Comprehensive school reform: A longitudinal study of school improvement in one state

We report on comprehensive school reform (CSR) reform in 48 schools over 6 consecutive years. In 1998, a total of 24 schools received CSR awards to improve student achievement. Control schools were carefully matched on 26 demographic variables to form a comparison group. Students' average performance, as represented in publicly available school report card data, increased across the combined sample of CSR and non-CSR schools in nontrivial ways. CSR and non-CSR schools, however, were not differentiated on the basis of a discriminant analysis function both in terms of achievement (Stanford 9) and demographic variables. To examine effects over time, a mixed-design multiple analysis of covariance (MANCOVA) test was performed for third- through ninth-grade scores for each performance level on the Stanford 9 achievement test (reading, language, and math). In all three performance areas, the mean NCE scores for CSR and non-CSR schools had significant within-subject effects for all 6 years. However, there were no between-subjects effects in any performance areas for CSR or non-CSR schools. The combined gains for CSR and non-CSR schools across time were notable; average math performance increased 1.0 standard deviation. Mean scores in reading and language also increased (.62 and.11, respectively), with language performance the least malleable area.