In lieu of an abstract, here is a brief excerpt of the content:

195 Appendix B Measurement Details and Technical Issues T HIS APPENDIX provides detail on the Beginning School Study Youth Panel (BSSYP or Youth Panel) research design and elaborates on the measurement of key constructs not covered adequately in the text. We also provide a nontechnical overview of multiple regression analysis, the statistical technique used in chapter 7. Youth Panel Sample and Research Design The Youth Panel is a prospective longitudinal study of children’s academic and social development beginning in first grade and continuing through high school graduation and beyond. Data collection began in 1982 and concluded in 2005–2006 when 80 percent of the original group was interviewed at ages twenty-eight and twenty-nine. In tracking the sample’s progress through school we count time two ways, by grade or by year. All were in first grade in the fall of 1982, the project’s first year, but repeaters fall behind. In this volume, grade refers to the modal grade of enrollment. In 1982, a two-stage probability sample of children beginning first grade in the Baltimore City public schools (BCPS) was selected for study. Probability sampling is difficult to implement, but provides internal validity lacking in other study designs (Michael and O’Muircheartaigh 2008). First, a random sample of twenty schools was selected from a list of public elementary schools in the city, stratified by racial mix (six predominantly African American, six white, eight integrated) and by socioeconomic status (fourteen inner-city or working class and six middle class). The African American enrollment averaged 99.5 percent in the six African American schools, 6 percent in the white schools, and 48 percent in the eight integrated schools (ranging from 17 percent to 87 percent). Second, within each school, students were randomly sampled from the previous spring’s kindergarten enrollment, supplemented by fall class rosters to pick up new entrants. Over the summer, parents were visited in their homes to obtain permission for their child’s participation. Three percent declined. The final sample consists of 790 students beginning first grade for the first time (nonrepeaters) in the fall of 1982. The city public school system then was 196 Appendix B about 77 percent African American (U.S. Bureau of the Census 1983), but to sustain comparisons across social lines, the Youth Panel oversampled whites. At baseline, 55 percent of students were African American. Parents’ educational levels averaged 11.9, with a standard deviation of 2.6. Broken down by race, African American parents had a slightly higher mean level of education than whites (12.1 years versus 11.7), but whites were overrepresented at the high and low ends of the distribution: for example, among white and African American mothers, 14.7 percent versus 10.3 percent college graduates; 43.7 percent versus 33.9 percent high school dropouts. Sixty-seven percent of the families received free or reduced price meals at school, 77 percent of African Americans and 53 percent of whites. Overall, 56 percent of the sample lived in two-parent households at the beginning of first grade, 70 percent of whites and 44 percent of African Americans. Data Collection Procedures Beginning in the summer and fall of 1982, data were collected by face-to-face interviews with students and parents. Teachers responded to questionnaires and school records were examined for data on marks, test scores, and the like. In later years most parent questionnaires were answered by mail or by phone, but students were interviewed face to face. Parents were surveyed in the summer or early fall annually from year 1 through year 11 with the exception of year 5. Parent and student fall interviews always took place before the end of the first marking period and spring interviews took place between the third and fourth (final) marking periods. The timing of student and parent questionnaires ensured that measurement of parents’ and children’s mark expectations preceded receipt of report cards. Students were interviewed in the fall of 1982 and spring of 1983, and fall and spring of the following year (1983–1984). They were then interviewed twice in their fourth and sixth years (1985–1986, 1987–1988), the sixth year being the first year of middle school for most. Those who were delayed making the elementary to middle school transition were interviewed twice in year 7. For the others, an abbreviated interview was conducted in the spring of year 7. Thereafter, spring interviews were conducted annually through year 13.1 Teachers filled out three...

Share