Education | Educational Assessment, Evaluation, and Research
Our analysis shows how a longitudinal data mart can provide a simple and effective way to analyze student test performance over time. Our data mart in this case is a mega-table compiled from several years of archival student-level test data, where we have modified all of the fields so that they have a common meaning over time. Using this longitudinal data base we then compared the performance statistics and effect sizes of test results in math and reading on the Connecticut Mastery Test (CMT) series for grades 4, 6 and 8 and on the Connecticut Academic Performance Test (CAPT) in grade 10, from 2000 to 2007. We found that students tested sequentially in grades 4, 6, 8 and 10 achieve better performance in mathematics and reading at the State, ERG and school district levels, as compared to new incoming students who began the testing sequence sometime after grade 4. This suggests that mobility relates to lower student performance on our tests, a finding that others have reported (Bourque, Mary D., 2008, Rumsberger, 2002). We conclude that student mobility should be monitored and that academic and/or social interventions may be warranted. We also conclude that a longitudinal data mart may provide a practical way to look at student test performance over time particularly when vertical scaling or vertical modulation are not available. A data mart could also serve as a simple low-tech way to cross-validate results from these techniques.
Mooney, Richard F. and Beaudin, Barbara Q., "Using a Longitudinal Data Mart to Examine the Effects of Student Mobility on Test Performance Over Time" (2008). NERA Conference Proceedings 2008. Paper 16.