Assessing difference: examining Florida’s initial teacher preparation programs and exploring alternative specifications of value-added models

This study explores important statistical issues on the appropriate functional form and model specification of the value-added educational achievement equation. We also wish to estimate the causal effect of a teacher’s institution of academic preparation and pedagogical training. Standardized test scores, viz., the Florida Comprehensive Assessment Tests (FCAT), provide a measure of pupil academic achievement. Accordingly, this study uses a value-added regression model to establish whether there is a “college preparation effect” on the average pupil’s FCAT reading and mathematics scores. We find that value-added regression analysis fails to uncover robust and substantive college preparation effects. Regardless of race (African American, Hispanic, or white), male or female status, or FCAT mathematics versus FCAT reading, pupil academic achievement does not vary substantively according to a teacher’s college of preparation. Further, the statistical significance of teacher program effects also depends on the functional form and specification of the value-added model.