Table of Contents | Previous | Next |
B. DATA SOURCES AND METHODS
The national Early Head Start Research and Evaluation study was based on a rigorous experimental design. The study sample consists of approximately 3,000 families who applied to one of the 17 research programs and were randomly assigned to either a program group or a control group. The research programs are located in all regions of the country and in both urban and rural areas. They include all major Early Head Start program approaches (home-based, center-based, and mixed) and broadly resemble all the Early Head Start programs funded initially, in both program and family characteristics (Administration on Children, Youth and Families [ACYF] 1999).
MAKING A DIFFERENCE IN THE LIVES OF INFANTS AND TODDLERS AND THEIR FAMILIES: THE IMPACTS OF EARLY HEAD START A rigorous evaluation of Early Head Start services in 17 programs selected from the first groups of programs funded showed they had significant favorable impacts on a wide range of parent and child outcomes, some with implications for children’s later school success. Findings from the study (Making a Difference in the Lives of Infants and Toddlers and Their Families: The Impacts of Early Head Start), using data gathered when children were age 3 and had completed the program, show that the programs sustained and broadened the pattern of impacts reported when children were age 2 (Building Their Futures: How Early Head Start Programs Are Enhancing the Lives of Infants and Toddlers in Low-Income Families, 2001). All Early Head Start evaluation reports are available at [http://www.acf.hhs.gov/programs/opre/ehs/ehs_resrch/index.html]. Early Head Start Improved Many Outcomes. The national evaluation conducted by Mathematica Policy Research, Inc. and Columbia University’s Center for Children and Families at Teachers College, in collaboration with the Early Head Start Research Consortium, reported that 3-year-old Early Head Start children performed significantly better on measures of cognitive, language, and social-emotional development than a randomly assigned control group. In addition, their parents scored significantly better than control group parents on measures of many aspects of the home environment and parenting behavior. Furthermore, Early Head Start programs enhanced parents’ progress toward self-sufficiency. Early Head Start fathers benefited as well. Although these overall impacts were generally modest, the pattern of favorable findings across outcomes in a wide range of key domains is promising. Impacts varied among key program and family subgroups, and impacts were larger in some subgroups. Early Head Start Had a Few Impacts on Health Service Use and Health Outcomes of Children. Early Head Start had small but statistically significant positive impacts on the percentage of children who visited a doctor for treatment of illness (83 compared with 80 percent) and receipt of immunizations (99 compared to 98 percent). The program also decreased the likelihood of hospitalization for an accident or injury in the child’s third year (0.4 compared with 1.6 percent). Although the data on breastfeeding are limited to families who enrolled before their child was born, they suggest that Early Head Start may have increased the rate of breastfeeding to levels found among less disadvantaged families. Findings for impacts on health care service use and health outcomes may have been limited by the high rate of health care services received by both program and control groups and by recruitment strategies. Nearly all families and children in both the program and control groups received some health services, reflecting the accessibility of health services afforded by Medicaid and the State Children’s Health Insurance Programs. It also reflects the fact that many of the research programs recruited families at health clinics or WIC offices, where families were linked to health services before applying to Early Head Start. No Impacts on the Health Care Use or Health Status of Primary Caregivers Were Found. In nearly all families, family members other than the focus child received health services during the follow-up period, and the program impact on such services was not significant. Similarly, Early Head Start had no impact on the receipt of mental health services as reported by parents. We also found no statistically significant impact on the health status of the primary caregiver. |
The data used in this paper come from the approximately 1,500 families who enrolled in the study between July 1996 and September 1998 and were assigned to the program group, which was eligible to receive Early Head Start services. In addition to using baseline data collected using the Head Start Family Information System (HSFIS) Program Application and Enrollment forms, this paper makes use of follow-up data collected at multiple time points based on (1) the number of months since random assignment, and (2) the age of the focus child. Data on use of health services and some data on health status were collected in Parent Services Follow-up Interviews (PSIs) at selected intervals following enrollment—targeted for 6, 15, and 26 months and completed an average of 7, 16, and 28 months after enrollment. Children were, on average, 10, 22, and 32 months old at the time of these interviews. Other data, particularly those related to child development, safety practices, and hospitalizations, were collected in Parent Interviews (PIs) and child assessments on a schedule tied to children’s birth dates—targeted for 14, 24, and 36 months of age and completed on average at 15, 25, and 37 months of age.
The information on health and health care was provided by children’s primary caregivers, not obtained from medical records, so the data may contain discrepancies resulting from difficulty recalling health histories or health care visits, poor communication with health care providers, difficulty understanding medical terminology, or misunderstanding of children’s health conditions (Miller et al. 2001).
Response rates varied across data sources and as expected, declined somewhat over time. Response rates to the PSIs ranged from about 84 percent (6-month PSI) to 71 percent (26-month PSI) of all program families. Response rates to the PIs ranged from 79 percent (14-month PI) to 73 percent (36-month PI) of all program families. Response rates to the Bayley child assessments were somewhat lower, ranging from 64 percent (14-month assessment) to 58 percent (36-month assessment) of all program families. ACF 2002a, Volume I, describes the study design and data in more detail, and Appendix B in ACF 2002a, Volume II, describes the data collection and response rates in greater detail. The data used in this paper were weighted for differential nonresponse.
In this paper, the measures based on PSI data are consistently reported for families who completed the 26-month PSI and provided data for the full follow-up period. Thus, the reported changes over time reflect changes that occurred within the same group of families and children. The age-based measures are reported for all children for whom data were collected at 36 months of age.
All the differences highlighted in the text are statistically significant at the 10 percent level or greater. Statistical tests of differences in means and distributions were conducted using T-tests and chi-squared tests, respectively.
| Table of Contents | Previous | Next |