ACF OPRE: Early Head Start Research and Evaluation Project (EHSRE)

I. INTRODUCTION: BACKGROUND AND CONTEXT FOR THE EVALUATION

Early Head Start has become a major national initiative in the six years since its beginning. Following the Administration on Children, Youth and Families’ (ACYF) funding of 68 grantees in fall 1995, the program has grown to 664 programs that in 2002 serve more than 55,000 low-income families with infants and toddlers throughout the country. With an increasing share of the Head Start budget, up to 10% in 2002, Early Head Start is an ambitious effort in which ACYF is responding to the “quiet crisis” facing American infants and toddlers, as identified by the Carnegie Corporation of New York in its 1994 Starting Points report.1 The final report of the Early Head Start Research and Evaluation project traces the services that Early Head Start families in 17 programs received over approximately 26 months in the program, describes the differences that the programs made in the services families received, and examines their impacts on the children and families through the children’s third birthdays. This report builds on the Early Head Start implementation study, which is fully described in two reports: Leading the Way (Administration on Children, Youth and Families 1999a, 1999b, 2000a, and 2000b) and Pathways to Quality (Administration on Children, Youth and Families 2002).

This chapter begins with a synopsis of the findings and then reviews the history of the program and the policy, programmatic, and research context for both the program and its evaluation. We summarize the questions the evaluation addresses, the conceptual framework guiding this research, and the general hypotheses that underlie the analyses. We then describe the 17 research programs, their families, and their communities, and follow with a description of the design, sample, and analytic approaches taken in the study.

Subsequent chapters describe:

The evaluation methodology and analytic approaches (Chapter II)
The services received by Early Head Start mothers, fathers, and children, and the difference the programs have made in the rates, duration, and intensity of their participation in a wide range of services during the initial period following program enrollment (Chapters III and IV)
The programs’ influence on children’s development, parenting, and family development when the children were 3 years of age (Chapter V)
The differential impacts of programs offering different service approaches and achieving different levels of implementation result in (Chapter VI)
Variations in impacts among key subgroups of children and families (Chapter VII)
Implications of these findings for policy, practice, and research (Chapter VIII)

In text “boxes,” this report also incorporates findings related to the fathers of Early Head Start children and presents what we have learned about their involvement with the programs and with their children. Appendixes in Volume II describe aspects of the methodology in greater detail and provide supplementary tables of findings. In addition, findings and perspectives from local program and research partners are integrated throughout and highlighted in text “boxes.” Reports of the local research are presented in Volume III in greater depth.

A. OVERVIEW OF THE FINDINGS

Early Head Start programs had numerous consistent overall impacts on children, parents, and families when children were 3 years old. These findings in many ways continue the trends observed when children were 2 years old, as reported in the interim report, Building Their Futures (ACYF 2001). As we present the findings in subsequent chapters, we describe how they do—or do not, in some cases—replicate or continue the impacts at age 2. Highlights of these findings include the following:

The Early Head Start research programs substantially increased the services families received.
When children were 3 years old, the Early Head Start programs largely sustained the positive impacts on cognitive, language, and social-emotional development found at age 2. The program continued to have favorable impacts on a wide range of parenting outcomes as well. These include positive impacts on parental emotional support and support for language and learning and discipline practices. The programs also had important impacts on parents’ progress towards self-sufficiency.
Full implementation matters: programs that fully implemented key program performance standards had a stronger pattern of favorable impacts on child and parenting outcomes than those that did not reach full implementation.
All program approaches had positive impacts on child and parent outcomes, although mixed-approach programs had the strongest pattern of impacts.
Mixed-approach programs that were fully implemented early had a stronger pattern of impacts than those that became fully implemented later or did not reach full implementation, and home-based programs that were fully implemented had a stronger pattern of impacts than those that never became fully implemented during the evaluation period. There were too few center-based programs to conduct this analysis by implementation pattern.
Programs served families with diverse characteristics, and the programs were differentially effective for different demographic subgroups. Although patterns of impacts varied, Early Head Start programs improved some outcomes for nearly every subgroup in the study.
Patterns of program impacts varied by race/ethnicity. There was a strong pattern of impacts for African American families, a number of notable positive impacts among Hispanic families, but virtually no impacts on child and parent outcomes for white families.
Early Head Start programs improved child and parenting outcomes among some subgroups of difficult-to-serve families that have special policy relevance, including teenage mothers and parents who were at risk of depression at the time they enrolled.
Programs had positive impacts on several areas of fathering and on father-child interactions. Fathers and father figures from program families were more likely than those from control families to participate in program-related child development activities, such as home visits, parenting classes, and meetings for fathers.

The numerous Early Head Start impacts that span most important outcome areas at age 3, even though modest in size, represent a significant policy achievement, given the history of program evaluations demonstrating few positive impacts. Early Head Start programs have not produced impacts in every dimension of child development, parenting, and family functioning that they hoped to influence, however, and this report also describes areas in which programs could work to enhance their services. The differential impacts across subgroups of programs and families also have important implications for program improvement. Programs were particularly effective for some subgroups, while they are challenged to better serve families in other subgroups. We return to the details of these findings after reviewing the national program’s history, describing the research questions that the study addressed, summarizing the programs and their families and communities, and describing the evaluation’s design and methodology.

B. EARLY HEAD START, ITS HISTORY, AND ITS DEVELOPMENT AS A NATIONAL PROGRAM

Early Head Start programs are comprehensive, two-generation programs that focus on enhancing children’s development while strengthening families. Designed for low-income pregnant women and families with infants and toddlers up to age 3, the programs provide a wide range of services through multiple strategies. Services include child development services delivered in home visits, child care, case management, parenting education, health care and referrals, and family support. Early Head Start programs try to meet families’ and communities’ needs through one or more official program options: (1) home-based, (2) center-based, (3) combination (in which families receive both home visits and center experiences), and (4) locally designed. Because a program may offer multiple options, we characterized programs for research purposes according to the options they offer families. For the purposes of the research, programs were grouped according to three program approaches (home-based, center-based, and mixed-approach), which are described in Section D.

A number of key events and changes, both external to and within the Head Start/Early Head Start infrastructure, shaped the development of the programs during their first six years. Figure I.1 depicts the timing of these key events. We describe these and other events in the following sections.

1. The Role of Legislation and Advisory Committees

The federal Early Head Start program began with bipartisan support provided by the 1994 Head Start reauthorization. This legislation established the mandate for infant-toddler services within Head Start. The 1998 Head Start reauthorization propelled the program toward rapid expansion, which saw an increase from 68 programs in 1995, when the evaluation was getting underway, to 664 programs in spring 2002, serving some 55,000 children.

Leading up to these mandates, a comprehensive study of Head Start services by the Advisory Committee on Head Start Quality and Expansion called for developing a “new initiative for expanded Head Start supports to families with children under age three.” At the same time, the committee recommended actions to ensure that such services be of the highest quality and that new partnerships be forged to reduce fragmentation of services (U.S. Department of Health and Human Services [DHHS] 1993). In response to the 1994 reauthorizing legislation, the Secretary of DHHS appointed the Advisory Committee on Services for Families with Infants and Toddlers. It envisioned a two-generation program with intensive services beginning before birth and concentrating on enhancing development and supporting the family during the critical first three years of the child’s life (U.S. Department of Health and Human Services 1995). The Advisory Committee recommended that programs be designed to produce outcomes in four domains:

FIGURE I.1

KEY EVENTS IN THE HISTORY OF EARLY HEAD START

Jan. 1994			Advisory Committee on Head Start Quality and Expansion recommends serving families with children under 3
			Carnegie Starting Points report released
			Head Start reauthorized with mandate to serve infants and toddlers
			Advisory Committee on Services for families with Infants and Toddlers sets forth vision and names Early Head Start

Jan. 1995			First Early Head Start program announcement solicits first grant applications
			Federal Fatherhood Initiative formed
			Wave I: 68 new Early Head Start programs funded
			Oldest child in the research sample born

Jan. 1996			First Early Head Start programs began serving families, random assignment begins
			Welfare reform legislation enacted (PRWORA)
			Wave II: 75 new programs funded
			First round of research implementation study visits conducted
			Revised Head Start Program Performance Standards published for public comment

Jan. 1997			White House Conference on Early Childhood Development and Learning
			Wave III: 32 new EHS programs funded
			Second round of research site visits conducted

Jan. 1998			Revised Head Start Program Performance Standards take effect
			Monitoring visits to Wave I programs conducted
			Wave IV: 127 new EHS programs funded
			Youngest child in research sample born
			Wave V: 148 new EHS programs funded
			Head Start reauthorized by Coats Human Services Reauthorization Act
			Random assignment of research families concludes

Jan. 1999			Wave VI: 97 new programs funded
			Third round of research implementation visits conducted

Jan. 2000
Jan. 2001			Additional Early Head Start grantees funded, bringing total to 635
July 2000			National evaluation data collection concludes

Child development (including health and social, cognitive, and language development)
Family development (including parenting and relationships with children, the home environment and family functioning, family health, parent involvement, and economic self-sufficiency)
Staff development (including professional development and relationships with parents)
Community development (including enhanced child care quality, community collaboration, and integration of services to support families with young children)

The Advisory Committee also stressed continuous program improvement and recommended that both national and local research be conducted to inform the development of the new Early Head Start program. The committee specified that local programs conduct annual self-assessments and improve their services based on analysis of local data. Both the 1994 and 1998 Head Start reauthorizing legislation specified that an evaluation begin early to focus on learning about all the services being delivered to families with infants and toddlers and the impacts of services on children and families.

The evaluation reported here is the result of this early planning, as well as DHHS research and evaluation planning. In 1990, the Secretary’s Advisory Panel for the Head Start Evaluation Design Project (commonly known as the “blueprint” committee) concluded that it was important for evaluations to reject the generic question of “what works?” and move toward designs that would address questions on the theme of “what works for whom, and under what conditions?” In addition, the blueprint committee explicitly recommended that Head Start research be conducted through collaborative enterprises and have as one of its emphases providing findings that could be used by programs for their continuous improvement (U.S. Department of Health and Human Services 1990). All of these elements have been incorporated into the Early Head Start Research and Evaluation project from its very beginning.

2. The National Early Head Start Program

At the very outset of Early Head Start, ACYF created an infrastructure for supporting programs. This included the revised Head Start Program Performance Standards, an ongoing training and technical assistance (T&TA) system, and program monitoring. Early Head Start program guidelines also emphasized the importance of continuous program improvement, and built in research from the very beginning.

The Head Start Program Performance Standards, which have guided Head Start practice since the 1970s, were revised and published for comment in November 1996. The revised standards went into effect in January 1998, bringing Early Head Start programs under the Head Start standards umbrella. Between fall 1996 and January 1998, the Head Start Bureau worked with Early Head Start programs to clarify a number of the new elements in the standards. Within ACYF, the Head Start Bureau, under the leadership of the late Helen Taylor, emphasized the centrality of children’s development and stressed program quality through adherence to the standards. The bureau worked with both Head Start and Early Head Start programs to meet the standards, and some programs that were not able to improve have closed.

In 1995, ACYF created the Early Head Start National Resource Center (NRC) to provide ongoing support, training, and technical assistance to all waves of Early Head Start programs. Operated under contract by the ZERO TO THREE national organization, the NRC provided a range of services:

Week-long training in infant care (“intensives”) and annual institutes for all Head Start programs serving families with infants and toddlers
Provision of a cadre of infant-toddler experts for (1) working with ACYF regional offices and Indian and Migrant program branches, and (2) conducting one-on-one consultations
Coordination with ACYF’s regional training centers, the Head Start Quality Improvement Centers (HSQICs) and Disabilities Services Quality Improvement Centers (DSQICs)

The 1998 Head Start reauthorization included funding for a leadership position for Early Head Start programs within the Head Start Bureau, supporting the mandated expansion of Early Head Start and monitoring to ensure program quality. Through comprehensive on-site visits, monitoring teams review programs for standards compliance every three years.

3. The Program’s Policy Context

During the initial period of Early Head Start’s implementation, significant national, state, and local changes were occurring, potentially affecting the approaches taken by Early Head Start programs, the way families responded, and how programs and communities interacted. The increasing focus on the importance of early development (including brain development) attracted the attention and support of policymakers, program sponsors, and community members for Early Head Start services. Just at the time that Early Head Start began serving families, the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) enacted major reforms to the nation’s system for providing income support to low-income families. This caused some programs to adjust their service delivery plans to meet changing family needs. Because some states no longer exempted mothers of infants from work requirements, some parents became more receptive to employment-related services (including child care) and may have been less available to participate in some program activities. It became more challenging for programs to provide services through home visits.

In some states, changes associated with PRWORA have made it easier for families to obtain child care subsidies and have spurred states to improve and expand child care. Several states where Early Head Start research programs are located have increased funding for child care, aided centers seeking accreditation, or facilitated quality improvements for infant-toddler care. The expansion of prekindergarten programs in some states may have created opportunities for children’s transition to other programs when they leave Early Head Start, while new prekindergarten programs often compete for the same well-trained staff that Early Head Start programs need.

The federal Fatherhood Initiative has heightened attention to the role of fathers in a wide range of federal programs and has increased Early Head Start programs’ efforts to draw men into their program activities and into the lives of Early Head Start children. In addition, programs may have responded to PRWORA’s increased emphasis on establishing paternity and enforcing child support.

A strong economy with low unemployment rates throughout the period of the early development of Early Head Start programs probably helped them meet the many needs of their low-income families. While some of the families were eligible for health care assistance through the Children’s Health Insurance Program (CHIP), most were served by Medicaid. With CHIP, some states with Early Head Start programs have moved far in providing health services for all children.

4. The Research Context for the Early Head Start Program and Its Evaluation

Over the past decade, findings from a number of program evaluations have emerged that have a direct bearing on the Early Head Start evaluation. Some findings—particularly those from the Comprehensive Child Development Program (CCDP) and the Packard Foundation’s review of home-visiting programs—identified many of the challenges inherent in trying to make a difference for infants and toddlers in low-income families. The CCDP experience highlighted the importance of focusing program services on child development, while the home-visiting literature revealed the importance of understanding—and measuring—the implementation and intensity of services. These lessons influenced both the guidance that ACYF has provided to Early Head Start programs over the past six years and the design of this evaluation.

a. Brief Review of Evaluations of Other Infant-Toddler Programs

A number of evaluations of two-generation programs serving low-income families with infants and toddlers have been conducted over the last quarter century. Program effects have often appeared weak, but the findings are difficult to interpret because of the great diversity in program approaches, research methodologies and populations served across studies. Programs have varied in (1) the duration and intensity of services, (2) the timing of services, (3) their status as home- or center-based (or both), (4) the duration and intensity of the parenting component, (5) the extent of reliance on case management, (6) the nature of self-sufficiency (adult education and job training) components and (7) populations served. Many intervention programs have begun by focusing on a single population group or within a single community context. The research has also been variable, with differences in designs, domains assessed, timing of assessments, degree of information on program implementation, and extent of information on services received by control group families. Findings from seven major studies, or series of studies, are summarized here.

The Child and Family Resource Program was a comprehensive, two-generation demonstration program for families with infants and toddlers. The program produced significant effects on a number of parent outcomes after three years (employment or job training, coping skills, sense of control) and on parent-child teaching skills, but did not significantly affect children’s cognitive or social development (Nauta and Travers 1982).

Randomized studies of three Parent Child Development Centers (PCDCs) focused on mother-child interactions and infant/toddler cognitive development. Dokecki, Hargrave, and Sandler (1983) found impacts on positive maternal behaviors at two sites and significantly higher Stanford Binet scores for PCDC children at two sites.

Between 1972 and 1977, the Carolina Abecedarian Project enrolled 120 “high-risk” African American families in four cohorts. From these, 111 children were randomly assigned to the program, which included full-time child care beginning in the first three months of life, or to a control group. Families and children continued receiving services until age 5. The program, which also provided social supports for families, was highly successful in improving children’s cognitive development relative to the control group, with significant differences at 18, 24, and 36 months of age, and with an effect size of more than 1 standard deviation at 36 months (Campbell and Ramey 1994; and Ramey and Campbell 1991). The largest effects were found for children with the most extreme environmental risks. No effects were found on the families’ home environments. The intervention impacts appeared to be smaller when control group children enrolled in community child care (Guralnick 2000). Follow-up studies showed that program effects persisted at every assessment point through 16 to 20 years of age.

Olds’ Nurse Home Visitation Program is a model, designed some 20 years ago, in which nurses visit first-time mothers, beginning during pregnancy and continuing until the children are 2 years old, “to improve pregnancy outcomes, promote children’s health and development, and strengthen families’ economic self-sufficiency” (Olds et al. 1999). Results of two randomized trials show reduced rates of childhood injuries and ingestions (events perhaps associated with child abuse and neglect). For the mothers in one site, there were long-term reductions in child abuse and neglect, reductions in subsequent pregnancies, increased economic self-sufficiency, and avoidance of substance abuse and criminal behavior. At age 15, the children had fewer arrests, convictions, and other negative outcomes. However, “the program produced few effects on children’s development or on birth outcomes,” and the other benefits were found for the neediest families rather than the broader population (Olds et al. 1999). The long-term effects of the program were documented with a white, semi-rural sample of women in New York State. A subsequent trial of the program with a cohort of African American women in a city in Tennessee showed a smaller short-term effect and a somewhat smaller 3-year follow-up effect of the program than demonstrated in the white, rural sample (Kitzman et al. 2000). In the HV2000 project, Olds et al. (2001) found that children of mothers visited by nurses (but not by paraprofessionals) scored higher on the Bayley MDI at 24 months and were less likely to have language delays at 21 months than the control group.

Project CARE tested the effectiveness of home-based parent education and social services with and without full-time, center-based child care. At 2 years of age, differences in language and cognitive development significantly favored the group that had received child care combined with family education, and these differences continued to 4 years of age (although somewhat lessened) (Wasik, Ramey, Bryant, and Sparling 1990). Project CARE compared two treatments (child care plus family support, family support only) with a no-services control group. The group with child care plus family support performed significantly better than both the other groups (Wasik et al. 1990). This study was conducted with an African American sample.

The Infant Health and Development Program (IHDP) combined home visiting, center-based education, and family services to low-birthweight premature infants and their families during the first three years of life. At age 3, the program group scored significantly higher on the Stanford Binet and lower in behavior problems. The heavier low-birthweight infants benefited more at ages 2 and 3 than did the very low-birthweight children (Brooks-Gunn, Klebanov, Liaw, and Spiker 1993). Effects were sustained through age 8 for the heavier low-birthweight children (McCarton et al. 1997).

The Comprehensive Child Development Program (CCDP) was implemented in 24 highly diverse sites beginning in 1989 and 1990. Programs featured intensive social services and parent education, although direct child development services and program-sponsored child care were far less intensive than in the IHDP and Abecedarian programs. When children were 2 years old, the national evaluation (conducted in 21 of the sites) found that CCDPs significantly improved (1) mothers’ parenting skills and attitudes (for example, greater sensitivity to cues given by children in parent-child interactions and more appropriate responding to signals of distress), (2) parents’ economic self-sufficiency, and (3) children’s cognitive development (Bayley Scales of Infant Development) and social behavior (cooperation and following rules). (Language development at age 2 was not measured.) These effects largely disappeared by age 3 and were absent at age 5. At one site, however, significant and moderately large positive impacts were found at age 5 on children’s cognitive development, parenting skills, and several self-sufficiency outcomes (St. Pierre, Layzer, Goodson, and Bernstein 1997).

In a secondary analysis of CCDP’s 2- to 5-year impact data, Brooks-Gunn, Burchinal, and Lopez (2000) found that when sites were divided into two equal-size subgroups with more- and less-intensive parenting education (based on the average number of home visits families at each site received), the subgroup of programs with more-intense parenting education showed three important significant impacts relative to the control groups at those sites: (1) higher Bayley scores at age 2, (2) higher Kaufman Assessment Battery for Children (K-ABC) Achievement Scale scores at ages 3 to 5, and (3) higher Peabody Picture Vocabulary Test-R scores at ages 3 to 5. No impacts were found in the subgroup of sites where programs had less-intense parenting education.

Comparisons of the effects of home visiting and center-based programs are difficult to make. In a careful review, however, Benasich, Brooks-Gunn, and Clewell (1992) examined 27 studies and discovered that 90 percent of the center-based programs (compared with 64 percent of the home-based programs) produced immediate impacts on cognitive outcomes.

b. Building a Knowledge Base for Early Head Start

When they recommended Head Start services for infants and toddlers, the Head Start Quality and Expansion Panel and the Advisory Committee on Services for Families with Infants and Toddlers drew upon evidence of effectiveness in the existing research literature (including some of the findings cited here). The Advisory Committee on Services to Families with Infants and Toddlers consolidated knowledge from the research literature and from practice into nine principles to guide Early Head Start programs: (1) high quality; (2) prevention and promotion; (3) positive relationships and continuity; (4) parent involvement; (5) inclusion; (6) culture; (7) comprehensiveness, flexibility, responsiveness, and intensity; (8) transition; and (9) collaboration. These principles, along with the revised Head Start Program Performance Standards, set the stage for quality as they guided programs to implement specific practices (for example, low child-teacher ratios in relation to high quality).

Head Start advisory committees have called for research to understand the conditions under which programs are successful (and for whom programs can be more effective) and to promote continuous program improvement. The Early Head Start Research and Evaluation project, therefore, represents not only an evaluation of the initial years of the national Early Head Start program but an important step in expanding the Early Head Start knowledge base in very systematic ways. It attempts to do so by building in a number of features in response to the challenges of the new standards, guidelines, and principles and with the goal of overcoming shortcomings of previous studies. These features include:

A comprehensive implementation study to provide data on the services specified in the revised Head Start Program Performance Standards that Early Head Start programs delivered
Collection of extensive data on the services individual families received at specified intervals following random assignment, while also carefully and thoroughly documenting services received by control group families along the same dimensions and at the same intervals as the program families (see Chapter IV)
Documenting the overall impacts of Early Head Start on children and families (see Chapter V) and conducting analyses that take participation rates into account in testing for program impacts
Conducting subgroup analyses to examine the extent to which different program approaches have different kinds of effects on Early Head Start’s children and families (as described in Chapter VI)
Conducting subgroup analyses to examine the relationship between levels of program implementation and the impacts achieved (Chapter VI)
Conducting subgroup analyses to learn how the effectiveness of Early Head Start may differ according to the characteristics of the families being served (Chapter VII)
Collecting data directly from Early Head Start and control group fathers to learn more about the role of fathers and father figures in the lives of programs and families (highlighted in boxes in Chapters IV, V, and VII.)
Incorporating local research, as well as other local documentation (including from program staff), to supplement the cross-site national data collection and analysis (highlighted in boxes throughout this volume, with more-detailed reports in Volume III)

In addition, a longitudinal follow-up study is currently underway, as the first Early Head Start “graduates” began preschool in fall 2000.

C. RESEARCH QUESTIONS ADDRESSED IN THE EARLY HEAD START IMPACT STUDY

1. Central Questions of the Study

The national evaluation has two overarching goals: (1) understanding the extent to which the Early Head Start intervention can be effective for infants and toddlers and their low-income families, and (2) understanding what kinds of programs and services can be effective for children and families with different characteristics living in varying circumstances and served by programs with varying approaches. The study was designed to address several key questions:

How do Early Head Start programs affect child, parent, and family outcomes?
How do different program approaches and community contexts affect these outcomes?
How do program implementation and services affect outcomes?
How do the characteristics of children and families affect outcomes?

These broad questions are translated into more specific research questions as we approach the analysis of impacts on services, children, parenting, and families (and are presented within the appropriate chapters).

2. Conceptual Framework

Like its older sibling Head Start, Early Head Start has the ultimate goal of promoting children’s “competence,” in the fullness of Zigler’s original definition—children’s “everyday effectiveness in dealing with their present environment and later responsibilities in school and life” (Zigler 1973). Infants and toddlers, however, have unique qualities that are different from those of preschool-age children, including their period of rapid development and important developmental milestones (such as developing trust and language). Good nutrition and health are particularly important during the first three years of life, as are both emotional and cognitive stimulation. Infants and toddlers develop in the context of relationships, and interventions during this period typically focus on those relationships, especially the one between parent and child.

The five objectives of the Head Start performance measures also apply conceptually to infants and toddlers, even though they were designed for preschool-age children. The objectives describe both processes and outcomes of the program. One can visualize the conceptual framework as a pyramid, with program management and operations at the base, providing the foundation for delivering services, supporting child and family development, and creating the ultimate outcomes that support social competence (Administration on Children, Youth and Families 1998). The evaluation design (described in greater detail in Section E and in Chapter II) follows this overarching framework:

The evaluation of Early Head Start began by documenting and analyzing program implementation to ascertain whether the research programs were well managed and had the potential for making a difference in the lives of children and families.
We collected extensive data on program services for both program and control groups to determine the extent to which programs (1) provided children and families with the appropriate services, and (2) linked children and families to needed community services and resources.
We then measured children’s growth and development, along with their families’ functioning and strengths and, by contrasting them with the same measures in control group children and families, assessed the impacts the 17 research programs are having at this early stage in their development.

3. Overarching Hypotheses

As described in Section D, Early Head Start programs strive to influence children’s development, parenting, and family functioning through three main approaches (center-based, home-based, and mixed). Within these approaches, we see that programs may follow multiple pathways for achieving their outcomes. Although service delivery strategies are implemented in diverse ways, they reflect two primary pathways to achieving the ultimate enhanced development of infants and toddlers (these can also be thought of as alternative theories of change by which programs achieve their effects):

The direct child pathway, for which we hypothesize that impacts on children’s development will be either more probable or stronger than impacts on parenting, parent-child interactions, and family functioning. Programs emphasizing this pathway work with children and families primarily through child development centers. Caregivers interact directly with children to establish relationships, and conduct activities designed to enhance children’s health and their cognitive, social-emotional, and physical development. These programs also support families through social services, parent education, and parent involvement, but most services are child-focused.
The indirect child pathway through parenting and parent-child relationships, for which we hypothesize that impacts on parenting, parent-child relationships, and family functioning will be more common or stronger than the impacts on children’s development, at least during the first two years of life. We hypothesize that child development impacts will manifest themselves somewhat later than through the direct child pathway. Programs emphasizing this pathway work with children and families primarily through home visiting (combined with social supports and group socialization activities). Home visitors interact with parents with the aim of strengthening the parent-child relationship, enhancing parenting skills, and supporting their efforts to provide an educationally stimulating and emotionally responsive home environment. These activities are then expected to lead to changes in children’s health, cognitive, social-emotional, and physical development.

Programs may follow multiple pathways for achieving their desired outcomes. In practice, their emphasis on each pathway varies. Hypothesized impacts depend on the balance adopted by the particular program, that is, whether it takes (1) predominantly a direct child pathway, with some parent and parent-child focus in the services offered (center-based programs); (2) predominantly an indirect pathway through parenting, with some direct child services added (home-based); or (3) a balance of these two pathways (mixed approach). Program impacts may also vary depending on the emphasis placed on the indirect pathways through family support. Programs whose theory of change follows either a direct or an indirect path to child development also strive to strengthen family self-sufficiency and resources so that parents are better able to provide emotional and educational stimulation for their children and to interact with them in positive ways.

In general, programs that emphasize creating a balance of both direct and indirect pathways would be expected to have stronger impacts on parenting and family outcomes than programs that emphasize the direct child pathway. They would also be expected to have stronger child development outcomes than programs that emphasize the indirect pathway through parenting. Because little research has been conducted with programs that emphasize both pathways, the Early Head Start evaluation examines more than one hypothesis. Programs emphasizing both pathways (the mixed-approach programs) may have more flexibility to respond to the varying needs of families, by providing predominantly home visiting, predominantly center care, or a mixture of the two that is tailored to the needs of the individual family. This flexibility may create a synergy that leads to effects greater than the effects of either of the two approaches alone. On the other hand, it is possible that in the short term, some dilution in both child and parent/family impacts could occur if emphasizing both pathways stretches the program’s resources or creates complex operational challenges.

In the context of this basic conceptual framework, Chapters V and VI(which describe program impacts on children, parenting, and families overall and for programs taking different approaches) begin with a discussion of hypothesized effects in each outcome area.

D. THE EARLY HEAD START PROGRAMS, FAMILIES, AND COMMUNITIES

The Early Head Start Research and Evaluation project was carried out in 17 sites that were purposively selected as generally reflective of all the Early Head Start programs funded during the first two funding cycles of Early Head Start. In the following subsections, we describe the types of approaches the research programs followed in delivering Early Head Start services, the families the programs served, the communities where the research programs operated, and how the research programs compared with all Early Head Start programs funded in Waves I and II. In Chapter 2, in the context of the study methodology, we provide a more in-depth discussion of how the research sites were selected.

1. The 17 Early Head Start Research Programs

Unlike some programs, Early Head Start does not embrace a particular program “model,” but asks each grantee to select service delivery options that will best meet the needs of the families and communities it serves. The period of dynamic change since the beginning of Early Head Start has provided ample opportunity for program adaptations over time. Each program has strived to implement the revised performance standards, find the approach (or mix of approaches) that will continue to meet changing family needs, and strengthen strategies that will promote children’s development. Early Head Start programs try to meet families’ and communities’ needs through one or more official program options: (1) home-based, (2) center-based, (3) combination (in which families receive both home visits and center experiences), and (4) locally designed.

Because a program may offer multiple options, we characterized programs for research purposes according to the options they offer families:

Center-based programs, which provide all services to families through the center-based option (center-based child care plus other activities) and offer a minimum of two home visits a year to each family
Home-based programs, which provide all services to families through the home-based option (weekly home visits and at least two group socializations a month for each family)
Mixed-approach programs, which provide services to some families through the center-based option and to some through the home-based option, or provide services to families through the combination or locally designed option (services can be mixed in the sense of programs targeting different types of services to different families or in the sense that individual families can receive a mix of services either at the same time or at different times; thus, in different ways, programs adjust the mix of home- and center-based services to meet the needs of families); these programs may also include child care provided directly by the Early Head Start program or through partnerships with community child care providers.

The 17 programs selected to participate in the national Early Head Start Research and Evaluation Project include 16 Wave I programs (the 68 programs funded in 1995) and 1 of the 75 Wave II programs funded in 1996. They are located in all regions of the country and in both urban and rural settings, and include all major Early Head Start program approaches. The families served are highly diverse, as described later.

When funded, the research programs were about equally divided among the three program approaches (Figure I.2). By fall 1997, the home-based approach predominated, having increased from five to seven programs (four were center-based and six were mixed-approach in fall 1997). Program approaches continued to evolve, and by fall 1999, most home-based programs had become mixed-approach.

This evolution in program approaches occurred as programs responded to changing family needs, particularly the increasing need for child care. Some programs changed their approaches in fundamental ways; others significantly altered services within their basic approach. Details of this evolution are described in the Pathways to Quality report, but we summarize key changes here. Comparing programs in 1997 and 1999 (the two periods in which we obtained detailed implementation data from site visits), we see that:

The four programs that had a center-based approach in 1997 remained center-based throughout but enhanced their programs in a variety of ways, such as achieving NAEYC accreditation; strengthening staff development; adding more classrooms; reducing group sizes; making changes that promoted greater continuity of care; collaborating more closely with welfare-to-work case managers; and expanding health, nutrition, and mental health services.
Two of the seven home-based programs continued to provide home-based services to all families while adding enhanced support for families’ efforts to use good-quality child care.

FIGURE I.2

THE EVOLUTION OF PROGRAM APPROACHES OVER TIME
[D]

Five of the seven home-based programs expanded services options to such an extent that by fall 1999 they had become “mixed” in their approach to serving families. The changes included (1) helping families find good child care and paying for quality child care that some home-based families used, (2) adding a child care center to serve a small portion of the enrolled families for whom the home visiting approach was not appropriate, (3) working with community partners to improve community child care, (4) visiting children in their child care settings as well as in their homes, and (5), in some cases, contracting with community child care partners for center-based services that met the Head Start performance standards.
The six mixed-approach programs continued taking a mixed approach, but by 1999 they had expanded some service options, including obtaining state funding to enhance the program’s ability to provide child care assistance, increasing home visit time spent on parent-child activities, taking formal steps to ensure that child care providers used by Early Head Start families met the revised Head Start Program Performance Standards, adding child care classrooms, and forming collaborations with state child care administrators.

Research programs varied along a number of dimensions that provide important context for their evaluation. One dimension is the variety of experiences programs brought to their new mission as Early Head Start grantees. Nine of the grantees had operated Head Start programs (four of these had not offered infant-toddler services before); one had previously operated a Parent Child Center (PCC) as well as Head Start; seven had been Comprehensive Child Development Programs (CCDPs) (five of these were new to Head Start but had served infants and toddlers); and three of the grantees without Head Start, PCC, or CCDP experience had operated other community-based programs. Many of the grantee agencies had experience offering infant-toddler services.

2. The Families That Early Head Start Research Programs Served

Table I.1 displays key characteristics of the 1,513 Early Head Start families at the time they entered the research programs. At the time of enrollment, primary caregivers were diverse:2

Early Head Start applicants (99 percent of whom were mothers) were on average 23 years old. The mean age across the programs ranged from 18 to 26. About 62 percent were first-time parents.
One-fourth of the primary caregivers lived with a spouse. Slightly more than one-third lived with other adults, and a similar proportion lived alone with their children.
Teenage parents headed slightly more than one-third of families enrolled in Early Head Start. The percentage ranged from 19 to 90 across the 17 programs.
Overall, one-third of families were African American, one-fourth were Hispanic, and slightly more than one-third were white (with a small percentage in other groups). Twelve programs were relatively homogeneous, with two-thirds or more of the families representing a single racial/ethnic group (four programs enrolled predominantly African American families, three were predominantly Hispanic, and five were predominantly white); in five, the racial/ethnic composition was diverse.
Overall, one-fifth of the Early Head Start primary caregivers did not speak English as their primary language, although in two programs more than half reported not speaking English well.
Nearly half the Early Head Start primary caregivers did not have their high school diploma at the time they enrolled (however, in four programs, two-thirds were high school graduates, and in three programs two-thirds were not).
At enrollment, 45 percent of primary caregivers were employed or in school or training.

Most families were receiving public assistance of some kind (77 percent were covered by Medicaid, 88 percent were receiving WIC benefits, almost half were receiving food stamps, just over one-third were receiving AFDC or TANF, and 7 percent were receiving SSI benefits).

TABLE I.1

KEY CHARACTERISTICS OF FAMILIES ENTERING THE EARLY HEAD START
RESEARCH PROGRAMS
Primary Caregiver's Race/Ethnicity
	All Research Programs Combined (Percentage)	Range Across Research Programs (Percentage)
Primary Caregiver (Applicant) Is Female	99	97 to 100
Primary Caregiver Is a Teenager (Under 20)	39	19 to 90
Primary Caregiver Is Married and Lives with Spouse	25	2 to 66
African American	34	0 to 91
Hispanic	24	0 to 90
White	37	2 to 91
Other	5	0 to 14
Primary Caregiver's Main Language Is Not English	20	0 to 81
Primary Caregiver Does Not Speak English Well	11	0 to 55
Primary Caregiver Lacks a High School Diploma	48	24 to 88
Primary Caregiver's Main Activity
Employed	23	11 to 44
In school or training	22	4 to 64
Neither employed nor in school	55	24 to 78
Primary Caregiver Receives Welfare Cash Assistance (AFDC/TANF)	36	12 to 66
Number of Applicants/Programs	1,513	17
SOURCE: Head Start Family InformationSystem application and enrollment data.

Approximately one-fourth of primary caregivers enrolled while they were pregnant. The percentage that were pregnant ranged from 8 to 67 percent across the programs.
HSFIS items relating to families’ needs and resources indicated that the greatest self-reported needs of parents were for adequate child care (34 percent of families overall, ranging from 11 to 68 percent across the programs); transportation (21 percent, ranging from 12 to 35); and medical care (14 percent overall, ranging from 3 to 36 percent).

To supplement the baseline data available through the HSFIS, several local research teams worked with their program partners to collect information about their families that would provide a richer understanding of their characteristics. Eight teams obtained comparable maternal mental health data using the CES-D (Center for Epidemiological Studies Depression) scale, which provides information on the mothers’ risk for depression. Across these eight programs, on average, 48 percent of parents scored in the at-risk range; this percentage ranged from 34 to 73 percent across the eight programs.

To be eligible for the research, the primary caregiver in the research program families had to be pregnant or have a child younger than 12 months of age. About 25 percent of the families enrolled while the mother was pregnant. The Early Head Start children who were born by the time of enrollment had diverse characteristics:

They varied in age, with almost half under 5 months.
Sixty-one percent were firstborn children.
About 10 percent were low birthweight (under 2,500 grams), although the figure was 24 percent in one program.
About 20 percent might have had—or were at risk for—a developmental disability.3

3. The Communities Served by Early Head Start Research Programs

The 17 research programs are distributed across the major regions of the country—six in the West, four in the Midwest, four in the Northeast or Mid-Atlantic, and three in the South. About half are in urban areas and half in small towns or rural areas, with home-based, center-based, and mixed-approach programs in each. Most programs are located in areas of low unemployment (the median 1998 unemployment rate was 3.8 percent, and, the national unemployment rate was about 4.5 percent). Four of the research programs are in cities or areas where unemployment exceeded 5.5 percent in 1998; the rates across those sites ranged from 5.5 to 10.4 percent. In these communities with higher unemployment rates, staff described job and job-training opportunities as inadequate.

Welfare reform influenced the community context in several ways. One key factor affecting Early Head Start families was whether or not the state (or, in some cases, the county) exempted mothers of infants under 12 months of age from the work requirements. Seven of the research programs operated in areas where there was no exemption. In these areas, mothers were expected to enter the workforce when their babies reached ages ranging from 6 weeks to 9 months.

A few programs described their communities as “service rich,” yet all identified some services for low-income families that were inadequate or lacking. As Chapter IV documents, families in the control group, who did not have the benefits of Early Head Start, generally received substantially fewer services. During implementation study visits, staff reported the major service inadequacies in communities to be lack of affordable and high-quality child care, insufficient affordable housing, and poor public transportation.

4. How Early Head Start Research Programs Compare with All Funded Programs from Which They Were Selected

The 17 selected research programs reflect the populations served by all Wave I and II programs from which they were selected (Table I.2).4 For example:

The average number of families enrolled in the research programs (85) is very similar to the number in Wave I (81) and Wave II (84) programs.
The racial/ethnic distribution is similar, but the research programs have a slightly larger percentage of African American families and a slightly smaller percentage of white families.
The percentage of single- and two-parent families in the research programs is similar to the average percentage in the Wave I and II programs.
About the same percentage of primary caregivers are in school or training.

Although the findings reported in subsequent chapters are not statistically generalizable to all Early Head Start programs because they were not randomly selected (see Chapter II), they are relevant to the rest of the programs because (1) the research sites include the full range of locations and program approaches, (2) the families served by the research programs resemble the families served by other Wave I and II programs, and (3) the research sites encompass variations on other key dimensions that ACYF considered in funding Early Head Start programs (e.g., variations in race/ethnicity of families served, former auspice, experience serving infants and toddlers directly, and years in operation). Thus, the lessons drawn from the experiences of these programs are likely to be applicable to the others.

TABLE I.2

COMPARISON OF RESEARCH PROGRAMS AND WAVE I AND II PROGRAMS
Total ACYF-Funded Enrollment
	Wave I Programs (Percentage)	Wave II Programs (Percentage)	Research Programs (Percentage)
10 to 29 children	6	0	0a
30 to 59 children	14	9	6
60 to 99 children	62	64	65
100 to 199 children	15	27	29
200 to 299 children	3	0	0
(Average)	-81	-84	-85
Race/Ethnicity of Enrolled Children
African American	33	21	34a
Hispanic	22	27	24
White	39	46	37
Other	6	5	5
English Is the Main Language	85	79	80
Family Type
Two-parent families	39	46	40
Single-parent families	51	46	52
Other relativesb	7	5	3
Foster families	1	1	0
Other	1	1	5
Employment Statusc
In school or training	20	22	22
Not employed	48	48	55
Number of Programs	66	11	17
SOURCE: Program Information Report data (columns 1 and 2) and Head Start Family Information System application and enrollment data (column 3). NOTE: The percentages for the Wave I and II Early Head Start programs are derived from available Program Information Report (PIR) data. The percentages for the Early Head Start research programs are derived from Head Start Family Information System application and enrollment data from 1,513 families. Percentages may not add up to 100, as a result of rounding. aThe data for the research programs refer to families instead of children.(back) bThe HSFIS data elements and definitions manual instructs programs to mark “other relatives” if the child is being raised by relatives other than his/her parents, such as grandparents, aunts, or uncles, but not if the child is being raised by his/her parents and is living with other relatives as well.(back) cThe research program data and PIR data are not consistent in the way that they count primary caregivers’ employment status, so it is not possible to compare the percentage of caregivers who are employed.(back)

E. OVERVIEW OF THE EVALUATION

1. Description of the Evaluation

The National Early Head Start Research and Evaluation Project is a cross-site national study conducted by Mathematica Policy Research, Inc. (MPR) and Columbia University’s Center for Children and Families at Teachers College, in collaboration with the Early Head Start Research Consortium (staff of the 17 research programs, local researchers, and federal staff). All together, the study encompasses the following components:

Implementation Study. Issues related to program implementation have been addressed in the Early Head Start implementation study and reported in two sets of reports; see Leading the Way (Administration on Children, Youth and Families 1999, 2000a, 2000b) and Pathways to Quality (Administration on Children, Youth and Families 2002).
Continuous Program Improvement. Throughout the evaluation, reports and presentations have provided new information that all Early Head Start programs can use to enhance their ability to meet their families’ needs.
Impact Evaluation. Program impacts are the focus of this report and of the interim report, Building Their Futures (Administration on Children, Youth and Families, 2001) .
Local Research Studies. Elements of these are integrated in this report, in boxes throughout the chapters of this volume and in Volume III. The local university research and program teams will report other local findings independently.
Special Policy Studies. These include studies of issues relating to welfare reform, children’s health, child care, and fatherhood. Key findings from the Early Head Start Father Studies are presented in this report. Special reports on child care and on children’s health will be issued separately, as will additional reports focused on particular issues related to father involvement.

The impact analyses (reported here) focus on program impacts on children and families; analyses of outcomes in the staff and community development areas are reported in the Pathways to Quality implementation report (Administration on Children, Youth and Families 2002). The study is grounded in an experimental design in which 3,001 families across the 17 program sites were randomly assigned to participate in Early Head Start or to be in the control group. The impact analyses benefited from partnerships with 15 local research teams that contributed site-specific findings from local research and brought the perspectives of researchers and program staff at the local level to the interpretation of the cross-site impact findings.

2. The Early Head Start Research Consortium

Under its contract with ACYF, MPR worked with the local research teams, the program directors from the research sites, and ACYF to create the Early Head Start Research Consortium. Beginning in April 1996, shortly after the local research grants were awarded, the consortium met two or three times each year to review evaluation plans (including instruments, data collection procedures, and data analysis plans) and collaborate on various reporting and dissemination activities. As described in Appendix B, in all but one of the sites, local researchers were responsible for all data collection (conducted under subcontract to MPR). The consortium created several workgroups to carry out research activities related to special topics, such as welfare reform, fatherhood, disabilities, and child care. The evaluation reports (including this one and those listed on page ii) embody the spirit of collaboration, as committees of consortium members reviewed the plans for and early drafts of this report and local research and program partners contributed brief reports of local studies, which have been incorporated into this report. The consortium members and their member institutions are listed in Appendix A.

3. Overview of the Implementation Study and Its Findings

The national evaluation includes a comprehensive implementation study that measured the extent to which programs had become “fully implemented” in 1997 and 1999. The assessment of implementation was based on 24 selected key elements of the program guidelines and the revised Head Start Program Performance Standards, as described in Leading the Way (Administration on Children, Youth and Families 1999 and 2000) and Pathways to Quality (Administration on Children, Youth and Families 2002). Data were collected in three rounds of site visits, and a panel of site visitors, national evaluation representatives, and outside experts, using a consensus-based approach, assessed the degree of implementation both overall and separately for the child and family development areas, as well as staff development, community partnerships, and some aspects of program management (see Appendix C).

One-third (six) of the programs were judged to be fully implemented overall by the fall 1997 implementation visits and continued to be fully implemented in late 1999 while still expanding the numbers of families served. We refer to these as the early implementers. By fall 1999, two-thirds of the programs were fully implemented. We refer to the six that reached this level after 1997 as the later implementers. The remaining five programs did not achieve ratings of “fully implemented” during the evaluation period. We refer to them as the incomplete implementers, all of which nevertheless made strides in particular program areas and, in fact, showed a number of strengths. In general, these programs were not rated as “fully implemented” in child development and health services but tended to have strong family development services.

As part of the implementation rating process, we also rated the degree of implementation of child development and health services, which included programs’ efforts in (1) conducting developmental assessments, (2) individualizing child development services, (3) involving parents in child development services, (4) promoting group socializations, (5) providing child care that meets the performance standards, (6) supplying health services for children, (7) offering frequent child development services, and (8) providing services for children with disabilities. Eight programs achieved a rating of “fully implemented” in this area in 1997, a number that increased to nine by 1999.

In the area of implementing family partnerships, we considered programs’ progress in (1) Individualized Family Partnership Agreements; (2) availability of services; (3) frequency of services; and (4) parent involvement in policymaking, operations, and governance. In fall 1997, 9 programs were rated as “fully” implemented in family partnerships; this increased to 12 programs by fall 1999.

The implementation study also assessed key aspects of the quality of both home- and center-based child development services. We assessed the quality of child care received by program families, including the care provided in both Early Head Start centers and other community child care settings. See Pathways to Quality for a detailed description of our assessment of these data (Administration on Children, Youth and Families 2002).5 Assessments of the child care arrangements used by program families are based on both field staff observations of child care settings and data collected from program staff during site visits. Observations of child care settings were made in conjunction with the study’s 14-, 24-, and 36-month data collection and included use of the Infant/Toddler Environment Rating Scale (ITERS)6 and the Family Day Care Environment Rating Scale (FDCRS),7 as well as observed child-teacher ratios and group sizes.

The quality of child care provided by Early Head Start centers during their first two years of serving families was good.8 All nine programs that operated centers from the beginning scored above 4 (the middle of the minimal-to-good range) on the ITERS, with the average being 5.3 (in the good-to-excellent range). Observed child-teacher ratios (2.3 children per teacher in 1997-1998 and 2.9 children per teacher in 1998-1999) and average group sizes (5.3 children in 1997-98 and 5.9 in 1998-1999) were well under the maximum allowed by the revised Head Start Program Performance Standards (below 4.0 children per teacher and 8.0 children per group).

Children in programs that did not offer center care often attended child care in community settings. The quality of care received by Early Head Start children in community child care centers varied widely across sites, with average ITERS scores ranging from 2.9 (minimal) to 5.9 (good to excellent) in 1998-1999. Overall, the average ITERS score in community child care centers was 4.4 (minimal to good). Average FDCRS scores ranged from 2.0 (inadequate to minimal) to 4.5 (minimal to good) across sites in 1998-1999; the average FDCRS scores were 3.3 (minimal) in 1997-1998 and 3.5 (minimal to good) in 1998-1999. However, observed child-teacher ratios and group sizes were in most cases lower than those set by the Head Start performance standards for infants to 3-year-olds (3.3 children per teacher in 1997-1998 and 4.2 in 1998-1999). The average group size in the family child care settings that we were able to observe was 4.5 children in the first year and 4.8 children in the second year. Some of the community settings were formal partners of Early Head Start programs and agreed to follow the performance standards; in other cases, parents found community child care on their own.

In fall 1999, 12 of the research programs operated Early Head Start centers. Most of them received good or high ratings on several factors that may be responsible for child care quality—curriculum, assignment of primary caregivers, and educational attainment of teachers. Among all research programs, between one-fourth and one-half received good or high ratings in monitoring and in training and support for child care providers.

Since the study was not able to observe home visits directly, we rated quality of child development home visits by considering program factors that are related to service quality. These included supervision, training, and hiring of home visitors; planning and frequency of home visits; and the extent to which staff reported that home visits emphasized child development and were integrated with other services. By fall 1999, 11 of the 13 programs that served some or all families in a home-based option received a good or high rating of quality, up from 9 programs in 1997.

The implementation study provided a solid foundation on which to build the impact evaluation. We learned that all programs were able to implement key features of the performance standards but that programs varied considerably in both their rate and completeness of implementing those standards. We learned much about the variation in services that programs following different approaches offered, and saw strengths and challenges in center-based, home-based, and mixed-approach programs. We also saw the great diversity in the families that the 17 Early Head Start programs served. These programmatic and family variations enabled the evaluation to learn much about what kinds of programs are effective, how variations in program strategies and implementation are associated with differential effectiveness, and how the programs are differentially effective for different types of families. After describing the evaluation’s design and methods in the next chapter, we then report the findings—both overall and in relation to subgroups of programs and families—in Chapters III through VII.

1The 1994 and 1998 Head Start reauthorizations directed that the percentage of the annual Head Start budget allocated to the new Early Head Start program was to begin at 3 percent in 1995 and increase to 9 percent for 2001 and 10 percent for 2002 and 2003. (back)

2We describe program and family characteristics at the outset of the study based on data from the Head Start Family Information System (HSFIS) application and enrollment forms that families completed at the time of application to the program. Programs submitted these forms to MPR for random assignment, and the date of the families’ random assignment is used as the starting point for considering the timing of services and events captured by the evaluation. In most cases, program enrollment occurred within a month of random assignment. (back)

3This percentage is an estimate. In Chapter III, we present information that the primary caregivers supplied approximately 6, 15, and 26 months after random assignment. The HSFIS contains more detailed data about the health and developmental conditions that are often associated with diagnoses of disabilities in young children. (back)

4This analysis compared family characteristics of the 17 research programs with those of all Wave I and II programs using the ACYF Program Information Report (PIR) database. (back)

5A special policy report on child care in Early Head Start will be produced in 2002 that includes a more extensive analysis of child care use and quality. (back)

6The Infant/Toddler Environment Rating Scale (ITERS) (Harms, Cryer, and Clifford 1990) consists of 35 items that assess the quality of center-based child care. Each item is ranked from 1 to 7. A ranking of 1 describes care that does not even meet custodial care needs, while a ranking of 7 describes excellent, high-quality, personalized care. (back)

7The Family Day Care Environment Rating Scale (FDCRS) (Harms and Clifford 1989) consists of 35 items that assess the quality of child care provided in family child care homes. Items in the FDCRS are also ranked from 1 to 7, with 1 describing poor-quality care and 7 describing high-quality care. (back)

8Because response rates were low in some sites, we may not have information for a representative sample of Early Head Start children’s child care arrangements. (back)

Table of Contents | Previous | Next