Table of Contents | Previous | Next |
III. THE NEXT WAVE OF DEMONSTRATIONS
A. CRITERIA FOR JUDGING NEW DEMONSTRATION APPROACHES
Demonstration projects and social experiments are important mechanisms for learning about the relative effectiveness of alternative social strategies and innovative program services, and about expanding their scope. At the same time, demonstrations often are costly and sometimes take many years to yield reliable results. Given these realities, it is critical to test the most promising interventions in ways that can be rigorously evaluated. In developing ideas for the next wave of demonstrations on marriage, employment, and family functioning, we will draw on the conceptual framework, experience with prior demonstrations, and observational research on the interactions between marriage, employment, and family functioning.
Proposed demonstrations should meet several criteria. They should (1) have a solid theoretical rationale linked to a conceptual framework, (2) build on successful elements of past demonstrations and not replicate approaches that have proved ineffective in other demonstrations, (3) have attributes that are consistent with policy implications of the empirical literature, (4) have the potential of significantly improving outcomes along more than one domain—such as marriage and employment or marriage and family functioning, (5) be subject to rigorous evaluation, and (6) have the potential to be replicated and implemented on a large scale, if successful. We favor focusing resources on groups for whom gains in marriage, employment, and family functioning are most urgently needed so long as interventions have a reasonable chance of positively influencing the groups’ outcomes. We see advantages in using current operational venues—such as local programs, providers, community activities, and existing demonstrations—to minimize start-up time, to limit costs, and to reach significant numbers of people. However, we recognize the possibility that new, stand-alone demonstrations may sometimes be appropriate.
Demonstrations can meet these criteria using a vast number and mixture of interventions. Even among strategies focusing on employment and earnings, there are wide differences in approach, intensity, duration, scale, delivery mechanisms, involvement of partners, and target groups. Program goals often differ as well, for example, over such issues as the relative emphasis placed on current employment versus long-term careers, or the emphasis on gender equity versus earnings gains. The employment area offers examples that try to influence outcomes through various aspects of the framework described above. Interventions can be found that emphasize incentives (wage subsidies), skills (classroom and on-the-job training, job-search skills), information (job openings and the career outlooks for various occupations), attitudes (work experience and work-readiness programs), or context (programs that try to alter the participant’s peer group or change the hiring and training practices of a local industry sector). Often, programs use a combination of these strategies.
B. SOME RELEVANT EXPERIENCE FROM SELECTED DEMONSTRATIONS
1. Employment-Related Demonstrations
The evidence from social experiments and program experience is mixed. Wage subsidies exert positive impacts on employment, hours worked, and total income, especially for single parents and where the subsidy increases the returns to workers above their wage rate (EITC, Minnesota Family Investment Program, Self-Sufficiency Project, and New Hope). It is hazardous to generalize about classroom training, since the intensity and duration of programs vary from a few months to a few years. Evaluations of randomized experiments typically show modest gains from programs targeted at low-income workers. One comprehensive evaluation—based on operational programs under the Job Training Partnership Act (JTPA)—found that earnings increased about 15 percent for adult women and 8 percent among adult men, enough to justify the program’s costs (Orr et al. 1996). Especially effective were the gains to those expected to use on-the-job and job-search training. Unfortunately, the programs did nothing to raise the earnings of young men and women. Moreover, more comprehensive reviews of government-funded training programs for the disadvantaged show limited gains. A meta-analysis of 31 evaluations of these programs found that annual earnings gains were about $1,400 (1999 dollars) for adult women, $300 for adult men, and zero or negative for youth (Greenberg, Michalopoulos, and Robins 2003).
Although gains from training programs are uneven, nonexperimental evidence shows substantial increases in earnings associated with years of general and vocational education. In addition, intensive job-search programs, especially those that teach people how to find their own jobs, have shown positive impacts. Like the JTPA effects on adults, these gains from job-search programs are small but sufficient enough to offset their modest costs. Subsidized jobs and work-experience programs, often involving jobs that gradually increase in difficulty and stress, raise earnings, especially during the period when these jobs are available. The gains beyond the subsidized job period have varied, depending on the target group and combination of activities.
Many training programs for low-income individuals begin with life-skills training aimed at changing attitudes about the importance of work and about the habits necessary to succeed in the workplace. Although we should attach some weight to the consensus of practitioners about the importance of these aspects of pre-employment training, we know of no studies that have documented the impact of this program component. One initiative that attempts to alter the context of at-risk youth is Job Corps. Individuals receive housing, education, training, health care, and other services, mostly at residential Job Corps centers (Johnson et al. 1999). Although the targeting of the program puts at-risk youth into an environment populated mainly by other economically disadvantaged youth, the centers attempt to positively alter the context within which participants learn, work, and interact. For the first few years after entering the program, the impacts on earnings were positive (Schochet, Burghardt, and Glazerman. 2001). Job Corps exerted favorable impacts on obtaining a GED and occupational certification, and on curbing criminal activity. But, the combination of earnings and other gains were insufficient in offsetting the social costs of the program (Schochet, McConnell, and Burghardt 2003). Moreover, while there is some indication that the Job Corps context matters, there were no statistically significant differences in impacts between those in a residential center (away from neighborhood peers) and those in a nonresidential setting.11
Using an industry context for employment interventions is the emphasis of sectoral strategies. Under this approach, workforce programs target an industry (or subset of an industry), become a strategic partner of the industry by learning about the factors shaping the industry’s workforce policies, reach out to low-income job seekers, and work with other labor-market groups, such as community colleges, community nonprofits, employer groups, and policymakers. The Aspen Institute and the Urban Institute have conducted studies of the operations of sectoral projects along with some analysis of data on the earnings of workers before and two years after participating (Blair 2002; Pindus et al. 2004). The goal is to link the training and career strategies for low-income job seekers to the industry’s needs. By design, the programs deal with a particular industry and thus generalizations are hazardous. But, the nonexperimental evidence indicates that the six sectoral programs taking part in the Sectoral Employment Development Learning Project (SEDLP) have yielded impressive results (Blair 2002). Earnings jumped by 73 percent in two years for the 95 percent of participants employed two years. These results must be interpreted with caution because the project lacked a control group. However, the increases between the year before and the two years after participation were much higher in the SEDLP than in demonstrations with comparable groups of workers. Although most of the gains came from higher work levels, wage rates increased by 23 percent. Moreover, two years after training, 69 percent of participants were employed in occupations related to their training. The focused nature of the training, the links with employers, the development of pathways for entry-level workers, and the expertise gained by the training organizations probably all have contributed to the apparent success of the sectoral strategy approach.
A traditional sector-based approach with a long-term track record of success in raising earnings through targeted training is the apprenticeship system. Apprenticeships involve intensive work-based learning and classroom courses. Employers are central to the process, setting up the programs and paying the apprentices during their work-based learning. Although formal, registered apprenticeships are most common in the construction and manufacturing industries, the role of apprenticeship is expanding in other occupations and industries, including metalworking, nursing, information technology, and geospatial occupations.12 One recent nonexperimental study found that apprenticeship training generated substantial gains in earnings, especially for those completing the program (Washington State Workforce Training and Education Coordinating Board 2004). Relative to a statistically matched comparison group that registered for services with the state employment service, those participating in apprenticeships raised their employment rate by 5 percent.13 Employed workers who were in apprenticeships earned nearly $2,000 more a quarter than this primary comparison group. These earnings gains are nearly double the comparably estimated gains for participating in a vocational degree program from community colleges.
Another program linked to specific industry sectors is the Career Academy. While operating within schools and as part of a local school system, career academies are high schools organized around an occupational or industry focus, such as health care, finance, and tourism. They try to weave related occupational or industrial themes into a college preparatory curriculum. An experimental evaluation using random assignment has documented some striking gains (Kemple 2004). Although career academy participation did not increase the earnings of women, young men assigned to career academies achieved an extraordinary 18 percent average gain in earnings compared with the control group over the four years after scheduled high school graduation. The career academy group earned an average of $1,373 a month, $212 more than the $1,161 a month earned by the control group. The earnings gains were concentrated among students with a high or medium risk of dropping out of high school.
One small but innovative employment demonstration tested the impact of providing job-search and job-readiness services to help both members of 17–24-year-old couples obtain a job or a better job (Gordon and Heinrich 2007). The Jobs for Youth (JFY)/Chicago’s Full Family Partnership (FFP) project operated in the Chicago area, mostly with low-income African American couples, beginning in 1998. The couples had to be in stable relationships, in which at least one partner was a parent and on TANF. The program enhanced the standard set of JFY services (job-readiness workshops of 10–15 days, GED instruction, and job-search assistance) to include one-on-one counseling. The idea is that the partners can support each other, recognize the challenges faced by their partners, offer specific supports like driving or watching a child, and provide appreciation, encouragement, and monitoring of skills. Although enhanced services were received by both parents in the couple only 60 percent of the time, mothers who participated in the FFP achieved higher employment and earnings gains at exit than mothers in the standard JFY approach or in the JTPA program. However, their earnings advantage eroded over time, partly due to more new pregnancies and higher child care burdens. Fathers showed less robust but still positive gains from participating in FFP instead of only JFY. The researchers also found that the couples approach was linked to completion rates 20 to 30 percent higher than among parents with similar characteristics who participated in the JFY program. But much of the advantage in completing the program was the result of higher levels of services provided through the FFP. When both parents completed the FFP, their earnings jumped by over $4,000. Some evidence indicates a feedback between use of program completion, increases in earnings, and marital stability. Among parents who both completed the program, nearly 90 percent remained together at least a year later. Gains in earnings were associated with relationship stability, but the causation may run from earnings to couple stability or the other way around.
2. Demonstrations Linked to Marriage and Family Functioning
Several demonstrations have tested or are testing ways of improving the health of marriages and the broader functioning of families. In the marriage area, programs providing premarital education, premarital counseling, and marriage preparation for couples have been subject to extensive research. Carroll and Doherty (2003) conducted a meta-analysis of 13 experimental studies and concluded that, “premarital prevention programs are generally effective in producing immediate gains in communication processes, conflict management skills, and overall relationship quality and that these gains appear to hold for at least 6 months to three years.” One of the studies reviewed (Markman et al. 1993) indicated that premarital prevention could reduce the likelihood of divorce. While these studies generally had small samples and short follow-ups, and rarely included low-income populations, the results from the broader literature on healthy marriage interventions have been sufficiently compelling to influence policy in the United States and other countries.14
Under the Healthy Marriage Initiative (HMI), the ACF is currently sponsoring demonstration projects with large samples and long follow-ups targeted at low-income populations.15 Two involve random assignment (Building Strong Families and Supporting Healthy Marriage) and one is focusing on community impacts (Community Healthy Marriage Initiative). The HMI projects cover a wide scope: improving financial incentives to marry, marriage education classes, mentoring programs involving married couples, training clergy and others to deliver marriage education, and courses in high school about healthy dating practices and information about the advantages of marriage. The intensity and duration of the interventions vary, with some lasting only a total of 10–12 hours of instruction. However, others are expanding their scope and beginning to connect with employment-oriented services as well as financial-literacy and matched-savings programs (such as the Individual Development Account programs).
One relevant set of results comes from an experimental project aimed at increasing fathers’ involvement and couple relationships. The Supporting Father Involvement program provided information, 16-week classes, and case management to 289 low-income families in four rural California counties (Cowan and Cowan 2007). The families were randomly assigned to a fathers group (where the classes and counseling tilted toward parenting), a couples group (where the classes and counseling tilted toward couple relationships), and a control group. The assessments of the treatment and control groups at 9 and 18 months after the baseline assignment revealed a number of significant positive impacts. Both the fathers and couples interventions increased father involvement and decreased parenting stress, anxiety, and conflict over the child. In addition, relative to the control group’s income trend, the impact on household income for those assigned to the couples program was a $3,770 gain per year (over an initial average of about $28,000); those assigned to the fathers group experienced a $2,505 additional gain over controls. The program developers view the central finding as showing that fathers are more likely to become involved with their children when they have a good relationship with the child’s mother. From our perspective, the program is an example of a family-functioning intervention that not only improved couple and parent-child relationships but also raised earnings.
Other initiatives have aimed directly at improving parenting to strengthen families. The interventions include direct training and mentoring through regular advice, and observations through nurses’ visits and classes at Head Start and other child care centers. In addition, some policies encourage parents to spend more time with their children, such as the mandate that employers offer family and medical leave and some programs to provide paid leave. Only some of these initiatives have blossomed into full-scale programs. Programs not primarily focusing on family functioning may nonetheless exert impacts for good or ill. New Hope offered low-income people in a set of zip codes in Milwaukee a package that included earnings subsidies to supplement the EITC, child allowances, a community service job at the minimum wage, subsidized health insurance, and subsidized child care. Eligibility for the benefit package extended to all types of households, including single individuals, childless couples, and families with children. Each participant had a project representative who helped them access benefits, served as an informal counselor, and encouraged participants. Wisconsin spent an additional $3,300 per year for each New Hope household (in 2006 dollars), mostly on added child care. The job and income stability provided through New Hope apparently generated positive effects on boys (Huston et al. 2003), while the strict work requirements imposed on welfare recipients might have harmed adolescents, though not younger children (Zaslow, McGroder, and Moore. 2006).
The Parents’ Fair Share (PFS) demonstration focused on one aspect of family functioning—increasing the financial and nonfinancial support of children by noncustodial fathers (Miller and Knox 2001). Given the likely links between work, support payments, and fathers’ involvement, the program provided a range of services, including employment and training, peer support, voluntary mediation between parents, and help in modifying child-support orders. PFS generated substantial initial gains in employment and earnings for the most disadvantaged fathers, probably as a result of on-the-job training and earnings during this component, but no significant gains for the full sample of fathers. PFS stimulated increases in child-support payments, but little additional father involvement. For another family-functioning outcome—communications between noncustodial fathers and custodial mothers—the level of disagreement increased, though this change may have resulted from more active interest by fathers who had not been closely connected with their children and the children’s mother.
The Nurse Home Visiting project has attracted wide attention for its ability to achieve significant gains in child and family functioning. The goals of these programs are to improve pregnancy outcomes, children’s health and development, and parents’ well-being. In a series of random-assignment demonstrations, Olds and his colleagues (1988) found several positive impacts of intensive nurse home-visiting services during pregnancy and through the child’s second birthday.16 The Elmira, New York, demonstration raised schooling and employment and delayed the second child. In Denver, the nursing intervention program component was linked to a delay in second births and a reduction in domestic violence, but to no other favorable effects (Olds et al 2004). The Memphis demonstrations enrolled young pregnant women who had no chronic illness linked to fetal growth, but nearly all were unmarried, poor, and teenagers (Olds et al. 2007). The evaluation of children and their mothers around the child’s 9th birthday documented several statistically significant impacts, including fewer second births, less use of welfare programs, a higher likelihood of marriage or cohabitation or other partnering with the child’s father, more months with an employed partner, and better academic outcomes for children. The projected led to an increase in the months spent with the mother’s current partner, although the program did not explicitly attempt to increase marriage and relationship skills.
Youth development is a part of the functioning of families. A number of projects have been undertaken to promote youth development in a variety of settings (Eccles and Gootman 2002). One interesting program with evidence-based effectiveness is the Carrera Program, an intensive, year-round, multiyear after-school program designed to promote positive youth development and positive reproductive health. The program employs a holistic approach involving school, family, supportive relationships, and social services and provides employment and academic assistance, family life and sexuality education, performing arts experience, sports training, and mental and physical health care. According to an experimental evaluation, participation in Carrera increased sexual health knowledge, receipt of health care and health behaviors, life skills, academic skills, and work experience (Philliber, Kaye, and Herrling 2001; Philliber et al. 2002). Participation also lowered levels of pregnancies and births and the likelihood of marijuana use in males.
3. Benefit-Related Policies to Raise Employment and Marriage
A major challenge of cash and in-kind benefit programs is to help low-income families without discouraging work and marriage. The EITC, subsidized jobs, and work requirements have achieved gains in employment, but structuring benefit programs that strengthen marriages is difficult. Until the income maintenance experiments of the 1960s and 1970s, the conventional view was that simply allowing low-income married-couple families to qualify for benefits on the same basis as one-parent families would eliminate marital disincentives. Helping couples with children achieve income stability was thought to reduce divorce and to increase marriage. However, evidence from the Seattle-Denver Income Maintenance Experiment showed that extending cash benefits to two-parent families did not increase, and may have even decreased, marriage (Cain and Wissoker 1990; Hannan and Tuma 1990). Helping single parents attain basic incomes was said to have increased their economic independence—they did not have to rely on a spouse or cohabiting partner for economic support.
Still, two recent demonstrations—New Hope and the Minnesota Family Investment Program (MFIP)—have shown that providing targeted benefits can sometimes increase marriage. The MFIP tested a welfare-to-work model with mandatory participation in work and training, consolidated and cashed out related benefit programs, enhanced child care subsidies, and reduced marriage penalties in the provision of transfer benefits.
New Hope achieved progress on several key goals, including gains in earnings of about $700 per year during the eligibility period, increases in family income by over $1,000 per year, declines in poverty rates by about 30 percent (from about 70 to about 50 percent), reduced reports of symptoms of depression, improvements in several dimensions of family functioning, and better outcomes for children (Huston et al. 2003). Most strikingly, marriage rates increased as well (Gassman-Pines and Yoshikawa 2006). At the five-year follow-up, marriage rates of never-married mothers in the New Hope treatment group were almost double those of never-married mothers in the control group (21 percent to 12 percent). The study does not identify the mechanism by which marriage rates increased, but one possible explanation is that New Hope offered a degree of income security not available to control-group members. One possibility is that the availability of assured jobs and earnings subsidies can increase marriage rates. During the first three years after entering the MFIP, single parents raised their earnings and showed modest increases in the likelihood of marriage. Over the subsequent three years, both dissipated; no significant effects were evident in earnings or marriage at the six -ear follow-up. A key element of the MFIP was the coverage of low-income two-parent families. Although the MFIP did not end up raising the incomes of these families (higher benefits were offset by lower earnings of women in two-parent families), it did lower the rate at which married couples divorced (Gennetian, Miller, and Smith 2005).
4. A Mix of Strategies
The experience of demonstrations and programs suggests a role for a mix of strategies, including changing incentives, skills, information, attitudes, and context. The programs that stand out in the employment arena assure the availability of a job and combine work incentives and work experience with learning, in the context of a well-articulated career structure. In addition, some research and demonstrations suggest genuine complementarities; specifically, enhancing marital stability increases men’s employment, earnings, and family incomes, while improving the health of marriages yields measurable gains in child outcomes.
C. SUGGESTED AREAS FOR MAJOR DEMONSTRATION PROJECTS
Even with the experience of past demonstrations and programs and the knowledge of existing demonstrations, it is a complex task to devise a sensible mix of new demonstrations aimed at increasing healthy marriages, employment and earnings, and well-functioning families. There are three types of outcomes of primary interest, five components of a causal framework, various target groups, and a multitude of program instruments and combinations of instruments. The primary question is, which combination of approaches is most likely to achieve more of our primary objectives?
This section presents several suggestions for types of demonstrations that widen the mix of service approaches, venues, target groups, motivations, and expected outcomes. The recommendations take account of common difficulties in recruitment of participants and in administering interventions. Underlying the proposals is the notion that interventions should become more holistic and deal with a broader mix of challenges faced by individuals, couples, and parents. Thus, the demonstration ideas entail incorporating effective employment services into marriage-oriented programs and for incorporating marriage education, relationship skills, and family-functioning interventions into employment programs. In this section, we offer five concrete proposals, providing a brief case for each and an outline of the way the demonstration could be evaluated—usually with an experimental design. These proposals are by no means an exhaustive list of possible program ideas. Rather, they illustrate how our conceptual framework may be applied to well-founded program models and how these models may be revised to address healthy marriage, employment, and family well-being outcomes.
1. Adding Effective Employment Services to Marriage-Oriented Programs
Under this type of demonstration, sites offering marriage education and relationship-skills programs would expand the scope of the initiative to include employment-oriented services. The specifics of the services are important and several approaches are promising. We recommend two employment-oriented strategies. The first involves offering both members of a couple combinations of wage subsidies, counseling, and community service jobs. The concept builds partly on the experience of New Hope and the Full Family Partnership programs. We suggest using wage-rate subsidies partly to avoid imposing marriage penalties, incorporating community-service job models used in other programs, and including both partners in the provision of job-related assistance.17 If the wage subsidy were available to both partners in a couple and only phased out with wage rates, then a low-wage working man or woman would not be penalized by marrying another worker.18
The second employment component would involve offering participants in the marriage programs access either to local sectoral industry programs or to apprenticeship programs. Nonexperimental evaluations indicate that both have a good record of improving the earnings of at-risk participants. In both cases, the interventions would integrate the topics in marriage education with the noncognitive employment-related skills, such as the those specified in SCANS, the Secretary’s Commission on Achieving Necessary Skills (1992). These approaches respond to the concern that low-income women report wanting marriage to coincide with financial stability and a good living standard. Many low-income mothers want to marry a man who has a decent job but also want a job of their own to avoid excessive dependence on a man who may not be reliable (Edin and Kefalas 2005).
The demonstration would target individuals who sign up for marriage education classes and who are cohabiting, in a close romantic relationship, or married. The presence of children would not be an eligibility requirement for participation. If couples do have children, then parental job sharing, parenting education, and backup emergency child care provision could be made available. Offering the employment- and income-related services in the context of ongoing programs should ease recruitment. In fact, a number of the sites that are currently funded by the Office of Family Assistance (OFA) through Healthy Marriage Demonstration Grants would be ideal candidates for piloting this type of intervention.19 They are currently offering marriage education in a variety of settings, often embedded in community organizations that already offer employment assistance. Some sites encourage asset building by low-income couples.
Ideally, the demonstrations would distinguish between impacts of the marriage component alone and impacts of the combined marriage-employment component. However, such a goal requires careful thought on the method of randomization, since contamination is likely if some couples within the same marriage class have access to employment services and the others do not. One possibility is to randomize individuals when they sign up for services to control status (no services), a marriage education class that offers the employment services and wage subsidies, and a class that does not offer employment services and wage subsidies.
In terms of our conceptual framework, this demonstration model focuses on enhancing marriage and job skills, but also aims to improve work and marriage incentives, deliver accurate information, influence preferences, and respond to preferences for near-term rewards. The model may affect the context of individuals and couples to the extent that marriage classes create a peer group for couples. With the employment component added, the group will have employment as well as relationship issues to discuss with each other. The demonstration is well grounded in theory and builds on successful elements of New Hope and other employment-related interventions. It meets the tests of influencing marriage, employment, and family-functioning outcomes, of evaluability, and of potential replication and scale.
The demonstration would test impacts on marital outcomes, relationship quality, and attitudes about marriage, employment, unemployment, wage rates, and earnings. Given the New Hope impacts, we might expect family-functioning benefits as well in improved parent-child relationships and child well-being (such as school-based and behavioral outcomes).
The comparisons of participants in a variety of settings will answer a variety of questions. By randomly assigning individuals to one of three groups, we can examine the relative impact of marriage education with or without the special employment services. Although comparisons between those only receiving marriage-related services (marriage only) and controls will take place in the Building Strong Families and Supporting Healthy Marriages demonstrations, retaining this comparison makes sense to control for the geographic setting when making broader comparisons. The first comparisons will examine the impact of marriage education (relative to control status) on relationship skills, employment, marriage, and family functioning. The second comparisons will focus on differential impacts resulting from the combined marriage education and employment services models on the same set of outcomes.
The employment and subsidy components would substantially reduce the economic barriers and the financial disincentives to marry.20 The combination package of marriage education and employment components would become increasingly attractive, given the expanded emphasis on jobs and income alongside marriage education and relationship skills. Unlike New Hope, the package would include marriage education and a somewhat more favorable schedule of wage subsidies. Still, based on the experience of New Hope, the program’s help in achieving income stability for couples is likely to encourage entry into marriage, discourage divorce, and possibly improve some aspects of family functioning.
The demonstration evaluation should include a cost-benefit analysis to determine whether the combined program represents a good investment and specifically whether any gains from incorporating employment components are sufficient to offset the added costs. Unlike some partial assessments of programs that include subsidized jobs, the analysis would estimate the value of production generated by workers in subsidized jobs. Of course, valuing reductions in divorce and improvements in marriage quality and family functioning would be difficult.
2. Offering Marriage-Related Services to Targeted Unemployment Insurance Recipients
Job loss is associated with a host of negative family functioning and marital outcomes. By helping those who lose jobs maintain healthy marriages, we can support marriage, improve family function, and improve long-term employment outcomes. The basic idea would be to offer marriage education classes to married men and women who claim unemployment insurance (UI) benefits and who are likely to experience long-term unemployment. Already, the state profiling systems within the UI system are able to identify UI claimants who are likely to exhaust their benefits and remain unemployed for over 26 weeks. In addition, the U.S. Department of Labor is sponsoring a Reemployment and Eligibility Assessment (REA) project in 20 states, which provides intensive job counseling and job-finding services to at-risk UI applicants.
Under the project, the Departments of Labor and of Health and Human Services would jointly sponsor another set of REA demonstrations that incorporate the offer (but not the requirement) to take a marriage education program. Even in the current REA demonstrations, some states are using random-assignment approaches to determine the relative effectiveness of alternative treatments. In the context of developing employment plans (sometimes about a month after applying for UI), REA staff would randomly assign claimants in the REA program to treatment or control status. Controls would go through the standard REA program. For those in the treatment group, REA staff would explain the availability of marriage education classes, their general rationale in strengthening marriages, their particular importance in the context of the strains that arise during unemployment, and how to access the marriage education program. Once claimants in the REA program are identified as in the experimental group, organizations providing the services would be encouraged to contact and to recruit them energetically. As in the case of other demonstrations in which participation of the treatment group is well under 100 percent, the experiment would be examining how access to marriage education affects outcomes. In addition, the demonstration would provide a rigorous test of the impact of recruitment on participation.
The demonstration focuses on improving the skills with which couples and families deal with unemployment and potential financial distress. It would offer information in a context in which individuals are primarily concerned with reemployment, but would have no direct effect on incentives. The venue (REA programs within UI programs) would simplify recruitment to the marriage education program. The demonstration deals with a concern about family functioning identified in the empirical literature. The intervention can potentially influence marriage, family functioning, and possibly employment outcomes, and can be readily evaluated and expanded to a large scale.
The key outcomes for the program would be the rate of participation by treatment group claimants, the reduction of risk of divorce, the reduction of family strains commonly associated with unemployment, the speed at which individuals return to work, and the effect on UI benefits. The study would measure various family-functioning outcomes, including marriage relationship-quality indicators and parent-child interactions. The demonstration would show the extent to which claimants will respond to the offer of marriage education classes and thus the desirability of offering marriage services to recruit from this pool at these times.
The evaluation would include a cost-benefit analysis that captures the additional resource costs, the impact on earnings, and the impact on marriage and family functioning. As in the prior proposal, it will be difficult to place a value on reductions in divorce and on improvements in marriage quality and family functioning.
3. Adding Marriage Education to Job Corps and Other Selected Out-of-School Youth Programs
The Job Corps, which serves more than 60,000 new participants per year at a cost of about $1.5 billion, is the nation’s largest and most comprehensive job training program for disadvantaged youth. It is also one of the most expensive federally sponsored education and training programs. This group is at very high risk for having a nonmarital birth (over 25 percent of female participants already have children, nearly all outside marriage), becoming noncustodial parents, and not having a stable marriage. As noted above, despite generating significant gains in education and in earnings lasting up to four years after entering the program, the fade out of the earnings advantage for Job Corps meant that the benefits were not sufficient to offset the social costs of the program. One possible reason for its limited success may be the negative effect on earnings of unhealthy couple relationships. The Job Corps itself recognizes the behavioral challenges that arise in center residential halls and the limited training of residential hall staff to cope with these challenges.
This demonstration would add a marriage education component to the Job Corps experience. The intervention could take place center-wide or be offered randomly to individuals within centers. The random assignment might be most appropriate for Job Corps participants in nonresidential programs. A special curriculum would be tailored to those who are not yet in couple relationships or parents. However, the program would also cover those already in couple relationships and parents. The initiative would complement the Career Success Standards launched by the Job Corps a few years ago to promote a “positive normative culture.” Programs have been revising the required employability skills to include communication, problem solving, conflict resolution, financial management, independent living, and career planning. However, the list does not deal directly with many elements commonly covered in relationship-skills programs.
The demonstration has several rationales. First, in the absence of any intervention relating to marriage, a very large share of participants will go on to have children outside marriage and either not marry or have an unstable marriage. Second, the program offers a setting in which individuals have time to learn marriage and relationship skills and a low-cost method for recruiting potential participants. Third, enhancing marriage and relationship skills is likely to complement efforts to improve job-market outcomes. Research indicates that marriage can contribute to improved employment outcomes for men, including minority and less-educated men. Fourth, the program can be evaluated rigorously through random assignment of participants. A second evaluation approach would deal with models that incorporate marriage education fully into the curriculum at some centers but not others. Difference-in-difference methods and analyses of peer effects would characterize this evaluation. Random assignment of Job Corps centers might be feasible as well, though such an approach might require a large number of participating centers (Schmidt Baltussen, and Sauerborn 2000). Job Corps is a federal program that has already conducted random-assignment activities that raise more sensitive issues regarding the exclusion of potential participants. Fifth, if the program were successful, it could be expanded nationally to all Job Corps programs and potentially to other youth programs such as YouthBuild, the National Guard Youth Challenge Academy, and the Youth Conservation Corps.
The one drawback is that several important impacts may not materialize for several years and may require a long-term follow-up. However, this problem will be present in any program that aims at the critically important task of preventing young people from becoming unwed parents and developing unhealthy couple relationships in the first place. As with the other recommended interventions, the evaluation of this demonstration should include cost-benefit analyses as well as impact analyses.
This demonstration proposal relates to several aspects of the conceptual framework. It would attempt to affect participants’ preferences, skills, and information about marriage and do so in two different contexts—one in which participants live away from their neighborhoods but in a supervised setting with other disadvantaged youth and one in which participants continue to live at home. It would have no direct effect on work or marriage incentives. Recruitment costs would be modest, given the ready identification of and easy access to applicants. The theory and empirical bases for the demonstration include the well-documented impacts of marriage on increasing men’s earnings and on reducing criminal activity. Moreover, past Job Corps results suggest the importance of improving long-term impacts of the program. Thus, the intervention could potentially affect marriage, employment, and family functioning. Finally, the demonstration could be subject to rigorous evaluation and, if successful, readily expanded to scale.
4. A Holistic Marriage-Employment-Family-Functioning Demonstration for Offenders and Ex-Offenders
Given the high rates of imprisonment in the United States, especially for minority men, dealing effectively with the most serious family and employment problems requires improved outcomes for the offender and ex-offender populations. Offenders have low levels of education, do poorly in the job market after release, often lose contact with their children and become uninvolved fathers, and have low rates of marriages and low rates of stable marriages. On the other hand, the evidence suggests good jobs can reduce recidivism and that stable marriages can improve job outcomes as well as reduce recidivism. Moreover, the children of offenders and ex-offenders account for a sizable share of the children most at risk of educational and behavioral problems.
The Administration for Children and Families has already signaled its recognition of the critical importance of reaching this target group by funding the marriage and incarceration demonstrations as part of the OFA Healthy Marriage Demonstration program. Other demonstrations are moving forward as well—including the Serious and Violent Offender Reentry Initiative projects sponsored by the National Institutes of Justice and some foundation-funded demonstrations testing supported work. This proposal recommends another demonstration targeted on these groups, but involving a comprehensive mix of marriage education, employment services (including preemployment training, wage subsidies and subsidized jobs), and family-related services to limit the disruption of family life when the offender enters prison or is released from prison. As in our other suggestions, the theory is that the mix of marriage, employment, and family services may be highly complementary—the presence of marriage education may enhance the impacts of employment-wage subsidy benefits and vice versa.
The potential target populations would be offenders who are romantically involved or married and who are entering prison for short-term stays (2 years or less), offenders entering work-release programs, and offenders reentering the community. Although some might favor focusing only on fathers, we believe that reaching men before they become fathers is desirable. Moreover, it may be of value to men who have not yet fathered a child to learn about the child-support obligations of noncustodial fathers.
Work-release programs offer an especially good target of opportunity. Offenders are near the end of their incarceration. The programs already have an employment component and perhaps others. The demonstration can substantially enhance the employment components by incorporating mentoring and wage subsidies as well as adding marriage education. As noted above, the topics covered in the marriage education program are likely to improve the individual’s noncognitive, job-related skills.
In one possible variant, the program might incorporate a mandatory jobs component for individuals on parole. If only job search and training were available, reentering offenders might otherwise take too long to find a job and ultimately return to criminal activity. Under the approach suggested by Mead (forthcoming), not only would ex-offenders be provided with help in seeking jobs, but jobs would be guaranteed. This assured availability of jobs could be used to make work mandatory—those not accepting some job (including the guaranteed job) would be subject to sanctions such as more stringent parole or being returned to prison or jail. Mead quotes Christopher Jencks as arguing that if jobs were guaranteed to jobless adults of ghetto areas, community pressure would induce many to take work seriously and accept jobs, even if they are low-paying. Mead would include pre-employment training and initial support services (health, housing, transportation) along with the job guarantee.
Mead would extend coverage of the approach to noncustodial parents who do not pay child support. Again, the assured job component would permit agencies to make work mandatory. Already, in some jurisdictions, judges require noncustodial parents to find some way to pay child support or face jail. Such policies are difficult to enforce because of uncertainties about an individual’s ability to find and hold a job. The assured job provisions would increase the credibility of the sanctions, since judges would know that the obligor is choosing not to work and thus not to pay support even though a job is readily available.21
The programs are likely to prove easiest to evaluate if they take place before an individual is fully released into the community. In general, officials running programs for offenders and ex-offenders are receptive to incremental funding and willing to undertake experiments. In the case of the work-release centers, a random-assignment experiment would have to involve randomly assigning individuals to work-release centers with and without the combined mix of services. In some ways, this model would simultaneously capture the impact of services as well as the impact of the context of being in a work-release center with enhanced services for everyone. In this respect, the demonstration would resemble aspects of the Job Corps evaluation.
The demonstration’s family-functioning components would incorporate best practices in helping families adjust to the individual’s absence, to make it easier for families to remain closely connected to the offender, and in insuring a smooth transition from prison to civilian life. The project would include efforts to resolve past child-support arrearages and current obligations in ways that improve the work incentives of fathers while retaining their connections with and support for their children. Several initiatives are already taking place in this field, including demonstrations in Maryland and Minnesota and provisions in several states in which debt forgiveness can take place as noncustodial parents establish a record of meeting current obligations (Ovwigho, Saunders, and Born 2005; Pukstas et al. 2004). The problem of high arrearages deterring men from working in the mainstream economy is particularly important for ex-offenders, since very few states suspend the payment of child-support orders while individuals are in prison (Holzer et al. 2005). Thus, a holistic program should deal with this issue while individuals are learning job skills and gaining a foothold in the job market.
This demonstration attempts to modify preferences, change incentives, increase information, and enhance skills. The context for the program poses advantages and disadvantages. Some advantages are the ease of recruitment of a critical target group and the integration of marriage, employment, and family-functioning approaches that avoid major gaps which might prevent participants from leading constructive and productive lives. The disadvantage is that the venue is associated with the stigma of criminality and has a concentration of people at high risk of returning to crime and of influencing peers to do so as well.
In terms of demonstration criteria, the initiative has a solid theoretical basis and empirical data documenting the needs of ex-offenders and the value of marriage in increasing earnings and reducing crime. However, there is no good record (say, from past demonstrations) that these components will succeed for the target population. In principle, however, the intervention could have substantial effects on marriage, employment, and family functioning. The demonstration could be subject to rigorous evaluation and, if successful, readily expanded to scale. Current OFA demonstrations might provide opportunities for pilot testing these approaches and determining their feasibility.
5. Strengthening the Functioning of Families for Parents Working Nonstandard or Irregular Hours
About one in three employed individuals work on a weekend day and about 15 percent of full-time employees do not work on a daytime schedule.22 Parents who work nonstandard or irregular hours face special challenges in maintaining a healthy marriage, parenting effectively, and dealing with child care requirements. Although the evidence is mixed and sometimes nonstandard hours can be helpful, some studies find that shift work, night shifts, and irregular hours raise the likelihood of separation or divorce (Presser 2000; Stradzins et al. 2006; White and Keith 1990). If family-functioning problems arise from these special circumstances for combining work and family, then family problems may spill over to the work lace in the form of increased turnover and reduced productivity. Presumably, employers use nonstandard hours and irregular shifts to run their operations effectively, often to provide customers with services at night or to operate their plants and equipment at full capacity. Many pay a night-shift differential to compensate somewhat for requiring work during nonstandard hours. While workers voluntarily take these positions and about 8 percent report that nonstandard hours allows for better family and child care arrangements, about 8 percent report it is the only job available. Others may lack sound information for judging potentially negative family consequences or the skills to cope effectively with them.
Given legitimate concerns about potential harmful impacts of nonstandard work, there is a case for a demonstration to determine whether a well-designed set of services and incentives can prevent negative consequences on families and can benefit employers as well. One option is to work with multisite employers who use nonstandard and irregular schedules for many employees. The organization putting together the demonstration would approach these employers and offer to provide services and incentives to some workers but would require some financial participation by employers, such as paid time to attend classes. Another project component could involve training supervisors and managers about how to mitigate harmful impacts on workers and their families while maintaining high productivity. Initially, the demonstration would recruit only parents (including married, cohabiting, and single parents) and would provide such services as marriage education or relationship-skills classes, classes that offer referrals to services and other means of coping effectively with nonstandard hours, and possibly a modest stipend that parents can use to deal with special needs. Among the skills taught would be how to structure child care with friends, families, and neighbors. The demonstration would not use random assignment within the work site, in part because of the likelihood of contamination (treatment group parents receiving services would commonly interact with control group parents in the same work site) as well as likely employer opposition. Instead, the analysis would use a difference-in-difference analysis to test whether family outcomes improved in the sites offering services relative to matched comparison sites. The follow-up data would cover not only those still with the organization, but also those who become unemployed, left the work force, or took other jobs. The evaluation would also attempt to determine whether employers experience any positive (or negative) impacts of the services. Ideally, if the demonstration proved sufficiently effective for families and employers, many employers might choose to sponsor similar services.
An alternative demonstration strategy would be to use a randomized encouragement approach to test the provision of services for parents working nonstandard hours. In a community with an organization funded to provide services, a survey firm would conduct a household survey to determine eligibility for services and, at random, to encourage some eligibles to take up the services and to provide no information or encouragement to others. The evaluators would follow both the encouraged households and eligible households not encouraged to participate for at least two years. This approach has advantages and disadvantages over the employer-based method. It offers the chance to obtain experimentally-based and unbiased estimates of the impact of encouragement and information on participation in the program and, assuming some positive encouragement effects, of the impact of services on family functioning. On the other hand, the approach requires a large sample, its effectiveness depends partly on the impact of encouragement, and it provides only partial evidence about potential impacts on employers. Further, the employer-based services might generate peer effects by linking participants with participating co-workers and might offer participants the assurance of employer support. Finally, since identifying gains for employers might be one way to generate long-term funding for the services, a demonstration not directly linked to employers may be disadvantageous.
The demonstration fits the framework by recognizing potential gaps in information and skills faced by parents working nonstandard hours. It draws on empirical evidence about one source of potential family-functioning problems.
6. Adding Marriage Education and Relationship Skills to the Nurse Home Visiting Intervention
The increasing evidence that the Nurse Home Visiting Program is highly cost-effective (Aos et al. 2004) has led to proposals for expanding the intervention nationwide (Isaacs 2007). Yet, although the program as constituted appears to yield some benefits for children and mothers, the gains might be enhanced significantly by combining the intensive visits with marriage education or relationship-skills classes. While less than 2 percent of participating mothers were married at the time of the intervention, 15 percent were married and over 75 percent had a partner at the time of the follow-up. It is well-known that tensions between partners rise soon after the birth of a child. Thus, it makes sense to combine advice on taking care of infants and toddlers with information and skill-building to sustain close partner relationships. Certainly, nurses should be sure to make mothers aware of the importance of marriage and father involvement for the long-term economic and social health of children. Any expansion of the Nurse Home Visiting approach should be informed by the potential effectiveness of high-priority, complementary services, such as marriage education and relationship-skills training.
A demonstration could test the existing Nurse Home Visiting Program against a Nurse Home Visiting Program enhanced by marriage education and relationship-skills components. Evidence from the Supporting Father Involvement Demonstration (Cowan and Cowan 2007) indicates that emphasizing relationship skills can have as positive or even more positive impacts on children than emphasizing parental skills alone. The additional components could work as follows: (1) train nurses to learn about and communicate the long-term benefits for children of marriage and healthy fathering for children, (2) attempt to engage fathers in the standard array of nurse home-visiting activities, (3) give participating mothers (and, where appropriate, fathers as well) formal invitations to participate in marriage education/relationship-skills training, (4) use existing marriage education/relationship-skills providers to contact and recruit participants in the enhanced program, (5) train nurses to learn about the programs and explain their value, and (6) insure that sufficient marriage education/relationship-skill classes are available for participants who choose to take advantage of these services. The demonstration would randomly assign potential participants to either the standard program or an enhanced set of services. In this case, the enhancement would be a well-structured marriage education component. The follow-up interviews will take place in the context in which both groups will have been receiving services and thus should reach a high share of the sample.
The demonstration would yield answers to several important questions. Can the Nurse Visiting Program effectively encourage participation in marriage education/ relationship-skills classes? To what extent do the offers and encouragement stimulate participation in these classes? If the enhanced component increases participation in these classes and related components, do these services increase relationship quality, stability between partners, cohabitation, and marriage? Are the gains for children higher in the enhanced Nurse Visiting Program than in the standard program? The proposal fits well into the conceptual framework. The program may alter a mother’s preferences by raising the priority she places on a good relationship with her partner or husband, as well as provide her with the skills to be effective in the dual roles of mother and partner. The marriage component can also deliver information to mothers about the importance of stable families for childrearing.
7. Linking Marriage Education, Mentoring, and Expanded Work-Based Learning for Youth
The decline in teen pregnancy is noteworthy but so is the continuing reality that 34 percent of girls become pregnant during their teenage years, almost always outside of marriage. Moreover, half of all first nonmarital births are to teenagers (Whitehead and Pearson 2006). These early actions complicate the route to a healthy, stable marriage and the raising of children in two-parent families. Once a woman has a nonmarital birth, her likelihood of having a long-term marriage declines substantially (Bennett, Bloom, and Miller 1993). If she marries a man who was not the father of her first child, child outcomes may suffer because of the poorer outcomes for children in stepparent families as opposed to biological or adoptive parent families. It is critical to increase the share of young people who recognize these realities, who do not engage in unhealthy relationships, and who delay childbearing until after marriage.
Any demonstration that attempts to address teen pregnancy and early unwed childbearing should recognize the lessons of youth-development programs. The best strategies help young people achieve their own positive goals rather than simply suppressing problem behaviors (Moore and Zaff 2002). They try to use positive peer influences and long-term mentoring. The best programs for disadvantaged youth address several barriers and deal with the whole person. Although programs can focus on the disadvantaged, they should not involve a concentration of youth criminal offenders.
Young people are making important transitions that involve increased responsibility for their own decisions—in sex lives, couple relationships, living arrangements, marriage, career choices, postsecondary education, and financial independence. Educational institutions and families provide young people with only some preparation for these life events. Critical gaps exist, especially in preparing many youth for productive careers, healthy marriages, and healthy parenting.
This demonstration would test a multipronged strategy to help young people develop achievable pathways toward rewarding careers, see the value of delaying childbearing until marriage, and gain the job-related and relationship skills necessary for success in career and family pursuits. The components of the program would involve both employment-oriented and marriage/family education elements. One potential venue is Career Academies, noted above as a successful intervention raising the earnings of young men, especially students with a high or medium risk of dropping out of high school. However, the venues could vary from Career Academies, to other youth job-focused programs, to community colleges, or to family planning clinics dealing with teens. To ensure the demonstrations are well-targeted, they should operate in areas with a concentration of the target population.
The employment component would focus on expanding work-based learning, ideally in an apprenticeship context. All participants would be given access to a combined work-based and school-based program. Such a model will provide youth with productive experiences, natural mentors, and constructive interactions with adults. Learning in context will help students see the relevance and gain the self-confidence many at-risk students lack. Placing inner-city youth in jobs in fields that can lead to rewarding careers will reduce the disadvantage they may face with respect to information and informal channels to jobs and careers. Linking youth with employers will expose inner-city youth to constructive adult peers who will often become informal mentors. The approach gives youth a try-out period with employers, many of whom will hire the youth once they graduate from high school.
Along with this employment component, participants would be exposed to an enhanced marriage and relationship curriculum that (1) teaches about healthy dating and uses experiential learning in the teaching of relationship skills, (2) offers abstinence or comprehensive sex education, and (3) teaches about love and marriage and about the importance of raising children in two-parent families. Several curricula are addressing these issues and more will emerge as many of the ACF Healthy Marriage grantees undertake programs within high schools. For example, Opportunities Industrialization Centers of America (OICA) is implementing an OFA grant to provide education in high schools on the value of marriage relationship skills and budgeting, and to provide marriage education, marriage skills, and relationship-skills programs for nonmarried pregnant women and nonmarried expectant fathers. OICA expects to serve 1,500 high school youth and 50 expectant women and unmarried fathers with instruction and support services over 12 weeks. Adding a strong career-focused component to the program could generate synergy in the mix of skills required in relationships and jobs. In addition, the two components could complement each other in giving youth a realistic way to reach both career and family-formation goals.
In developing a holistic approach to teaching life skills to participants, the demonstration would build on the experience of prior youth-development programs and on the teaching of skills documented in the reports of the Secretary’s Commission on Achieving Necessary Skills (1992) and of the Partnership for 21st Century Skills.23 Some of the skills in relationship curricula—especially in communication, listening, problem solving, and allocating resources—parallel those emphasized as required for careers. The emphasis on work-based learning will provide youth with the opportunity to practice these skills in context.
This holistic youth demonstration emphasizes the information and skills aspect of our framework, but includes an effort to change preferences, improve incentives, and utilize constructive contexts as well. Given the improved and increasingly visible career options, young women would have more of an incentive to delay pregnancy to pursue occupational outlets. The enhanced career success of potential male partners would encourage women to see marriage as a more viable and sensible option. The initiative builds on the theoretical and empirical youth-development field and on empirical data showing the importance of early interventions. Evidence from other demonstrations indicates these components can be effective for the target population. Although the demonstration can ultimately exert significant positive effects on marriage, employment, and family functioning, the impacts will take time to materialize and document. Finally, the demonstration could be subject to rigorous evaluation and, if successful, readily expanded to scale.
| Table of Contents | Previous | Next |