Table of Contents | Previous | Next |
Appendix A: Descriptions of the FACES, NLSY79, SIPP, and PSID Databases
Descriptions of the Datasets
For each dataset, a subsample of families with 3- to 5-year-old
children was selected based on their age eligibility for Head Start. The
lone exception was the Head Start FACES dataset, which only included families
with enrolled Head Start children.
Head Start FACES. The Head Start FACES represents the first cohort
of a periodic, longitudinal data collection with a nationally representative
sample of Head Start families. This cohort of 3,200 children and families
was randomly drawn from 40 Head Start programs across the United States.
Using a broad array of measures, FACES focused on classroom quality, children's
experiences in Head Start, children's status at entry into and completion
of kindergarten, and characteristics of Head Start families and how local
programs serve them. Since this dataset includes findings on a nationally
representative sample of Head Start families only, it is provided as a
reference for judging the findings about Head Start families in the other
datasets.
NLSY79: National Longitudinal Survey of Youth 1979. The National Longitudinal Surveys, sponsored by the Bureau of Labor Statistics, U.S. Department of Labor, are a set of surveys designed to gather information at multiple points in time on the labor market experiences of diverse groups of men and women. The National Longitudinal Survey of Youth 1979 (NLSY79) consists of a national probability sample drawn of young men and women living in the United States and born between January 1, 1957 and December 31, 1964. The sample included an overrepresentation of Blacks, Hispanics, and economically disadvantaged non-Black, non-Hispanics. The NLSY79 sample was first interviewed in early 1979 and has been re-interviewed 17 times since.
The NLSY79 Children is a child sample, which began in 1986, and targets children born to female NLSY79 respondents. During the 1996 Survey, interviews were completed with 7,103 children, or more than 87% of the children born to interviewed NLSY79 mothers. The NLSY79 children included in these analyses were 3- to 5-years old at the time of the 1996 survey (conducted between April and October, 1996). The respondents to the survey were the mothers of these children, i.e., the NLSY79 female respondents. These women ranged in age from 31 years to 38 years old at the time of the 1996 survey.
SIPP: The Survey of Income and Program Participation. The SIPP, sponsored by the United States Census Bureau, is a continuous series of national panels, with a sample size ranging from approximately 14,000 to 36,700 interviewed households. The duration of each panel ranges from 2 ½ years to 4 years. The SIPP sample is a multistage-stratified sample of the United States civilian non-institutionalized population. From 1984 to 1993, a new panel of households was introduced in February of each year.
The SIPP content is built around a "core" of labor force, program participation, and income questions, designed to measure the economic situation of persons in the United States. Each national panel consists of a series of data collection waves. Waves are conducted every four months, and cover the preceding four months (with the exception of the Head Start and child care questions, which cover the previous one month only). Within each wave, approximately the same numbers of interviews are conducted during each month of the 4-month period. The current project used data from the SIPP 1993 Panel, Wave 9, which was conducted during the period October 1995 through January 1996.
In addition, the survey was designed to provide a broader context for analysis by adding questions on a variety of topics not covered in the core section. These "topical modules" are assigned to particular interviewing waves of the survey. Among the topics covered by the modules are personal history, child care, wealth, program eligibility, child support, disability, school enrollment, taxes, and annual income. The 1993 Panel, Wave 9 included a section on use of child care.
All household members 15-years old and over were interviewed; proxy response was permitted when household members were not available for interviewing. This study provides information on the characteristics of the respondent identified as the designated parent or guardian of one or more Head Start-age children.
PSID: The Panel Study of Income Dynamics. The PSID, initiated in 1968, is a longitudinal study of a representative sample of U.S. individuals (men, women, and children) and the family units in which they reside. It emphasizes the dynamic aspects of economic and demographic behavior, but its content is broad, including sociological and psychological measures. As a consequence of low attrition rates, the success of re-contact efforts, and the growth of participating families, the sample size has grown dramatically in recent years, from about 7,000 core households in 1990 to almost 8,700 in 1995. The study is conducted at the Survey Research Center, Institute for Social Research, University of Michigan.
Starting with a national sample of 5,000 U.S. households in 1968, the PSID has re-interviewed individuals from those households every year since that time, whether or not they are living in the same dwelling or with the same people. Adults have been followed as they have grown older, and children have been observed as they advance through childhood and into adulthood, forming family units of their own. Information is collected each year about the original 1968 sample individuals, their current co-residents as well as their children, regardless of whether they currently live with the respondent. The core sample is representative only of everyone who was in the United States in 1968 or those who have been born to such persons. As a result, recent immigrants to this country are under-represented in the sample. To remedy this shortcoming, a representative national sample of 2,000 Latino households was added to the study in 1990. That sample was differentially sampled to provide adequate numbers of Puerto Ricans, Mexican-Americans, and Cuban-Americans.
The PSID gathers information about families and all individuals in those families through annual interviews. A single primary adult - usually the male adult head of household, if there is one - serves as the sole respondent. Sometimes the wife (or cohabitor, referred to as "wife") of the head of household agrees to grant an interview when the household head does not. The single household respondent provides information about him/herself and about all other family members. The central focus of the data is economic and demographic, with substantial detail on income sources and amounts, employment, family composition changes, and residential location.
The Head Start participation question first appeared on the 1995 PSID Survey, but the most current final release PSID data was the 1993 Survey. Therefore, in order to analyze issues related to Head Start participation, it was necessary to identify Head Start participation by means of the 1995 Survey, and then examine family characteristics using the 1993 Survey dataset. This issue is discussed further in the next section.
Caveats About Each Dataset
As noted earlier, readers are cautioned that differences across the datasets preclude the direct comparison of findings. The following discussion illustrates this concern through the examination of one specific issue: the differences in the reported proportion of Head Start-enrolled children among ALL children in the study. The proportion reported for the PSID Sample (18.4%) is much higher than the proportions reported for the SIPP (6.0%) and NLSY79 (4.1%) samples. This discussion explores whether this finding is best explained by differences in the data or by the method of administration of the Head Start enrollment question across datasets. For each dataset, the specific Head Start enrollment question and the sample selection process are discussed.
NLSY79. The 1996 NLSY79 Survey (Child Supplement), which was administered April through October 1996, asked the following question regarding children who were under 8 years of age at the time of the interview/child assessment:
- Has [Child Name] ever been enrolled in the Head Start Program?
To construct this sample, all families with 3- to 5-year-old children were selected, and then sorted on the Yes /No responses to this question. Overall, 4.1% of these children were reported to have ever been enrolled in Head Start.
SIPP. The SIPP 1993 Panel, Wave 9, conducted during the period October 1995 through January 1996, contained the following question regarding children who were less than six years of age at the time the interview was conducted:
- During a typical week in [last month)] please tell me if [respondent] used any of the following arrangements to look after [name of child] while [parent] was working/at school: (The response options included "the Federally supported Head Start program")?
To construct the sample, all families with 3- to 5-year-old children were selected, and then sorted on the Yes /No responses to the Head Start question. Overall, 6.0% of these children were reported to have been enrolled in Head Start in the previous month.
PSID. The PSID question regarding Head Start
was first included during the 1995 survey, and asked the following of
all family members between the ages of 5 and 49:
- [Has he/Has she/Have you] ever been enrolled in Head Start?
Currently, only Early Release Data are available for the 1995 survey, and while those data are reasonably reliable for a variable such as this Head Start question, the Early Release Data generally are considered not yet adequately reliable with regards to income variables. The most current Final Release Data available for the PSID are from the 1993 Survey (which reports upon activities during 1992). Therefore, the sample was constructed of families with children who were 6-, 7-, and 8-years old when the 1995 Head Start question was administered. These children would have been 3-, 4-, and 5-years old in 1992. Analyses were then conducted on the 1993 Survey data for the families of these children. Overall, 18.4% of these children were reported to have ever been enrolled in Head Start.
Discrepancies Among the Datasets. One explanation for the discrepancies in overall Head Start enrollment among the data sets may be found in how and when the enrollment question was asked. While the SIPP asked about Head Start attendance of 3- to 5-year olds during the previous month, the NLSY79 and the PSID asked whether the child was ever enrolled in Head Start. Further, the NLSY79 asked the "ever been enrolled" question when the children were 3- to 5-years old, while the PSID posed this question when the children included in this study were 6- to 8-years old. The retrospective method, asking the "ever been enrolled" question for 6- to 8-year olds, gave them a larger window of time to have "ever been enrolled" than was provided for the 3- to 5-year olds studied in the NLSY79. This is expected to result in a greater number of children ever enrolled.
The key to this difference is the period of time during which the child had the opportunity to ever be enrolled. Under the NLSY79 scenario (i.e., asking the question regarding current 3- to 5-year olds), a 3-year old has up to one year to ever have been enrolled, a 4-year old up to two years, and a 5-year old up to three years to ever have been enrolled in Head Start. Similarly, for the SIPP, the time period covering Head Start enrollment was very restricted, covering only one month. Under the strategy applied (out of necessity) to the PSID data (i.e., asking the question of 6- to 8-year olds), all of the children effectively have a three year period during which they could have ever been enrolled, because they have all exceeded the upper age limit. Given the greater time period during which children in the PSID sample could ever have been enrolled, it seems reasonable to expect that the proportion of PSID children identified as enrolled in Head Start would be higher.
Another point is that the same reason that results in PSID respondents having greater opportunity to report Head Start enrollment also increases the chance for respondent error in recall. The longer the time that has passed since a child was enrolled in preschool, the greater the chance for a recall error. Such an error may occur as a false positive - reporting the child was in Head Start when he/she was not enrolled - or as a false negative - reporting the child was not enrolled when he/she was actually enrolled in a Head Start program. The PSID also used fathers as primary respondents, and there is no current evidence on the accuracy of their recall of their children's preschool experiences that occurred one to three years earlier. The fact that mothers are more likely to be involved with preschool and that they are typically the respondents to such questions suggests that fathers may be prone to increased error in recall.
This discussion illustrates the need to approach the data as independent sources, each with its particular strengths and limitations.
| Table of Contents | Previous | Next |