GSE Publications

Document Type

Journal Article

Date of this Version

July 2004


Education is a cumulative process. Yet while students' knowledge and skills are built up over time, educational researchers are rarely afforded the opportunity to examine the effects of interventions over multiple years. This study of the America's Choice school reform design is just such an opportunity. Using 11 years of student performance data from Rochester, NY - which includes several years of data before America's Choice began working in the district - we examine the effects of America's Choice on student learning gains from 1998 to 2003. Employing a sophisticated statistical method called Bayesian hierarchical growth curve analysis with crossed random effects, we compare the longitudinal gains in test performance of students attending America's Choice schools to those of students attending other Rochester schools. Our analytical method allows us to examine student test performance over time, account for the nested structure of students within schools, and account for the very real problem of within-district student mobility.


Reprinted from Consortium for Policy Research in Education, July 2004, 66 pages.
Publisher URL:



Date Posted: 29 May 2007

This document has been peer reviewed.