Table of Contents | Previous | Next |
2.0 Case Study Methodology
2.1 The Sample
The FACES case study sample was a randomly selected, representative1 sample of 120 Head Start families. The sample included three families from each of the 40 Head Start program sites that participated in FACES. The sample was stratified by the age of the Head Start child. Two out of three families selected by site had 4 year olds in Head Start during the 1997-1998 school year, while the other families had a 3-year-old child. Exhibit 2-1 displays basic demographic data of the FACES case study sample.
| Variable | Percentage |
|---|---|
| Children's Race | |
| African American | 28.2 |
| White | 40.2 |
| Hispanic/Latino | 17.1 |
| Native American | 0.9 |
| Asian | 0.9 |
| Other | 12.8 |
| Marital Status | |
| Single | 39.3 |
| Married | 38.5 |
| Separated | 10.3 |
| Divorced | 11.1 |
| Widowed | 0.9 |
| Language other than English spoken in the home | 28.2 |
| Number of Grades Completed | |
| 8 or less | 9.4 |
| 9 | 5.1 |
| 10 | 8.5 |
| 11 | 12.0 |
| 12 | 65.0 |
| Degrees Obtained | |
| No degree | 79.5 |
| AA: | 11.4 |
| BA/BS: | 6.8 |
| Graduate: | 2.3 |
| Financial Resources | |
| Currently Employed | 52.1 |
| Receiving Welfare Benefits | 31.9 |
| Receiving Food Stamps | 55.6 |
| Receiving Medicaid or Medical Assistance | 59.5 |
| Monthly Household Income | |
| Less than $500 | 10.3 |
| $500-999 | 17.9 |
| $1000-1499 | 19.7 |
| $1500-1999 | 12.8 |
| $2000-2499 | 4.3 |
| More than $2500 | 35.0 |
The sample was developed in two stages. The first stage occurred during the field test in the spring of 1997. A sample of 40 families (one from each site) with 3-year-old children who were new to Head Start in the fall of 1996 was randomly selected to participate. Only families with 3-year-old children were chosen in the first stage of sampling in order to maximize the number of children in the sample who would continue to their second year of Head Start (when they were 4-years-old). The second stage of the sampling was completed in the fall of 1997 when an additional 80 families were added to the sample. The 80 new families were comprised of two families from each site (one family with a 3-year-old-child and one family with a 4-year-old child) who were new to Head Start in the fall of 1997. This increased the total sample to 120 families. Families from the original field test sample of 40 that did not return to Head Start in the fall of 1997 (n = 11) were replaced with families with 4-year-old-children starting Head Start in the fall of 19973. There was an overall attrition rate of 12% over the course of the study, with 14 families leaving between the fall of 1997 and December of 1998. Most of these families moved and could not be tracked.
2.2 Measures and Data Collection Procedures
The design of the FACES case study involved the following four primary data collection components4:
- Home visit parent interviews;
- Home and neighborhood observations;
- Monthly telephone interviews; and
- Community agency interviews.
Copies of all data collection instruments can be found in Appendices D1 and D2.
Home Visit Parent Interviews
Each of the three major data collection points (home visits in the spring of 1997, fall of 1997, and spring of 1998) included semi-structured, open-ended interviews to discuss with parents their perceptions of themselves and their families, their experiences with Head Start, and their neighborhoods. Two home-visit parent interview instruments were created. The first instrument was adapted from the "Getting to Know your Family" introductory interview developed by Ramey and Ramey (1992). This instrument was used to develop rapport with the families during the first home visit and to obtain valid and reliable data in a brief amount of time (i.e., 30 minutes). Parents were asked open-ended questions covering the following topics:
- How they would describe their children;
- What were the primary reasons for enrolling their children in Head Start;
- What were their short- and long-term educational and occupational hopes and goals for their children; and
- What they saw as their family strengths and areas of needed improvement.
Home visit interviews were completed with all 40 families in the field study in the spring of 1997 and with all 120 families in the fall of 1997. The second home visit parent interview instrument, used in the spring of 1998, included open-ended questions designed to allow parents to discuss, in greater detail, topics such as the nature of the children's and families' participation in Head Start activities, their parenting beliefs, satisfaction and goals, as well as questions regarding their neighborhoods. Home visit interviews using this second instrument were conducted with 101 of the 110 families remaining in the study in the spring of 1998. This interview took approximately 30 minutes to complete.
Home and Neighborhood Observations
Each of the three major data collection points also included observations of the families’ homes and neighborhoods. The 10-item home observation measure included seven items from the Home Observation for Measurement of the Environment (HOME) physical environment subscale (Caldwell & Bradley, 1984) and 3 items regarding the families' housing type and circumstances. Interviewers also were asked to describe the families' homes using an open-ended question format. A 19-item neighborhood observation checklist was completed by the interviewers in spring and fall of 1997 and asked of the parents in the spring of 1998. Interviewers (and families) were asked to indicate the presence or absence of items in the families' immediate neighborhood.5 Items included neighborhood resources, such as parks, libraries, schools, and grocery stores as well as physical and social neighborhood quality indices, such as abandoned or boarded up buildings, vandalism, graffiti, or loitering. Parents also were asked to rate the overall safety of their neighborhoods on a 5-point scale.
Monthly Telephone Interviews
These interviews provided monthly updates on changes in the families' household composition, child care arrangements, employment status, health status, and Head Start participation that occurred between home visits. In addition, measures of social support (Chen, Telleen & Chen, 1995), psychological well-being (Center for Epidemiology Studies Depression Scale –CES-D -Radloff, 1977), family resources (Dunst & Leet, 1987), and parents' satisfaction with Head Start and transitions to kindergarten were rotated into the interview (one measure each month) throughout the study. The monthly telephone interview was developed to be brief (i.e., 10 minutes), to be easy to administer over the phone (mostly close-ended questions), and to parallel the questions asked in the FACES parent interview. Monthly telephone interviews conducted over the course of 15 months were completed, on average, with 58% of the families. Monthly response rates varied from 48% to 68% over the course of the study.
Community Agency Interviews
Telephone interviews were conducted with 200 community agencies in 10 Head Start FACES sites regarding the amount and overall nature of collaboration between their agency and the local Head Start program. The methodology and findings from this study are presented in Section V of this technical report.
Staffing and Data Collection Procedures
Case study data were collected at each of the 40 FACES sites. For each of the three waves of data collection, FACES site managers conducted the home visits as well as the home and neighborhood observations. The site managers also maintained monthly telephone contacts with the families. In order to develop and maintain rapport with the families over time, families were assigned to the same site manager for all of their interviews. Site managers also sent birthday and holiday cards to the families in their caseload to maintain contact over the year. Families were given small, child-oriented gifts at each home visit for their participation in the study. All data were quality checked and organized for data analyses by the FACES case study managers.
2.3 Data Analyses
The FACES case study integrated three methodological or analytic strategies for identifying or confirming emergent themes within and across the Head Start families in the study. The first strategy involved content coding of open-ended responses to questions posed to each of the families during home visit interviews. The second strategy involved descriptive analyses of quantitative data collected in monthly telephone interviews, with a particular focus on changes over time within families. The third strategy integrated qualitative and quantitative data from the case study with data from the FACES parent interview, teacher ratings, and child assessments, to produce a narrative for each family in the case study. The intent of these three strategies -- the family narratives, the content coding, and descriptive analyses -- was to integrate them into an overall analytic strategy to identify or confirm important themes both within and across the families in the study.
Content Coding and Analyses
The second analytic approach involved content analyses of open-ended responses from the home visit interview. Content analysis is an approach that has been used in a large number of studies to systematically organize and categorize textual information in a standardized way that allows researchers to make inferences about the information (Weber, 1990). This process involves content coding words or even sentences of text into a coding scheme or classification system with many fewer categories that are organized around the content of the text. In this case study, content codes for the home visit interviews in the spring and fall of 1997 were adapted from content codes used in a previous study that used the "Getting to Know Your Family" introductory interview instrument (Ramey and Ramey, 1992). These data were content coded by the two FACES case study managers, first separately, and then together, to refine and reach agreement on aspects of the content coding schemes and/or ways that particular text should be coded. The spring 1998 home visit interviews were content coded in the same way, except that there were no previous content coding schemes, so coding schemes were developed based on a small sample of the cases (i.e., 25%) and were adapted to the data, if necessary, during the coding process. Once all the cases had been content coded, each case was stored and codes were organized and analyzed using NUD*IST qualitative software (QSR, 1995) to identify predominant themes or issues related to particular research questions in the study. The content codes for both the fall 1997 and spring 1998 home visit interviews are found in Appendix D4.
Descriptive Analyses of Monthly Telephone Interview Data
The third analytic approach involved conducting descriptive analyses of the monthly telephone data collected while following the FACES case study families over a 14-month period of time with 12 monthly data collection points. These involved simple descriptive analyses across families as well as analyses examining change over time, such as the percentage of families each month who experienced particular health problems. A critical role of the telephone interview analyses was to examine and understand changes over time and better understand the amount of change these families were experiencing regarding the key questions of interest. Whenever possible, monthly telephone data were linked with parent interview data on key indices to enhance the ability to assess changes within these families.
Family Narratives
Narratives for the FACES case study families were developed using an iterative process integrating qualitative and quantitative descriptive data from the home visit interviews, monthly telephone interviews, as well as parent interviews, teacher ratings, and child assessments from the larger study. The first step in the process was to organize the structure and content (Exhibit 2-2) around the four primary areas or themes of the FACES case study:
- The Head Start child;
- The Head Start family;
- The family's interactions with Head Start; and
- The family's home and neighborhood.
Data from each of the families' home visit interviews, parent interviews, child assessments, and teacher ratings, as well as monthly telephone interviews were used to create a first draft narrative. After a first draft was completed, it was refined and read by the two case study managers to identify and highlight themes within each family narrative. The final narrative was used to identify emergent themes within each family or across families within each of the primary themes of the study. The emergent themes then became the organizing structure of the narrative chapter. Chapter 5 of Section IV contains several examples of final narratives completed on FACES families and demonstrates how this iterative process can become a useful vehicle for identifying and confirming emergent themes within families as well as beginning to identify themes across families. This methodology and iterative analytic process emphasizes first developing each case (family narrative) as the unit of analysis and conducting analyses and building patterns of explanations and emergent themes within each case as a building block to comparing themes across cases (Yin, 1984). With this approach, the themes drawn from the multiple cases can then become the emergent themes for the overall study.
 |
The Head Start Child
The Head Start Family
The Family’s Interactions with Head Start
The Family’s Home, Neighborhood, and Community
|
|
2.4 Limitations
There were no statistically significant differences found between the case study sample and the larger FACES sample of families on basic demographic information, such as household income, marital status, ethnicity, educational attainment, employment status, receipt of welfare, Medicaid or food stamps, or language spoken in the home. It should be noted, however, that because of the reduced sample size from the main sample, the findings presented in this section are not considered representative of the entire Head Start population. Caution should be used in generalizing these findings to the entire Head Start population.
1There were no statistically significant differences between the case study sample and the larger FACES sample of families on basic demographic information, including: household income, marital status, ethnicity, educational attainment, employment status, receipt of welfare, Medicaid or food stamps, and language spoken in the home.(back)
2Demographic data presented in this exhibit is from the FACES fall 1997 parent interviews – 117 of the 120 case study families completed parent interviews.(back)
3Approximately one half of the selected families had a 4-year-old child in the study and one half of the selected families had a 3-year-old child in the study during the 1997-1998 school year.(back)
4A complete description of the overall design of the FACES case study can be found in Appendix D1.(back)
5Immediate neighborhood was defined as within six blocks or 2 mile of the family=s home.(back)
| Table of Contents | Previous | Next |