Pennsylvania

The research design for this project is both longitudinal and cross-sectional, and is designed to generate evidence that is associative/correlational, and involves secondary analysis of a national database, the National Center for Education Statistics’ (NCES) nationally representative Schools and Staffing Survey (SASS) and its supplement the Teacher Followup Survey (TFS).

Analysis involves longitudinal comparisons of nationally representative cross-sectional samples of subgroups of teachers, schools, and districts (within and across states) before and after the advent of NCLB. We utilize multivariate regression methods—primarily multinomial logistic regression. These will involve analyses of the relationships between a range of measures of school accountability and also a range of measures of teacher preparation and the likelihood of individual teacher turnover, while controlling for the background characteristics of districts, schools, and teachers. The dependent variable, teacher turnover, will be a trichotomous variable based on whether each teacher remained with their school, moved to another school, or left teaching in the year of the survey. The independent variables for the regression analyses will be drawn from the extensive SASS data obtained by the survey questionnaires.