ERIC Number: ED456153
Record Type: Non-Journal
Publication Date: 2001-May-1
Reference Count: N/A
Detecting a Change in School Performance: A Bayesian Analysis for a Multilevel Join Point Problem. CSE Technical Report 542.
Thum, Yeow Meng; Bhattacharya, Suman Kumar
To better describe individual behavior within a system, this paper uses a sample of longitudinal test scores from a large urban school system to consider hierarchical Bayes estimation of a multilevel linear regression model in which each individual regression slope of test score on time switches at some unknown point in time, "kj." Preliminary evidence based on a sample of 58 elementary schools in Chicago, Illinois, suggests that change points in test score trends indeed differ from school to school. In addition the estimated posterior distribution of the change points suggests that, while the estimated timings of change in performance do not contradict the claim that a well-publicized intervention at time "t" may have been a contributive factor, changes have not been uniformly positive and do require further scrutiny. Additional results are explored using models that accommodate case weights and shorter time series. (Contains 8 figures, 3 tables, and 35 references.) (SLD)
Descriptors: Bayesian Statistics, Educational Change, Educational Trends, Elementary Education, Estimation (Mathematics), Longitudinal Studies, Urban Schools
CRESST/CSE, Graduate School of Education & Information Studies, University of California, Los Angeles, CA 90095-1522. Tel: 310-206-1532.
Publication Type: Numerical/Quantitative Data; Reports - Research
Education Level: N/A
Sponsor: Office of Educational Research and Improvement (ED), Washington, DC.
Authoring Institution: National Center for Research on Evaluation, Standards, and Student Testing, Los Angeles, CA.; California Univ., Los Angeles. Center for the Study of Evaluation.