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

This section summarizes some of the differences among two-generation programs based on examinations of six such programs (Table 1). These six were selected because they represent large and well-known federal, state, and locally developed programs; they exhibit variation in programmatic approach; and they have been or are being evaluated via randomized trials.²¹ These six programs as as follows:

• Avance.²² The Avance Family Support and Education program began in the 1970s in San Antonio, Texas. It seeks to help children succeed in school by teaching parents to teach their children and by meeting parents' educational and job training needs. About 2,500 families participated in Avance programs during 1994.

• Child Family Resource Program (CFRP).²³ The Child and Family Resource Program (CFRP) was funded from 1973 to 1983 by the U.S. Department of Health and Human Services. CFRP was based on the premise that the best way to promote children's growth and development is by supporting families and helping parents become more effective caregivers and educators. Comprehensive social and educational services were provided by 11 projects to more than 1,000 families per year.

• Comprehensive Child Development Program (CCDP).²⁴ The Comprehensive Child Development Program (CCDP) began in 1990. Five-year grants were made by the U.S. Department of Health and Human Services to 34 projects which provide comprehensive, continuous, coordinated social, health, and educational services to low-income families with a newborn child for up to five years. More than 4,000 families participated in CCDP during 1994.

• Even Start.²⁵ The Even Start Family Literacy Program was first funded in 1990 by the U.S. Department of Education. It allows great local discretion to more than 400 grantees but mandates that participating families take part in early childhood education, parenting education, and adult education. About 30,000 families participated in Even Start during 1994.

• Head Start Family Service Centers (FSCs).²⁶ Head Start FSCs began in 1990. Sixty-six projects were funded by the U.S. Department of Health and Human Services. Each provides regular Head Start child development and parenting, augmented by case managers who assess adults' needs, deliver services to adults, and make referrals for literacy, substance abuse, and employment training. The FSCs served more than 4,000 families in 1993.

• New Chance.²⁷ New Chance was a comprehensive program for disadvantaged young mothers and their children. Funded with a combination of public and private dollars, it operated between 1989 and 1992 at 16 locations, serving more than 1,500 families. Program components included case management, intensive educational services, and free child care.

The Early Childhood Component

Two-generation programs vary in the age of the children they serve, the duration and intensity of the services offered, and the way in which services are delivered. For example, some programs target three- and four-year-olds (Head Start FSCs), and some focus on children from birth to school entry (CFRP and CCDP), while others specify a wider age range, such as birth through age eight years, but leave the exact age and duration of service to the discretion of local grantees (Even Start). Some programs intend a set period of service duration ranging from one year (Head Start FSCs) to five years (CCDP), while others allow this dimension to vary according to family needs (Even Start).

Child-focused services are usually delivered via home visits (CCDP, CFRP) or in centers (Avance, New Chance), with center-based programs usually providing more intensive services. For example, CCDP provides relatively low-intensity services to children from birth through age three using biweekly home visits for a maximum of 30 minutes during which the focus of instruction is on teaching parenting skills. CFRP's early childhood component was even less intensive, providing 15 minutes of child development during monthly home visits. However, most CCDP and CFRP projects and all Head Start FSCs enroll their four-year-old children in Head Start which provides a year-long, half-day, center-based program.

Avance offers educational child care for infants and toddlers for three hours per week while their mothers are in parenting classes. Similarly, New Chance provides free child care in high-quality centers while mothers are taking part in program activities.

Even Start projects exhibit huge variation in intensity of early childhood services across sites. Children in the middle 50% of the projects (in terms of number of hours of early childhood services received) receive between 21 and 330 total hours of service during their period of participation.

The Parenting Education Component

Most two-generation programs include a set of services for parents designed to enhance parenting skills, increase parent involvement in schools, improve parental self-esteem and coping skills, alleviate parental depression, and develop parents as teachers and role models. Programs, however, exhibit the same variation in service delivery mode, intensity, and duration of parenting services as do child-focused services.

Avance takes a high-intensity approach in which mothers attend a center for three hours per week for one school year to strengthen parenting skills, learn about toy making, and become familiar with community resources. New Chance is another higher-intensity parenting program in which mothers attend full-day classes; parenting education is included as part of the daily curriculum.

The Head Start FSCs, which rely on the basic Head Start model of involving parents in governance and service delivery, offer a lower-intensity parenting intervention. CFRP and CCDP both deliver parenting education through home visits (monthly for CFRP and biweekly for CCDP). Even Start allows projects flexibility in designing the parenting component. The average Even Start adult receives 58 hours of parenting education over a seven-month period, sometimes delivered in the home, sometimes in group sessions.

The Adult Education/Employment Training Component

Programs also vary considerably in how they deliver adult education or job training services. New Chance has a two-phase approach in which full-day, full-week, classroom-based adult education leading to GED attainment is emphasized early on, followed by vocational training, internships, and job placement assistance. Even Start mandates that parents take part in adult education (adult basic education, adult secondary education, GED preparation, or English as a Second Language courses); but local projects vary in the degree to which they emphasize this component and in the intensity and duration of services provided. The Head Start FSCs typically do not offer adult literacy or employment training services directly but provide them through case management and referrals to other community agencies. CCDP sites rely heavily on case manager intervention with parents to help them obtain adult literacy education, vocational training, employment counseling, and job training. The intensity of this component varies substantially across projects.

Still other programs are relatively weak in this area, providing little more than a referral service (CFRP). Avance added a low-level adult literacy and job training component to its existing services to provide basic literacy instruction for mothers who complete the parenting program; if desired, these services are available for multiple years.

In sum, while all six of these two-generation programs offer early childhood, parenting, and job training or adult literacy services, they vary considerably in how these services are delivered and in their intensity and duration. Such wide variation in services and in delivery amounts may well lead to variation in program effects.

The six programs reviewed in this article have been or are being evaluated using high-quality, randomized experimental studies (see Table 2). Child outcomes that were investigated and that are summarized below include cognitive and literacy skills. Outcomes for parents include assessments of attitudes toward parenting, parenting behaviors, and self-sufficiency behaviors such as attainment of GED and literacy skills. At this time, most of these evaluations have reported only on the relatively short-term effects of two-generation programs, having followed families from the birth of their children through the children's fifth birthday (Figure 1).

Short-Term Effects on Children

On the whole, the studies reviewed here show small or no effects on child development (see Table 2). Neither Avance nor CFRP had any effects on several measures of child development. CCDP had a small positive effect on child development for children at age two years. Even Start had a medium-sized effect on a measure of school readiness skills nine months after entry to the program; however, children in the control group caught up once they entered school. Even Start children also demonstrated gains on a test of language development; however, control group children achieved similar gains. Neither the New Chance nor the Head Start FSC evaluations measured child development outcomes.

Short-Term Effects on Parenting

Several programs had positive effects on parenting. For example, Avance reported positive effects on the home learning environment, child-rearing behaviors and attitudes, maternal role as a teacher and sense of parental efficacy, and use of community resources; New Chance showed positive effects on child-rearing attitudes and on emotional support for children; Even Start had a positive effect on the presence of reading materials in the home; CFRP had positive effects on attitudes and practices relevant to child rearing and on parent-child interactions; and CCDP had positive effects on parenting attitudes, expectations for child's success, time spent with child, and mother-child interaction.

Short-Term Effects on Adult Education/Job Training

Even Start, New Chance, and Avance led to large increases in the percentage of mothers who attained a GED certificate (22% of Even Start mothers versus 6% of controls; 37% of New Chance mothers versus 21% of controls). Unfortunately, attainment of a GED was not accompanied by corresponding positive effects on standardized tests of adult literacy.

Participants in two-generation programs increased their use of federal benefits such as AFDC and food stamps (CCDP, CFRP, New Chance). In each instance, this finding resulted either from increased numbers of families eligible for benefits because of increased participation in educational classes, or simply through increased awareness of the availability of federal benefits.

None of the studies that measured annual household income (Even Start, New Chance, CCDP, CFRP) found a positive effect on this variable, and only CFRP had a positive effect on employment. Further, none of the programs made a measurable difference on variables such as maternal depression, maternal self-esteem, or the use of social supports.

Who Benefits Most?

Some of the evaluations included analyses to determine what types of families benefit most from participation in a two-generation program and what types of services are most beneficial. These analyses are correlational in nature, and conclusions drawn from them are subject to competing explanations; for example, self-selection of families into service categories or amounts of service, or differential motivation of families. However, the findings are provocative and suggestive of potentially important trends.

Evaluations of Even Start, New Chance, and CFRP all found a positive relationship between amount of participation and program benefits. The Even Start evaluation found that high levels of participation in early childhood education were associated with larger gains in school readiness skills and vocabulary, and that high levels of participation in adult education were associated with larger gains in adult literacy and increased rates of GED attainment. The New Chance evaluation found that amount of participation was positively related to GED attainment and to reduced pregnancies.

Results of the Even Start evaluation also suggest that the children of parents who participate the most in parenting education show greater gains in language ability than children with less participative parents. This suggests that, even in the short term, it is possible to find the parent-to-child linkage hypothesized by two-generation programs.

Several of the evaluations conducted analyses to determine whether program effects varied across demographic subgroups such as teenage versus non-teenage mothers, mothers who completed high school versus mothers without a diploma, male children versus females, relatively high-income versus relatively low-income families, and so on. None of the evaluations in which these questions were addressed (CCDP, CFRP, New Chance, Even Start) found that effects differed across subgroups, except for Avance, which found better educational and parenting outcomes for mothers who were better educated and married at enrollment.

Each evaluation in which the effects for individual sites (Even Start, New Chance, CCDP, CFRP) were investigated found large site-to-site variation. One of the strengths and, at the same time, one of the weaknesses of large-scale studies conducted on ongoing demonstration programs is that such studies are based on data from individual projects that vary widely in the quality of implementation and in the activities implemented from project to project. Naturally, the evaluations described here average the results from sites that have taken different approaches to implementation.

These program outcomes were achieved at widely varying costs, both in direct program costs and in the costs of all leveraged or brokered resources used by the program in question. CFRP, CCDP, Head Start FSCs, and Even Start are federally funded programs which require that grantees not use federal funds to duplicate services that can be obtained locally. Rather, they must build on existing services and use program funds only to "fill the gaps" in service provision and to ensure that families receive required services. Thus, these programs often use Head Start to provide the early childhood portion of the program for four-year-old children, and Even Start and CCDP use local adult education programs to satisfy their adult literacy components. Calculating the costs of these leveraged or brokered services is necessary to understand the full cost of two-generation programs.

Direct Costs

Direct costs for two-generation programs vary widely, both on a per-year basis and in terms of the number of years that a family might participate in the program. (Table 1 contains cost estimates, converted to 1994 dollars.) At the upper end of the cost spectrum, CCDP costs $8,632 per family per year, with the intent that a family participates for five years, and New Chance costs $8,311 per family for 6 months, the average length of participation out of a possible 18 months. Even Start, CFRP, and the Head Start FSCs each cost between $2,500 and $3,500 per family per year. Avance costs about $1,600 per family per year. Families in CFRP and CCDP can participate for up to five years, families in the Head Start FSCs can participate for three years, families in Avance participate for up to two years (53% participate for nine or more months), and there is no set length of participation in Even Start (the average is less than one year).

Per-family costs typically vary substantially across program sites. In Even Start, for example, about half of the projects spent $2,000 to $4,000 per family, but 22% spent less than $2,000 per family, and 33% spent more than $4,000 per family. For New Chance, site-level costs ranged from a low of $4,758 per family to a high of $16,846. In CCDP, costs ranged from $4,592 to $13,413 per family across 24 program sites.

Full Program Costs

The direct costs discussed above do not include the costs of referred or brokered services. Studies of two programs (New Chance and Even Start) calculated the costs of those additional services. In 1991, leveraged or referred services added 54% to federal program costs for Even Start sites: federal Even Start costs were $2,663 per family per year; referred and leveraged services such as adult education, meals, or transportation cost $1,438, yielding a total cost of $4,101 per family per year (in 1994 dollars). For New Chance, the cost of services supplied by other agencies added 18% ($1,496) to the cost of services supplied by the sponsoring agency ($8,311), so that the total cost per family was $9,807 for six months of service.

Allocation of Program Costs

Even Start and New Chance varied in how they allocated program costs among service components in 1991. Although both spent about two-thirds of program funds on the direct provision of services and about one-third of program costs on other functions, there are differences in how costs are allocated within those two categories. Even Start emphasizes early childhood education (31%), adult education (15%), and parenting education (9%), while New Chance emphasizes child care (29%) and case management (27%).

In contrast, in 1991, the overall allocation of costs in Head Start programs was also about 70% for direct service provision and 30% for other costs, but again, the distribution of costs within categories varied with the majority being spent on education (41%), administration (13%), and occupancy (13%).

Comparative Cost of Single-Component Programs

One of the most intensive and effective child-focused programs (the Infant Health and Development Program) costs about $10,000 per family per year, and the intent is that families participate for a three-year period.²⁸ Head Start, the largest and oldest child-focused program in the nation, costs about $4,000 per child per year.²⁹ Because few families have more than one child in Head Start at the same time, $4,000 is considered to be Head Start's annual per-family cost. At less than $1,000 per year, the Missouri PAT program—which involves monthly, 60-minute home visits, primarily for parenting education, and monthly parent group meetings—is a much less intensive and less expensive child-focused program.³⁰

A range of estimates is available for the per-participant cost of an adult education/job training program. Basic adult education programs funded by the federal government cost quite little—between $100 and $500 per participant.¹⁷ Costs for Job Opportunities and Basic Skills (JOBS) training programs range from $100 to $1,000 per participant for less comprehensive programs¹⁴ to $1,400 to $3,900 per participant for more comprehensive versions of the programs.^31,32 California's Greater Avenues for Independence (GAIN) program costs about $3,000 per participant.³³