U.S. Department of Health and Human Services

Head Start: What Do We Know About What Works?

Sharon M. McGroder

Office of the Assistant Secretary for Planning and EvaluationU.S. Department of Health and Human Services

March 29, 1990

PDF Version: http://aspe.hhs.gov/daltcp/reports/headstar.pdf (24 PDF pages)

---

This report was prepared by the U.S. Department of Health and Human Services (HHS), Office of Family, Community and Long-Term Care Policy (now the Office of Disability, Aging and Long-Term Care Policy). For additional information about this subject, you can visit the ASPE home page at http://aspe.hhs.gov. The Project Officer was Sharon McGroder.

The opinions and views expressed in this technical analysis paper are those of the author. They do not necessarily reflect the views of the Department of Health and Human Services, Office of Management and Budget, or any other funding organization.

---

TABLE OF CONTENTS

I. BACKGROUND

II. WHAT DOES THE RESEARCH SHOW?

Head Start Evaluations

Other Preschool Program Evaluations

Program Characteristics Associated with Positive Outcomes

III. WHAT RESEARCH IS NEEDED

Follow Through

Project Developmental Continuity

Current Initiatives

IV. CURRENT AND PLANNED RESEARCH EFFORTS

Head Start Evaluation Design Study

Long-Term Benefits and Cost-Effectiveness of Head Start

V. ADDITIONAL TOPICS RELATED TO HEAD START'S EFFECTIVENESS

Coordination with State Preschool Programs

Coordination with the Family Support Act

Allocating Expansion Funds for Quality, Enrollment, and Management

REFERENCES

NOTES

I. BACKGROUND

Head Start programs provide comprehensive child development, educational, health, nutritional, social and other services to predominantly low-income preschool children and their families. In recognition of the important role parents play in their child's development, Head Start programs are required to provide for the direct participation of parents in the development, conduct, and direction of local programs.

In FY 1989, over 1900 grantees and delegate agencies provided services to over 450,000 children, 65 percent of whom were four-year-olds. These agencies operated 24,000 classrooms and employed 80,000 staff nationwide. As a result of Head Start's emphasis on parental involvement, 32 percent of Head Start staff were parents of current or former Head Start children. In FY 1989, Head Start programs spent over $1.2 billion, for an average annual cost-per-child of about $2,660.

The President's FY 1991 budget provides a $500 million increase for Head Start. This would mark the single largest funding increase for Head Start since it began providing full-year services in 1966. This expansion of Head Start, taking place in the context of many other program and policy developments, has led policy-makers to explore more closely the evidence of its success.

II. WHAT DOES THE RESEARCH SHOW?

Head Start Evaluations

The Westinghouse study, 1969

In 1969, the Westinghouse Learning Corporation completed the first major evaluation of Head Start. Summer programs were found to have no lasting impact. Full year programs resulted in cognitive and language gains at the first grade level but appeared to "fade out" by second or third grade. Gains were particularly noteworthy among blacks and among children attending Head Start in central cities and in the Southeast.

This study has been criticized for having a faulty design, leading to biased results. First of all, the comparison group selected may not have been as disadvantaged as those in Head Start. Thus, results showing little difference in the outcomes between the two groups may simply reflect the higher initial starting point of the comparison group rather than the lower achievement of the Head Start group. Consequently, the impact of Head start may be systematically underestimated.

Second, the design did not control for environmental circumstances--either at home or at school--leading to uncertainty as to whether short-lived child outcomes truly reflect effects of Head Start or reflect subsequent effects of the home and school environments. For example, some have argued that since low-income children compete for slots in compensatory programs, Head Start children (upon entering public school) are no longer among the most needy. Available slots in compensatory programs are then targeted to those disadvantaged children who did not have the benefit of participating in Head Start. Thus, it may not be that Head Start gains "fade-out"; it may be that the former comparison group catches up.

In addition, this study did not evaluate the health and nutritional components of Head Start programs, nor did it assess the effects of parental involvement in Head Start on children's cognitive and socioemotional outcomes or on parent outcomes. Yet these components are as integral to the Head Start program as is the education component. Consequently, this study yields incomplete information on Head Start's effectiveness.

Head Start Synthesis Project, 1985

There has not been a major evaluation of Head Start since the Westinghouse study. However, there have been many small evaluations of selected aspects of various Head Star programs across the country. Despite the wide array of research designs, sample sizes, and outcome measures used, a "meta-analysis" was conducted in 1985 to glean basic findings from these diverse studies. This "Head Start Synthesis Project" reviewed over 210 reports of research on the effects of local Head Start programs and found that Head Start results in "significant, immediate gains in cognitive test scores, socioemotional test scores, and health status, (though) in the long-run, cognitive and socioemotional test scores of former Head Start students do not remain superior to those of disadvantaged children who did not attend Head Start." A small subset of these studies found former Head Start children were more likely to be promoted to the next grade and less likely to be assigned to special education classes (McKey, et al., 1985).

Unlike the Westinghouse study, this synthesis project did evaluate the health and nutrition components, and found that "Head Start is very successful in improving the general health of the children it serves, providing needed health care, and improving existing health care within communities (though) it appears less successful in its health education efforts and in its efforts to influence better home health practices."

In addition, this synthesis explored the effectiveness of parental involvement and found "evidence that parents who actively participate in the program have high levels of psychological well-being, improve their economic and social status, and have children with high levels of developmental achievement." Unfortunately, the study found that it is usually the same small number of parents who actively participate on an on-going basis. Furthermore, parent education programs designed to influence child-rearing practices in the home have had mixed results.

Finally, the synthesis study assessed the effect of Head Start on communities and found "Head Start has been associated with positive changes (which) usually result in increased and more comprehensive social and health services for the poor and in more responsive educational programs. Greater visibility and greater parent participation both increase the effectiveness of a local Head Start program in affecting community institutions." Based on these findings, this study recommended more extensive collaboration with the public schools to bridge the "discontinuity in educational approaches between Head Start and public schools."

This synthesis has been criticized, mainly for its use of meta-analysis. This technique of standardizing findings across many studies of various designs may have the tendency of minimizing the results found in specific studies. Because it may "over-aggregate" results, studies using meta-analysis may conceal more distilled findings. But since this is likely to have the effect of underestimating effects, Head Start children may be doing even better than these results suggest.

Path to the Future: Long-term Effects of Head Start In the Philadelphia School District

This 1987 study used ten years of data on almost 15,000 children who attended Philadelphia Follow Through programs in 33 schools in the city's school district. (The Follow Through program serves children from kindergarten through third grade.) Since enrollment in Head Start is not a prerequisite for enrollment in Follow Through, this data included children who had and had not attended Head Start.

Results indicate that Head Start children more often avoided serious school problems, were less frequently retained in grade from kindergarten through sixth grade, had better attendance rates and lower attrition than non-Head Start children. Head Start children performed slightly (but non-significantly) better on achievement tests than their non-Head Start peers up to third grade, but there was no difference on achievement test scores from third to sixth grade (Copple, Cline, & Smith, 1987).

The Philadelphia data is based on a large number of children and schools, collected over a long period of time, and represents typical Head Start programs more accurately than the samples in previous large-scale studies. Unfortunately, the outcomes measured related only to school success and not to any of the more comprehensive goals of the Head Start program. In addition, the Philadelphia data base is--in a statistical sense--a "noisy" one: varied record-keeping over the ten years of data collection and policy changes during this period increases the variability in the data. Increased variability tends to depress the real magnitude of an effect size. However, this suggests that any findings are likely to underestimate the true effects of Head start, and the detection of statistically significant effects should be take seriously.

Program Model Effects Study

This 1987 study compared the relative effects on parents and children of three different Head Start delivery models (center-based, home-based, and a combination of center and home-based) of a grantee in rural Pennsylvania. Outcomes were measured from the time children were enrolled in Head Start through kindergarten.

No difference was found in children's cognitive development across the three delivery models. Parents enrolled in the home-based model demonstrated greater gains in academic stimulation of their children; in the use of toys, games, and reading material; and in encouraging their children to learn. Home-based parents also demonstrated greater growth in knowledge of child development and parent empowerment. (University of Delaware, 1988).

Oregon Longitudinal Study

This 1989 study is assessing former Head Start children, currently in kindergarten through fifth grade, on academic performance and placement indices. Results show that former Head Start children performed at grade level on achievement tests in reading, math, and language use in grades 3-5. In addition, former Head Start children were more likely to be enrolled in "gifted and talented" programs than their non-Head Start peers. Finally, former Head Start children were less likely to be enrolled in special education classes than the district-wide average enrollment rates (Norris, 1989).

"A Look at What Head Start Children Are Learning Now"

A recent University of Arizona study on Head Start transition effects used the "Head Start Measures Battery" (consisting of language, math, nature and science, perception, reading, and social development scales) to assess Head Start children's developmental level and provided an experimental group of elementary school teachers with this information so they could plan individually-tailored activities. Comparison with a control group revealed that significantly greater gains on achievement and cognitive tests are made when public schools were provided with information that enabled them to provide developmentally-appropriate programs to children (Bergan, Feld, & Saladeczek, 1989).

Head Start Family Impact Study

This 1988 study evaluated the effect of two different Head Start interventions on family functioning. Eighty-one children enrolled in a Minnesota Head Start program were divided into two groups: one group remained enrolled in the current Head Start program, and the other group experienced an "enhanced" program with intensive parental involvement. Parents in the enhanced program were involved in learning activities with their children 2-3 times as often as parents in the "regular" Head Start program. Family functioning, as measured by the FACES III (Family Adaptability and Cohesion Evaluation Scales) instrument, was assessed for both treatment groups and a control group, made up of 21 children on the Head Start waiting list.

Results indicate that family dysfunction scores diminished significantly for families in "enriched" programs, and diminished somewhat for families in the regular Head Start program, compared to control group families. In addition, mothers in both Head Start groups were more likely to assess their child as competent --and were more competent themselves--compared to control group mothers (Leik and Chalkey, 1989).

Other Head Start Studies

Finally, a number of studies (Cawly and Goodstein, 1966; Bee, 1981; Shipman, 1976; Hebbler, 1981; Copple, 1987) revealed that Head Start children have developed the necessary social competencies to adapt more readily to their school environment and experience more "real life" academic success than their no-treatment peers.

Other Preschool Program Evaluations

Findings from other evaluations of preschool programs suggest the potential effectiveness of high quality early childhood programs.

Consortium for Longitudinal Studies, 1982

The Consortium was formed to combine the projects of 12 researchers who had conducted preschool programs in the 1960s. A follow-up study using common measures was conducted in 1976-1977. Program design and curricula varied; included were both home-based and center-based programs. The 3,700 program participants were poor and minority families, with a heavy representation of black families. Children ranged from age three months to five years at the time of program entry. By the time of the Consortium's follow-up, they ranged in age from 9 years to 19 years.

Researchers found positive results from preschool attendance. Children who participated in preschool programs were more likely to succeed in school as measured by staying on grade level with their peers and avoiding inappropriate placement in special education classes. Lasting gains in academic achievement tests (particularly mathematics) were noted. However, intelligence gains (as measured by IQ tests) faded three years after the program had ended. In addition to these cognitive impacts, there were signs of favorable impacts on the child's self-concept, parental aspirations for the child's education, and on a family's achievement orientation. Parent participation was extremely high (reviewed in Collins, 1989).

These studies each employed strong research designs, many with random assignment. Little attrition was experienced in the follow-up study. However, one should be cautioned not to attribute the successes noted in this study to Head Start programs. First of all, these preschools had access to professional resources normally available only in university research projects and laboratory schools. In addition, teaching staff experienced intensive staff development and on-going support. Therefore, these programs do not demonstrate effectiveness of typical Head Start programs. However, results from this study are indicative of the benefits attainable from high quality, resource-intensive early intervention programs.

High/Scope Perry Preschool

One of the Consortium participants was the High/Scope Perry Preschool in Ypsilanti, Michigan. The program began in the early 1960s and involved 123 black preschoolers from low socioeconomic backgrounds. Teachers visited each child's home for two hours every week, and a majority of the program children attended the half-day program for two years. The researchers at High/Scope continued to follow program and control children to age 19 and looked at a broader set of outcomes, including teenage pregnancies, arrests, and unemployment.

The study found lasting benefits of preschool education in improving cognitive performance during early childhood; in improving scholastic achievement during the school years; in decreasing delinquency and crimes and in increasing high school graduation rates and frequency of enrollment in postsecondary programs and employment (Berrueta-Clement, et al., 1984).

Like the other Consortia studies, this project did not evaluate Head Start programs; therefore, at best, it is suggestive of possible outcomes of high-quality and resource-intensive early childhood education programs.¹ On the other hand, this study has some design problems that have implications for the validity of the findings.

Many criticize this study for its small sample size: only 58 children were in the "treatment" group (65 were in the control group). Small sample size limits one's ability to "slice" the data in many ways and still maintain enough variability between the treatment and control groups to be able to detect differences in outcomes.

Furthermore, although this study employed random assignment of children, subsequent reassignments were made if experimentals had employed mothers or no access to transportation.² This reassignment effectively destroys the true experimental design by introducing systematic biases. It is unclear whether these biases serve to under- or over-estimate the study's findings.

Finally, the evaluation design did not take into account the subsequent life events and environmental circumstances--at home, school, or the community--that are bound to affect the likelihood of school success, criminal activity, pregnancy, and welfare receipt. Generally, the ability to attribute long-term outcomes, which may (at best) only be distally related to a short-term intervention many years earlier, is suspect.

Program Characteristics Associated With Positive Outcomes

Hours of Contact

A number of studies indicate that longer hours of Head Start services produce more and longer-lasting gains for children (McKey, et. al., 1985; Copple, Cline, & Smith, 1987). In addition, the 1969 Westinghouse Study showed that summer programs (approximately 240 hours of services per year per child) provided insufficient contact with children and families to result in long-term benefits. Head Start has most recently responded to this research by outlining in regulation minimum annual hours of operations.

Group Sizes

Many studies attest to the fact that smaller class sizes effectively lead to more teacher attention per child, positively affecting children's developmental outcomes (National Day Care Study, 1980). The average group size in Head Start is 18 children.

The National Association for the Education of Young Children (NAEYC) recommends a group size of 14-20 for 3-year-olds and 1620 for four-year-olds. The Child Welfare League of America (CWLA) recommends maximum group sizes of 14 for three-year-olds and 16 for four-year-olds. The Federal Interagency Day Care Requirements (FIDCR) (which were never implemented) recommended maximums of 15 and 20, respectively.

Child-Staff Ratios

Related to group size, smaller child-staff ratios also enable teachers to spend more time with each child, positively affecting their developmental outcomes. Head Start classrooms contain two paid staff--a head teacher, a teacher's aide--and often also a volunteer for an average child-staff ratio of 6:1 to 9:1.

NAEYC recommends maximum child-staff ratios of 7:1 to 10:1 for three-year-olds and 8:1 to 10:1 for four-year-olds. The Early Childhood Environment Rating Scales (ECERs) recommend ratios of 7:1 for three-year-olds and 8:1 for four-year-olds. CWLA recommends 5:1 for three-year-olds and 7:1 for four-year-olds. FIDCR recommended 4:1 to 5:1 for three-year-olds and 7:1 for four-year-olds.

Teacher Qualifications

The National Day Care Study (1980) found that children in programs with staff trained in early childhood education had better relationships with the teacher and showed significant gains in skills and knowledge. NAEYC recommends that caregivers be trained in early childhood education or child development, from professional development courses for early childhood teaching assistants to a bachelor's degree for early childhood teachers. CWLA recommends teachers have professional education and experience in early childhood education or child development and teaching experience. Likewise, FIDCR recommended staff be trained or have experience in child growth and development.

Head Start created the Child Development Associate (CDA) training program in 1985 to increase the supply of teachers who are qualified to work with young children. CDA training combines education in child development with practical experience in the classroom to prepare enrollees to teach young children. Currently, 80 percent of CDA-credentialed teachers work in Head Start.

Recognizing the importance of having well-qualified teachers, Head Start recently promulgated regulations that would require all head teachers to obtain at least a CDA credential before beginning teaching and encourages teachers' aides to seek the credentialing. Currently, 41 percent of classroom staff have degrees in early childhood education or a CDA.

Consistency of teaching staff is also related to positive outcomes for children; therefore, minimizing turnover of qualified staff is important to Head Start's effectiveness. Unfortunately, average turnover in Head Start programs is about 20 percent annually. Low teachers' salaries are a major contributing factor to this high turnover rate. A 1988 study of Head Start salaries reveals that Head Start teachers work about the same number of hours per year as public school kindergarten teachers but earn substantially less. Head Start teachers with a bachelor's degree in early childhood education earn, on average, 63 percent of the beginning salaries for public school kindergarten teachers ($11,518 versus $18,350, respectively). The 1987 Head Start Recruitment and Enrollment Study reported the median annual wage of a Head Start teacher was $5,682, with few teaching' staff receive any fringe benefits.³

III. WHAT RESEARCH IS NEEDED

Despite the diversity in evaluations of Head Start programs, most available research shows that a Head Start experience results in positive effects on cognitive and socioemotional development for children. However, there remain many questions about what gains in child and family outcomes Head Start can realistically expect to produce, how to maintain these gains over time, and what specific program features contribute to Head Start's effectiveness.

What outcomes can Head Start expect to affect? The original intent of Head Start was to improve poor children's learning and social skills, as well as their health and nutrition, so that they would be better equipped to learn in elementary school. Improving "school readiness" would better the chances for breaking the cycle of poverty.

Despite the intentionally comprehensive nature of Head Start, most evaluations of Head Start have concentrated on using educational performance as the measure of effectiveness, often ignoring other outcomes measures such as:

• health and nutrition benefits to children and families,
• increased self-esteem of children and families, and
• improved child-rearing practices.

Failure to measure the effects of Head Start on these dimensions prevents researchers from testing Head Start's full impact. (Though many practitioners confirm that improved health and social skills are among the more immediate benefits to Head Start children.) Moreover, this bias in evaluations severely skews the perception of Head Start's effectiveness and may effectively redirect the goal away from comprehensiveness toward only school-related measures of success.

Results from subsequent evaluations of preschool programs suggest that an intensive preschool experience is associated with a reduction in teen pregnancies, crime, and welfare receipt (see Perry Preschool). Early childhood education advocates have used these results to support their call for more investment in early childhood education, including Head Start. However, should a short-term intervention, of varied quality, at age three or four be expected to reduce teen pregnancies, crime, and welfare receipt before it is termed a success? Citing these results may amount to over-promising the benefits of Head Start, especially when affecting these distal outcomes is not the primary goal of Head Start. Longitudinal evaluations may indeed try to capture longer-term effects, but research designs must take into account the likely impact of subsequent events and environments in which previously-enrolled Head Start children find themselves.

How can gains achieved by Head Start children be maintained beyond the first few years of elementary school? There is little research on why cognitive effects of Head Start appear to "fade out" by second grade. Some researchers hypothesize that competition for scarce compensatory resources, coupled with the fact that Head Start children (upon entering elementary school) are no longer the most needy and therefore do not continue to receive compensatory services, results in non-Head Start children "catching up" to Head Start children.

Moreover, practitioners admit that teachers often teach to the average abilities of students in the classroom. Therefore, there is a natural "tendency toward the mean," where the gap between children's achievement levels narrows over time as "below average" students catch up and "above average" students plateau. Since research indicates that Head Start children begin elementary school more advanced than their non-Head Start peers, apparent "fade out" of effects may really be the convergence of students' achievement levels.

Finally, gains achieved by Head Start can only be maintained if there is an effective transition of Head Start children and families to schools. The term "transition" has been defined as the strategies and procedures that are planned and implemented to ensure the effective placement and adjustment of the child as s/he moves from Head Start into kindergarten or elementary school. A 1974 survey of 144 Head Start grantees revealed that:

• 70 percent of the grantees implemented a variety of transition activities;
• Most Head Start programs provided parents with information about the school their child would be attending;
• About 50 percent of programs arranged for visits to the schools or meetings with the kindergarten teacher; and
• Generally, parents rated Head Start programs highly on preparing them for transition, but some wanted more information, more academic preparation, and more visits to the schools.

Effective transition is important for the continuity and progress of the children's and families' development. But it is important to recognize that Head Start (as a Federal program to local grantees) and the public schools (state- and locally-administered) have no common administrative linkage, making efforts to coordinate transition activities inherently difficult. Therefore, a key question is: What mechanisms are appropriate to facilitate effective transition?

Follow Through

The Follow Through program--administered by the Department of Education--was initiated in 1967 to provide comprehensive educational, health, nutritional, social, and other services primarily to children from low-income families in kindergarten and the primary grades who were previously enrolled in Head Start or other preschool programs of a compensatory nature. The intent was to aid in the continued development of these children to their full potential.⁴ Studies of Follow Through indicate this programs effectiveness for maintaining gains. one study revealed that Follow Through students demonstrated a significant superiority in three of five academic measures by the end of fifth grade (Zigler, 1976).

Project Developmental Continuity

In 1974, ACYF funded 15 Head Start grantees to participate in "Project Developmental Continuity." The purpose of the program was to stimulate the development of programs that coordinated educational and other services to children and families from Head Start through children's third grade. By specifying institutional features of the program,⁵ ACYF hoped to cause systematic changes in the behaviors of teachers and parents toward children, changes that would increase developmental continuity and enhance children's social competence.

This study yielded very little evidence that local PDC programs enhanced children's social competence or that PDC affected the behavior of former Head Start parents. PDC and non-PDC teachers and classrooms did differ; however, PDC classrooms had no detectable influence on measured child outcomes. Finally, PDC schools were found to differ from non-PDC schools in only three of the seven required institutional features (administration, parent involvement, and development support services).

The degree to which this evaluation adequately assessed the impact of the PDC is debatable for many reasons. First, the research design was not an experimental one; pre-treatment differences existed between PDC and non-PDC children that were not possible to control for statistically in the analysis. Moreover, post-treatment variables (such as the non-PDC schools implementing some of the seven institutional, changes, and implementation of P.L.94-142--Education of the Handicapped Act) may have further blurred any difference between the PDC and non-PDC children. In addition, small sample sizes and attrition may have damaged whatever rigorous design had previously existed. Finally, measures of children's social competence were inadequate (and sometimes non-existent) in some programs.

Current Initiatives

The Head Start Bureau is launching new initiatives addressing effective transition. They plan to convene a task force to explore ways of encouraging and supporting Head Start programs in developing linkages with community agencies and school systems. The Bureau is also funding 10 grants to states to hire one fulltime staff person in the Governor's office to coordinate state preschool, Head Start, and transition-to-elementary-school initiatives.

While coordination at the federal and state levels is desirable, the transition of children and families from Head Start to elementary schools occurs, by nature, at the local level. In an effort to encourage grantees to plan for transition, ACYF funded 15 grantees in 1986 to develop and test different ways of implementing transition programs.

An evaluation of these practices indicates that almost all the grantees engaged in joint planning with the school system and its staff. A majority of the grantees developed written agreements with the school system, most grantees provided some of the children's records to the schools. In some programs, parents were given information on how to deal with the school their children would be attending, were scheduled for visits to the public school to meet the kindergarten teacher, and were given information on how to help their children practice skills needed for elementary school and were given suggestions for summer activities.

In addition, current research indicates that transferring information on the developmental levels of Head Start children is also an effective transition practice that facilitates child-centered teaching and reduces inappropriate placements in special education classes by 20:1 (Bergan, Feld, and Saladeczek, 1987).

Head Start needs to continue to explore and evaluate new approaches to effective transitioning, including the approaches used by currently-operating Follow Through programs and including employing a mentoring or case management approach, to provide the continuity and the personal attention needed to assure effective transitions.

What are the additional benefits of two versus one year in Head Start? There is little evidence on the additional benefits to the child and family of two--as opposed to only one--year of Head Start. In the Perry Preschool Program, "one year of preschool produced the same effects as two." Research on this subject could explore potentially varying effects of Head Start for different populations of Head Start children. Results could inform programs' decision on when to enroll three-year-olds for whom an additional year of Head Start would prove beneficial.

What program components are effective? There is a lack of information on which programmatic features of Head Start (curriculum, home- vs. center-based program models, level and type of parental participation, involvement of community readers and organizations) contribute to successes, for which children, and under what conditions. Of particular importance is the relative effectiveness of the various forms of parental involvement in leading to positive child and family outcomes. Research in this area will allow encouragement of the most effective program designs for certain populations of Head Start children.

IV. CURRENT AND PLANNED RESEARCH EFFORTS

Head Start Evaluation Design Study

Recognizing a deficiency in information, the Head Start Bureau is exploring alternatives for designing a comprehensive, (possibly longitudinal) evaluation to answer the following questions:

• "Which Head Start practices or factors maximize benefits to which children and families?" and
• "How can Head Start effects be maintained?"

Head Start has assembled an advisory group with expertise in child development, early childhood education, psychology, pediatrics, Head Start, social policy, the testing of young children, and in designing evaluations of programs for disadvantaged children. The advisory group will recommend to ACYF a design for the study or set of studies to explore the relative short-term and long-term effects of various program features on an array of child and family outcomes. The recommended design will be presented in a final report by the fall of 1990.

Long-Term Benefits and Cost-Effectiveness of Head Start

High/Scope Educational-Research Foundation, with funding from OHDS, is conducting a follow-up study in three sites on former Head Start children who are now 20-22 years old. High/Scope plans to compare these former Head Start children to a control group for educational attainment, employment, criminal records, and other long-term outcomes. A cost-benefit analysis (similar to that employed in the Perry Preschool Project) will be conducted, and the final report delivered to OHDS by June 1990.

V. ADDITIONAL TOPICS RELATED TO HEAD START'S EFFECTIVENESS

Factors that determine the effectiveness of the Head Start program need to be assessed on a continuing basis in the determination of future policy and funding allocations. These factors include:

• Children served - What groups of children should be targeted for participation in Head Start? How many and which children should be served by Head Start as opposed to other pre-school programs? What degree and types of coordination is most appropriate between Head Start and other preschool programs, including State programs, in ensuring all disadvantaged children have access to preschool?

• Program quality - What is the appropriate policy and funding allocations for staff salaries and benefits, training and technical assistance, etc.?

• Program oversight - What is the appropriate level of monitoring and oversight necessary to ensure program quality, and how can we ensure availability of resources sufficient to meet these oversight needs?

Many of these factors are currently being addressed through the implementation of the $151 million appropriation increase in FY 1990. Of this increase, approximately $100 million is being used to expand enrollment by 37,500 children. Current Head Start policy is to serve eligible children for at least one year in the year before he or she enters public kindergarten. For most children, this will be when they are four years old. Thus, grantees have been encouraged to expand enrollment to four-year-olds currently unserved. The remainder of the FY 1990 increase will be used to improve program quality: $49 million will increase staff salaries by 5.4% on average, and $2 million will be used to improve training and technical efforts.

In addition, the FY 1991 Budget and Legislative program submitted to the Congress in January 1990 further implements policies to increase Head Start's effectiveness. The Administration has requested $500 million to increase enrollment by up to 180,000 children. Again, this increase will be targeted to reaching unserved children primarily in the year before they enter public school. In FY 1991, $90 million of the amount available to the Secretary to conduct discretionary activities will be used to encourage States to provide funds to further expand enrollment in Head Start programs. Finally, the Administration's Head Start Reauthorization bill includes provisions to ensure timely reviews of grantees, and improved coordination with other programs serving disadvantaged preschool-aged children.

Beyond 1991, there are several specific realities that need to be considered in determining policy and funding strategies to make Head Start as effective as possible.

Coordination with State Preschool Programs

States are becoming more involved in preschool: in FY 1988, 28 states had initiatives in early childhood education, 23 of which targeted disadvantaged or at-risk children between the ages of three and five. Nine states supplement funding of Head Start to increase the number of eligible children able to be served.

Given increased state activity in this area, Head Start expansion strategies should not ignore this reality. In an effort to provide a Head Start-like experience to all Head Start-eligible four-year-olds, strategies should explore:

• What existing state programs already offer the comprehensive services required by Head Start? How many Head Start-eligible children do they already serve?
• What opportunities exist to expand remaining state preschool programs to offer the full range of Head Start services? How many Head Start-eligible children could these programs serve?

Coordination with the Family Support Act

Drafters of the Family Support Act recognized that moving families to self-sufficiency requires appropriate support services--like child care--in addition to training, education, and employment services. They acknowledge that the families participating in the JOBS program are also likely to be eligible for, and could benefit from, Head Start. Accordingly, the Family Support Act requires that agencies implementing the Act coordinate with Head Start programs. Similarly, a current HDS legislative proposal mandates that Head Start coordinate with agencies implementing the Family Support Act.

Examples of coordination include referring Head Start families to the JOBS program, referring JOBS participants to the Head Start program, and exploring alternatives for "wrapping around" the standard half-day of Head Start with full-day child care services to meet the employment needs of JOBS participants. The new realities of the JOBS program and increasing maternal employment will also require experimenting with, evaluating, and encouraging new and innovative forms of parental involvement.

Allocating Expansion Funds for Quality, Enrollment, and Management

While the name "Head Start" connotes a high quality preschool program for disadvantaged children, the fact is that there is a range of quality in the 1900 Head Start programs. For example, while the average group size is well within recommended limits (18), group sizes range from 12 to 22. Moreover, while 41 percent of classroom staff have degrees in early childhood education or a CDA, this implies that a majority of classroom staff do not have even minimal training in child development. Programs with well-qualified staff, with lower turnover rates, with appropriate child-staff ratios and group sizes, and that actively involve parents and the community are likely to be more effective than programs that do not.

Simply increasing enrollment of current programs will not ensure that all children are receiving a quality experience that will m make them ready to enter school. Therefore, considering what we currently know about what constitutes an effective Head Start program, how should funding increases for Head Start beyond FY 1991 be allocated among the objectives of increasing enrollment, improving quality, and ensuring the effective management of Head Start programs?

• Strategies for increasing enrollment:
  • developing systematic, standard procedures for targeting, recruiting, and enrolling children most in need of services;

  • developing guidelines or policy concerning when to enroll only four-year-olds and when enrolling three-year-olds is allowable or even advisable.

• Strategies for improving quality:
  • increasing wages, salaries, and benefits of Head Start staff;

  • improving training and technical assistance;

• Ensuring effective management of this ever-growing program requires:
  • having sufficient resources to monitor grantees' programs once every three years (as proposed in the President's Head Start bill).

REFERENCES

Abt Associates, Inc. National Day Care Study, 1980.

Bee, C. K. "A Longitudinal Study to Determine if Head Start has Lasting Effects on School Achievement," University of South Dakota, 1981.

Bergan, Feld, & Saladeczek. "A Look at What Head Start Children are Learning Now," University of Arizona, 1989.

Berrueta-Clement, J.R., et al., Changed Lives: The Effects of the Perry Preschool Program on Youths Through Acre 19, High/Scope Educational Research Foundation, 1984.

Cawly and Goodstein. "An Assessment of Intelligence, Psycholinguistic Abilities, and Learning Aptitudes Among Preschool Children," Connecticut University, 1966.

Collins, R. "Head Start Research and Evaluation: Background and overview," Vienna, VA: Collins Management Consulting, 1989.

Copple, Cline, & Smith. "Path to the Future: Long-Term Effects of Head Start in the Philadelphia School District," 1987.

CSR, Inc., "The Transition of Head Start Children in to Public School (Final Report)," November, 1987.

Hebbler, K. "An Analysis of the Effectiveness of Head Start and the performance of Low-Income Populations," Montgomery County Public Schools, 1988.

Kresh, E. (Unpublished findings from the Head Start Salaries Study, 1988).

Leik, R.K. and Chalkey, M. A. "The Head Start Family Impact Project," University of Minnesota, 1989.

Love, J.M. and Grover, J. "Study of Head Start Recruitment and Enrollment: Final Report," RMC Research Corporation, February, 1987.

McKey, et al., "The Impact of Head Start on Children, Families, and Communities: Final Report of the Head Start Evaluation, Synthesis, and Utilization Project (Executive Summary)." CSR, Inc., Washington, D.C. June, 1985.

Norris, D. "A Jump on Academic Success," Albina Ministerial Alliance Head Start program, Portland, Oregon, 1989.

Shipman, et al. "Notable Early Characteristics of High- and LowAchieving Low SES Children," ETS, Princeton, NJ, 1976.

University of Delaware College of Human resources. "Individual And Family Studies, 1985-1987: Head Start Delivery Models Study," 1988.

Zigler, E. "Summary of Findings from Longitudinal Evaluation of Intervention Programs (mimeographed)," Yale Research Group, Yale University, New Haven, CT, 1976.

NOTES

1. Most notably, child-teaching staff ratios were limited to 5:1 or 6:1 (compared to 6:1 to 9:1 in Head Start) and average annual program costs were $4813 per child, or almost $11 per child-hour (compared to Head Start's average annual $2660 per child, or $3.50 per child-hour).

2. Because home visits were an important feature of this program, mothers of experimental group children who were employed were reassigned to the control group, and a non-employed mother from the control group was reassigned to the experimental group to facilitate home visits.

3. While lower than their public school counterparts, Head Start teachers' salaries are somewhat higher (and turnover somewhat lower) than their child care counterparts. The average wage of child care providers is about $5 per hour (compared to about $9 for Head Start teachers), and national turnover of child care staff is about 41 percent.

4. Section 662(a) and 662(c) of the Follow Through Act (Subchapter C of P.L.97-35: the Omnibus Budget Reconciliation Act of 1981).

5. The seven components were: (1) Administration--special function staff; (2) Education--continuous, developmentally-appropriate curricula from Head Start through third grade; (3) Bilingual/multicultural education; (4) Services for handicapped children; (5) Parent involvement; (6) Developmental support services--coordinated nutritional, medical, dental, mental health, and social services; and (7) Training--on-going training of teachers and parents relating to PDC components.