Unequal Childhoods - Annette Lareau [209]
Across the two sites, only one mother refused participation outright (some fathers eventually declined). In addition, two or three parents who had initially agreed to be interviewed were not interviewed because of scheduling difficulties. Overall, however, the response rate for families was more than 90 percent. I still did not have a sample with which I felt comfortable, however. Despite their racially diverse character, the two schools did not include enough middle-class Black and poor white children for the study. I broadened the sample to include third-graders at Swan and also recruited through fliers and informal networks.9 The bulk of the interviews with parents took place in 1993 and 1994, but some were not completed until 1997. An additional sixty interviews with educators and other adults working with children were also conducted. For example, the third-grade and fourth-grade teachers at the schools were interviewed (including in Lawrenceville), along with other school personnel (i.e., reading specialists, music and art teachers, the school nurse, bus drivers, and yard duty teachers) as well as numerous providers of children’s leisure services. Where possible, these interviews were focused on the children in our study. (A chronological overview of the study is presented in Table C9, Appendix C.)
RECRUITING THE FAMILIES
The classroom observations and interviews with parents were crucial in gaining access to families for the observation phase. Ms. Green’s third-grade classroom at Lower Richmond yielded seven of the twelve children: Brindle, Taylor, Irwin, Driver, Carroll, Yanelli, and McAllister. The fourth-grade classroom at Swan yielded two more: Tallinger and Handlon. Selecting the families involved a conscious, complicated calculus. The interview phase had helped identify certain kinds of experiences and family traits (especially the number of organized activities, the strength of kinship ties, and the depth of family-school relationships) as broadly characteristic of each social class. I wanted most of the families we observed to be as representative of these traits as possible. We tried to avoid selecting children whose parents were either unusually active in school or unusually quiet in their interactions with teachers. Among the middle-class families, we further limited the pool to only those with two parents in the home. Thus, in most cases, the research assistants and I had only three or four children per category to choose among.10
In making the final choices for the observational phase, I wanted to balance the sample by gender, race, and class. I wanted, as well, to look at children who, although from different races and social classes, shared key characteristics. For example, despite different class locations, some children shared church involvement, extended family nearby, or participation in organized activities. Overall, the research team and I tried to “mix and match” the children we chose in order to lessen the chance that differences in the behavior we observed were connected to some unknown variable such as parental involvement in the school.
I also sought “deviant cases.” In particular, I very much wanted to include a middle-class child who participated in no organized activities. I was unable to find a single such child among the families we interviewed, nor could Swan teachers or parents think of a possible candidate. I was more successful in finding deviant cases in terms of child-rearing strategies and family location, such as families with middle-class characteristics who live in working-class or poor neighborhoods. Two families offered this form of contrast. The Irwins, a deeply religious