The Future of Children, Princeton - Brookings: Providing research and analysis to promote effective policies and programs for children.

By design, two-generation programs strive to increase the participation of mothers and children in early childhood education, parenting education, and adult education/job training. Case management services are delivered, services are brokered, and support services are made available and utilized. These comprehensive, multigenerational programs can and have been implemented, with varying degrees of success, in a wide range of settings.

Evidence about the short-term effects of two-generation programs is mixed. On balance, the evidence supports the following conclusions:

• Two-generation programs increase the rate of participation of children and their parents in relevant social and educational services.
• As currently designed, two-generation programs have small short-term effects on a wide set of measures of child development.
• Two-generation programs have scattered short-term effects on measures of parenting including time spent with child, parent teaching skills, expectations for child's success, attitudes about child rearing, and parent-child interactions.
• Two-generation programs have large short-term effects on attaining a GED, but these are not accompanied by effects on adult literacy. There are few effects on income or employment. There are no effects on the psychological status of participating mothers as measured by level of depression, self-esteem, or use of social supports.
• Many correlational analyses show that amount of participation is positively related to test gains and GED attainment.
• There is little evidence that two-generation programs are any more or less effective for important subgroups of participants.
• Where we find positive effects, those effects are generally small (except for effects on GED attainment).

This is a mixed assessment of the short-term effects of two-generation programs. It says little about anticipated long-term effects, but many researchers believe it is not reasonable to expect long-term effects in the absence of substantial short-term effects. The Even Start evaluation provides some evidence supporting the hoped-for cross-generational effectiveness of these programs because it found the amount of parenting services provided to mothers to be related to children's vocabulary. However, this is not an experimentally derived finding, nor has it been replicated in other studies.

How can two-generation programs be improved to maximize their short-term success? There is substantial evidence that effects on children are best achieved by services aimed directly at children, and effects on parents are best achieved by services aimed directly at parents. There is only limited evidence to support the indirect method of achieving large effects, that is, achieving effects on children through earlier effects on parents. (See the article by Barnett in this journal issue.) Thus, it is important for a two-generation program to provide early childhood education services directly for the benefit of children and not to assume that it is just as good to provide parenting services to mothers who will then act as enhanced intervenors in their children's lives.

Intensity of services matters for each component of a two-generation program. No program reviewed here (and no other to our knowledge) provides anything close to the intensity of child-focused services delivered by the highest quality early childhood programs such as the Infant Health and Development Program (IHDP), which calls for a full-day, full-year, center-based program for children from age one through age three years.³⁴ The consequence of having a low-intensity early childhood component is that two-generation effects on children's cognitive development, while real, are small. High-intensity programs such as the IHDP have short-term cognitive effects on children which are on the order of 5 to 10 times as large as the effects of low-intensity programs.

The parenting component of the two-generation programs reviewed here also is relatively weak. Wasik and Karweit³⁵ compared the effectiveness of interventions that included parent-focused and child-focused components of different intensity levels. They classified as "low intensity" those parenting interventions that consisted mainly of weekly or biweekly home visits along with occasional parent meetings. This level of intervention sounds a lot like what is provided in many of the two-generation programs described in this article. Wasik and Karweit concluded that the most effective interventions included intensive child and parent services which involved a center-based program for children and meetings with parents on a weekly or semiweekly basis for at least a year. Low-intensity parenting components did not add much, if anything, to the effectiveness of a high-intensity child component.

The same worry applies to the intensity of adult education or job training services: two-generation programs may not deliver enough of any adult-focused service to match the service levels offered by the best of the single-component programs. Because all three components must be included in a two-generation program, along with case management and support services, two-generation programs are in danger of taking a broad-based approach that does not provide enough of any single service to be effective.

There are practical limits on the amount of public funds that program administrators (and taxpayers) are willing to allocate to disadvantaged families. Programs that provide high-quality infant stimulation as one component (IHDP) cost at least $10,000 per child per year. High-quality parenting programs such as Avance's add at least $1,000 per year to this total, and a high-quality adult education and job training program such as California's GAIN costs another $3,000 per year. This means that a two-generation program which incorporates three high-intensity components could easily cost $12,000 to $15,000 per family per year. This is 50% to 100% greater than the most expensive two-generation programs (CCDP and New Chance).

However, considering the high costs of several other educational or social "programs" may help put the costs of two-generation programs in perspective. For example, a year of public schooling costs about $6,000 to $10,000; a year of special education in the public schools costs about $20,000; a year of private schooling costs about $10,000 to $15,000; and a year of private higher education costs $20,000 to $30,000. Viewed from this perspective, the costs of high-quality two-generation programs do not seem so intimidating. Why should high-quality preschool, parenting, and adult education services, delivered to disadvantaged children and parents, cost less than high-quality educational services delivered in other settings?

The two-generation intervention strategy was initiated with limited evidence about the most effective way in which to implement the parenting and adult literacy components. A conservative strategy for enhancing two-generation programs would be to conduct planned variation research by building on known effective practices. A reasonable approach would be to conduct a series of small-scale research studies in which different promising parenting and/or adult education strategies are systematically appended to a single high-quality early childhood program to test experimentally their differential effectiveness in producing positive outcomes for parents and in enhancing children's development.

One problem with this approach is that, while considerable research evidence exists on what constitutes a high-quality early childhood program, no consensus exists on what constitutes a high-quality parenting or adult education/job training program. The research on adult education programs cited earlier in this article suggests that most adult education programs tend to replicate the poor high school settings in which participating adults initially failed. When this happens, the common two-generation strategy of using existing community-based adult education services is doomed to failure. Before high-quality adult education can be satisfactorily incorporated into two-generation programs, better approaches need to be developed.

The modest results described here are sobering but not surprising because two-generation programs may well be struggling to fix problems that are beyond their grasp. Many families recruited to participate in these programs are deep in poverty, facing the most adverse circumstances of substandard housing, substance abuse, inadequate incomes, and dangerous neighborhoods. Given the history of small effects commonly associated with programs aimed at alleviating poverty, and based upon the mixed evidence presented in this article, it is naive to think that there will be quick, easy, or inexpensive fixes. Further, current welfare reform efforts may face an uphill task. There is no evidence that the two-generation approach, as currently structured, can move substantial numbers of families from the welfare rolls in two years, or for that matter, in any amount of time. However, the results of the two-generation programs reviewed here follow children only through their fifth birthday, although some of the programs are designed to continue until children are age eight or older. Follow-up data should be collected regarding the families participating in these evaluations to determine whether hypothesized longer-term effects and linkages between effects on parents and effects on children do, in fact, occur.