Outcomes for Control Group Families and Program Impacts on Family Outcomes
What was the family context for these child impact findings? We turn now to an examination of outcomes for families. As we did for the child outcomes, we first consider how families were faring in the control groups, in the absence of JOBS programs, and then turn to an examination of program impacts.
A. Control Group Families
How were families faring economically in the absence of a JOBS program?
Analyses focusing on control group families indicate that, in the absence of welfare-to-work strategies under JOBS, sample families would have made some educational and economic progress over the two-year follow-up period, though many would still be considered disadvantaged.
Following random assignment, between 8 and 19 percent of control group mothers in the three sites received some sort of educational credential. Economically, whereas all sample families had applied for or were receiving AFDC at baseline, at the time of the two-year follow-up study, only 53 to 79 percent (with the percent varying by site) of control group mothers were receiving welfare. More mothers in the control groups (between 29 and 47 percent) were employed in the month preceding the survey than two years earlier. Employed mothers in the control groups had average earnings of approximately $6.10 an hour, but many (between 11 and 25 percent) were working at or below minimum wage.
Considering all sources of income(13), an estimated 69 (in Grand Rapids) to 77 percent (in Atlanta) of households in the control groups were living below the poverty line(14) at the two-year follow-up. Not surprisingly (given the population studied in the evaluation), these rates exceed the 17.4 percent of families with related children in the U.S. as a whole who were living in poverty in 1994. In addition, 11 percent in Riverside, 19 percent in Atlanta, and 21 percent in Grand Rapids, reported total net adjusted household incomes reflective of "deep poverty" (i.e., less than 50 percent of the poverty line). This compares to 7.2 percent of families in deep poverty nationwide (see Figure SR-5).
In the absence of a JOBS program, how were families faring in terms of non-economic aspects of family life?
There were also indications of a range of risk on the non-economic aspects of family life among control group mothers.
Depressive symptomatology. More than a quarter of the mothers in each site (from 29 percent in Riverside to 37 percent in Atlanta and Grand Rapids) reported "many" depressive symptoms, suggestive of clinical levels of depression. The rates in Atlanta and Grand Rapids are comparable to those reported by mothers in Atlanta in the months immediately following random assignment, when 42 percent of these mothers reported high levels of depressive symptoms (Moore et al., 1995). However, rates of "many" depressive symptoms are lower in the Child Outcomes Study sample compared to the more disadvantaged mothers enrolled in the New Chance Demonstration (all of whom were teenage mothers who had dropped out of school at the start of that evaluation); 52 percent of these mothers had high levels of depressive symptoms (Quint et al., 1997). Nonetheless, rates of depressive symptomatology suggestive of clinical depression among control group mothers in the Child Outcomes Study are higher than those typically found in community samples, which are closer to 20 percent for women (see review in Devins and Orme, 1985) (Figure SR-6).
Parenting. Interviewers rated a nontrivial minority of control group mothers (between 8 and 19 percent) as behaving in a harsh manner toward the focal child during the interview. Interviewers also rated control group mothers as moderately warm, on average, in interacting with their children, but there was a wide range in ratings of warmth, and ratings varied by site. On average, control group mothers did not rate themselves as particularly aggravated in the parenting role. However, about 10 percent of mothers in each site had scores indicating substantial aggravation and stress in the parenting role.
Father involvement as reported by mothers. We first examined the proportion of focal children living with their biological fathers. For those not living with their fathers, we note the proportion who had seen their fathers in the past 12 months. We also note the proportion of nonresidential fathers who had purchased clothing, toys or presents; had provided groceries; and who had served as a babysitter for the child. Few control group mothers (between 4 and 12 percent) were living with the focal child's biological father at the time of the follow-up survey, though over half of focal children who were not living with their biological fathers had seen them in the previous 12 months. In terms of other forms of support (buying clothes, toys, presents; providing groceries; occasional babysitting), focal children's biological fathers provided, on average, less than one of the three kinds of support.
Marital status and fertility. Whereas only 2 percent or fewer of control group mothers had been married at baseline, between 4 and 15 percent of control group mothers were married at the two-year follow-up. In addition, between 12 percent and 19 percent reported having had a baby between baseline and the two-year follow-up.
B. Program Impacts on Families
Program impact findings for families in the full sample of the National Evaluation of Welfare-to-Work Strategies have been presented elsewhere (Hamilton et al., 1997; Freedman et al., 2000). Here we summarize findings specifically for the Child Outcomes Study sample. These findings help ascertain whether the JOBS programs brought about changes in economic and non-economic aspects of family life specifically in families that entered the evaluation with preschool-age children. It is important to determine if the presence of such young children impeded program participation or prevented families from making economic progress. It is also important to ascertain whether being assigned to a JOBS program when at least one child in the family was of preschool age resulted in changes in parenting behavior or the home environment, or in maternal psychological well-being.
Did assignment to a JOBS program result in increased participation in program-related activities for mothers with young children?
Regardless of program approach, the six JOBS programs studied resulted in greater participation in job search activities. The largest impacts on participation in job search activities were concentrated in labor force attachment programs. Mothers in each site's labor force attachment program were about ten times as likely as those in the control groups to have ever participated (for at least one day) in job search activities since random assignment. Large impacts also occurred for mothers assigned to Riverside's human capital development program: 29 percent of these program mothers had participated in job search activities, compared to less than 5 percent of mothers in Riverside's (human capital development) control group.
Participation in basic education was higher for mothers assigned to a human capital development program than for control group mothers. The largest impact on participation in basic education activities (regardless of whether a degree was ultimately obtained) occurred for mothers assigned to Riverside's human capital development program: over half of these program mothers had participated in basic education activities since random assignment, whereas only 16 percent of control group mothers had done so.
Corresponding to the participation mandate faced by program, but not control group, mothers, all six programs in the Child Outcomes Study resulted in sanctioning of program families who failed to comply with the participation mandate. Sanctioning rates were highest for mothers in the two Grand Rapids programs (27 percent for those in the human capital development program and 38 percent in the labor force attachment program). Mothers in Atlanta's human capital development program were the least likely of the three human capital development programs to be sanctioned (15 percent as opposed to 27 percent in Grand Rapids and 24 percent in Riverside). Similarly, sanctioning rates were lower for mothers in Atlanta's labor force attachment than in the other two labor force attachment programs (11 percent in Atlanta; 38 percent in Grand Rapids, and 14 percent in Riverside).
Did assignment to a JOBS program result in impacts on targeted outcomes and/or on outcomes derivative of targeted outcomes for families with young children?
Educational attainment. Mothers in the human capital development programs, but generally not those in the labor force attachment programs, progressed in terms of educational attainment. This is in keeping with program goals: labor force attachment programs generally did not aim to increase educational attainment, while human capital development programs did.
Overall, mothers in human capital development programs in each site were more likely than controls to have received any educational credential since random assignment. By contrast, each site's labor force attachment program tended not to improve mothers' educational attainment (except in Atlanta, where mothers in the labor force attachment program were more likely than controls to have obtained a trade degree). Indeed, mothers assigned to Grand Rapids' labor force attachment program were less likely than controls to have obtained a high school diploma or GED, trade certificate, or any degree since random assignment.
Economic circumstances. Two years after random assignment, improvements in families' economic circumstances occurred mainly for mothers assigned to labor force attachment programs and for mothers assigned to Riverside's human capital development program. Impacts were largest and most numerous for the two programs in Riverside.
Program mothers in Atlanta witnessed relatively few economic impacts. From among multiple economic outcome measures, impacts were found on only a few in this site. As intended, both programs led to a reduction in AFDC receipt at the two-year follow-up. In addition, those in Atlanta's human capital development program were less likely to report earning the minimum wage or less and had somewhat greater earnings from their current jobs, and those in Atlanta's labor force attachment program were more likely to be employed, and less likely to be in deep poverty in the month prior to the two-year follow-up. Neither of Atlanta's JOBS programs had a statistically significant impact on total net adjusted household income at the two-year point.
The only statistically significant unfavorable economic impacts of any of the six JOBS programs in this study occurred for mothers in Grand Rapids' JOBS programs. Both JOBS programs in Grand Rapids' reduced the proportion of families at or above the poverty line. Mothers assigned to Grand Rapids' human capital development program were less likely to be employed 40 or more hours per week, and results for this program show an increase in the proportion of mothers working for less than the minimum wage. The only favorable economic impact of Grand Rapids's labor force attachment program was on the likelihood that program mothers were employed at some point since random assignment. It is worth noting that those in Grand Rapids' labor force attachment program had greater employment rates in the month prior to the two-year follow-up than those in any of the other sites' JOBS programs, but because control group mothers in Grand Rapids also had relatively high rates of employment (compared to control group mothers in the other sites), there was no net impact of Grand Rapids' labor force attachment program on current employment. Neither of Grand Rapids' JOBS programs had a statistically significant impact on total net adjusted household income at the two-year point.
Economic impacts in Riverside were more widespread. Mothers assigned to either of Riverside's JOBS programs were more likely than controls to have been employed at some point since random assignment, were more likely to have been employed in the month prior to the two-year follow-up, had more hours of employment, higher wages, and greater earnings. (Riverside's JOBS programs were the only ones in this study to have favorable impacts on hourly wage and hours employed.) In addition, mothers enrolled in Riverside's labor force attachment program were less likely to be receiving AFDC, and mothers in Riverside's human capital development program were more likely than their control group counterparts to have incomes above the poverty line. These impacts also occurred for many lower-risk and higher-risk subgroups. In addition, despite the absence of impacts on total income at the aggregate level, Riverside's human capital development program resulted in higher total household income for one lower-risk and two higher-risk subgroups of mothers. However, for another lower-risk subgroup, this program resulted in lower earnings. Also, despite the absence of impacts on deep poverty at the aggregate level, Riverside's labor force attachment program increased deep poverty for some higher-risk families. Nevertheless, employment and economic impacts of Riverside's JOBS programs were generally favorable, which is consistent with this site's focus and experience implementing work-oriented programs.
Employment-related child care. Program mothers (except those in Grand Rapids' human capital development program) were more likely than controls to be using child care for any of their children while they were employed at their current or most recent job. The largest impact on the use of employment-related child care for any of their children occurred for Riverside's labor force attachment program, where program mothers experienced the largest employment impacts (and where control group mothers reported relatively infrequent employment-related child care). Relatively large impacts also occurred in Riverside's human capital development program and in Grand Rapids' labor force attachment program, where program mothers were more likely than controls to have been employed at some point since random assignment. The small impacts in Atlanta likely reflect, in part, the absence of large employment impacts in this site.
Impacts on the current use of employment-related child care for the focal child were more limited and occurred only in Riverside: Only Riverside's JOBS programs increased the overall likelihood that mothers were both employed and using any non-maternal care for the focal child in the month prior to the two-year survey. Regardless of program approach, program mothers in this site reported more employment-related hours of child care for the focal child in the prior month, were more likely than controls to be employed and using informal care for the focal child, and to be employed and using care for the focal childduring irregular hours or with a varying schedule. These findings are consistent with the focus in Riverside likely experienced more intensively by program than control group mothers on helping mothers arrange low cost, primarily informal, child care, which provides greater flexibility as to hours of employment.
Did assignment to a JOBS program affect further, non-targeted outcomes for families?
There was less evidence of impacts on aspects of family life not targeted by JOBS welfare-to-work programs, yet such impacts did occur. When these impacts occurred, however, there were indications (discussed below) that they could be quite important to child outcomes.
Maternal psychological well-being. Specific aspects of maternal psychological well-being were affected in different ways by JOBS welfare-to-work programs, and impacts did emerge, especially for subgroups of mothers.
Assignment to a JOBS welfare-to-work program often led to increased feelings of time stress. Perhaps more worrisome, enrollment in a JOBS program increased depressive symptomatology, in the aggregate, for mothers assigned to Grand Rapids' labor force attachment program, as well as for a few lower-risk subgroups exposed to Grand Rapids' or Riverside's human capital development program, and a lower-risk subgroup assigned to Atlanta's labor force attachment program. Interestingly, Atlanta's labor force attachment programincreased the number of depressive symptoms among mothers reporting relatively many psychological problems at baseline, and decreased the number of depressive symptoms among mothers reporting no or relatively few psychological problems at baseline. This pair of offsetting impacts indicate that this JOBS program had completely opposite impacts (which resulted in an absence of impacts in the aggregate) depending on the initial level of mothers' depressive symptoms.
At the same time, findings indicate that, despite an absence of impacts at the aggregate level, four of the six JOBS programs led to greater feelings of control over one's life for certain (especially lower-risk) subgroups of mothers. This was especially true for Riverside's labor force attachment program, in which a majority of lower-risk subgroups examined had increases on the measure of subjective sense of control.
Parenting. In the aggregate, Atlanta's human capital development program led to greater warmth (as reported by the mother) and to higher scores on a summary measure of "favorable" parenting. Additional impacts on parenting emerged for a few subgroups, with Atlanta's human capital development program having both favorable and unfavorable impacts in both lower-risk and higher-risk subgroups. In the aggregate, Atlanta's labor force attachment program led to less harsh discipline, greater maternal warmth (as reported by the interviewer), greater verbal interactions with the focal child, and higher scores on a summary measure of "favorable" parenting. The additional impacts that emerged in a few subgroups were uniformly favorable in lower-risk subgroups and were both favorable and unfavorable in higher-risk subgroups.
There was only one aggregate impact on parenting in Grand Rapids: Mothers assigned to Grand Rapids' labor force attachment program reported less warmth toward the focal child than did control group mothers. Additional parenting impacts emerged in subgroups, and most were unfavorable regardless of risk level.
In the aggregate, there were no impacts of either of Riverside's JOBS programs on any measure of parenting. Nevertheless, a few impacts emerged at the subgroup level, with both favorable and unfavorable impacts of each program in lower-risk subgroups, but only favorable impacts of each program in higher-risk subgroups.
Father involvement as reported by mothers. Mothers in each of Grand Rapids JOBS programs reported receiving more forms of support from the focal child's biological father, such as purchasing groceries, toys, clothes, and babysitting, compared to control group mothers. This difference, while statistically significant, was small. Further, such support was reported to be occurring at low levels: program mothers in Grand Rapids still reported less than one kind of support, on average. Riverside's labor force attachment program decreased the likelihood that the focal child's biological father lived with the focal child at the two-year follow-up. Additional impacts emerged when subgroups were considered; notably, each of Atlanta's JOBS programs decreased the likelihood that the focal child's father lived with him or her in four higher-risk subgroups. In Grand Rapids, all impacts on father involvement in subgroups were favorable. By contrast, in Riverside, all impacts on father involvement in subgroups were unfavorable.
Marriage and fertility. In the aggregate, none of the six welfare-to-work programs had an impact on mothers' fertility, or their marital and cohabitation status, by the two-year follow-up, though a few scattered impacts emerged for five of the six JOBS programs in both lower- and higher-risk subgroups.