothers assigned to Riverside's JOBS programs were JOBS mandatory for less than half of the follow-up period. Nonetheless, mothers assigned to Riverside's LFA program still participated in only one-quarter of the months in which they were required to participate during the follow-up period, whereas mothers assigned to Atlanta's and Grand Rapids' HCD programs participated in over half of the months in which they were required to participate.
4.В Control group families were subject to the same guidelines about child care within each site as experimental group families. In Atlanta, subsidies could only be used for licensed child care (and this held for both experimental and control group families). In Riverside, both experimental and control group families were encouraged to use lower cost informal child care arrangements. In Grand Rapids, type of child care was left to the family (again in both experimental and control groups). However, we note that the enhanced case management received by experimental group mothers in each site might have provided a context for further discussions about child care issues.
5.В Data collection for the Child Outcomes Study was complete by January 1996. We note that the period of data collection for the Child Outcomes Study antedates and does not overlap with the period of implementation of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996.
6.В Much of this chapter draws from the report of the larger NEWWS titled, Evaluating Two Welfare-to-Work Approaches: Two-Year Findings on the Labor Force Attachment and Human Capital Development Programs in Three Sites (Hamilton, Brock, Farrell, Friedlander, and Harknett, 1997). For greater detail on how each treatment stream was implemented in each of the sites, see this report.