Table of Contents | Previous | Next |
Appendix 2.2: Determination of Head Start Participation
Chapter 2 provides information on the incidence of no-shows (children randomly assigned to the Head Start group but who failed to participate) and crossovers (children randomly assigned to the non-Head Start group but who participated in Head Start). These data are provided by age cohort and for both the total sample that was randomly assigned and the subset of children who are part of the Year 1 analysis sample. For this purpose, a child was considered a “no-show” if it was not possible to identify a time when he/she participated in Head Start during the 2002-03 program year after checking several data sources. Similarly, a child was deemed a crossover if information was obtained indicating that he/she participated in Head Start at any time during the 2002-03 program year.
The determination of participation status was first based on three sources of information:
-
EnrolOct. Information reported by site coordinators from a check in early November 2002 with the Head Start centers where random assignment was done on whether a child was enrolled and/or attending the center on or before Oct 15.
-
P3ENROL. Each parent’s response as to whether their child was currently attending Head Start. This information could come from either the fall 2002 (P3ENROL_fall) or the spring 2003 (P3ENROL_spr) parent interview or from both if available.
-
P4EVER. Each parent’s response in fall 2002 as to whether their child had ever attended Head Start.
The P3ENROL_fall variable was combined with the P4EVER variable to create a fall parent indicator of Head Start services (Parent_HS_fall). A cross-tabulation was then run on EnrolOct x Parent_HS_fall x P3ENROL_spr, and the results were examined to determine which cases received Head Start over the year and which required further investigation to determine Head Start services (Exhibit A.2.2.1).
Two pieces of corroborating data were needed to make a final determination of each child’s Head Start participation. For example, if the Head Start center where random assignment took place reported the child as attending Head Start, the fall 2002 parent interview reported the child in Head Start, but the spring 2003 parent interview reported the child in non-Head Start center-based care, the child was coded as receiving Head Start services at least sometime during the year.
Other cases were more difficult to determine, such as those with a single initial indicator that Head Start participation took place. For example, if the random assignment center reported the child as NOT in Head Start, and if the parent stated that the child received Head Start in fall 2002 and the child was not in Head Start in the spring, additional data were used to determine whether a child actually attended Head Start at any point (see below). There were 322 cases requiring further investigation (i.e., they could not be determined based on the “two corroborating pieces of information” rule.)
For the 322 cases requiring further investigation, the following additional data were examined to determine Head Start participation:
-
The type of setting that parents reported their child was currently attending in the fall and/or spring.
-
The name of the current setting provided by the parent from the fall and/or spring parent interview.
-
Parent-reported dates in fall when the child started and stopped Head Start services.
-
Parent-reported dates in fall when the child started and stopped another non-Head Start child care arrangement.
-
The location where the child was reported by the study’s in-person assessor to be receiving services at the time of the fall and/or spring assessments (N=114).
-
The location of the classroom where observation was done for a particular child by the study’s in-person assessor in the fall and/or spring (used in only a few cases).
Information obtained from the spring 2003 interview with care providers of children in non-center-based non-parental care at that time.
| Fall 2002 Program Attendance Record | Fall 2002 Parent Interview | Spring 2003 Parent Interview | ||
|---|---|---|---|---|
| Head Start | Not Head Start | Missing | ||
| Head Start | Head Start | Participant (N=1,750) | Participant (N=121) | Participant (N=94) |
| Not Head Start | Participant (N=9) | SB (N=31) | SF (N=1) | |
| Missing | Participant (N=77) | SS (N=7) | Participant (N=78) | |
| Not Head Start | Head Start | Participant (N=161) | SB (N=85) | SF (N=23) |
| Not Head Start | SB (N=82) | Non-Participant (N=1,205) | Non-Participant (N=94) | |
| Missing | SS (N=48) | Non-Participant (N=201) | Non-Participant (N=454) | |
| Missing | Head Start | Participant (N=37) | SB (N=9) | SF (N=3) |
| Not Head Start | SB (N=1) | Non-Participant (N=39) | SF(N=8) | |
| Missing | SS (N=12) | SS (N=12) | [Impute] (N=25) | |
| Cases requiring further investigation: SB – checked both fall and spring parent interviews during further investigation (N=208), SF – checked fall parent interview during further investigation (N=35), SS – checked spring parent interview during further investigation (N=79). |
The process used to investigate these cases was as follows:
-
We looked at the fall and/or spring parent interview care setting name, type of setting, Head Start start and end dates (if applicable), and current care provider start dates. In addition, we checked data on the location of fall and/or spring child assessments and from the spring non-parental care provider interview. Based on these data items, we made a determination whether the child received Head Start services (N=294) or concluded that no such determination could be made and imputed participation status instead (N=3).
-
We imputed participation status (without checking any other sources) for cases for which all three original data items (i.e., EnrolOct, Parent_HS_fall, and P3ENROL_spr) were missing (N=25).
| Table of Contents | Previous | Next |