January 1, 2004
Using Hierarchical Growth Models to Monitor School Performance Over Time: Comparing NCE to Scale Score Results
Authors:
Pete Goldschmidt, Kilchan Choi and Felipe Martinez
Monitoring school performance increasingly uses sophisticated analytical techniques and we investigate whether one such method, hierarchical growth modeling, yields consistent school performance results when different metrics are used as the outcome variable. Specifically we examine whether statistical and substantive inferences are altered when using normal curve equivalents (NCEs) vs. scale scores. The results indicate that the effect of the metric depends upon the evaluation objective. NCEs significantly underestimate absolute growth, but NCEs and scale scores yield highly correlated (.9) results based on mean initial status and growth estimates. Correlations between NCE and scale score rankings, based on fitted school initial status and growth values are generally over .94. Further, statistical and substantive results, using NCEs and scale scores, pertaining to school-wide program effects are highly correlated (.95) as well. NCEs and scale scores matched 99% of the time on whether or not the program indicator variable was statistically significant.
Goldschmidt, P., Choi, K., & Martinez, F. (2004). Using hierarchical growth models to monitor school performance over time: Comparing NCE to scale score results (CSE Report 618). Los Angeles: University of California, Los Angeles, National Center for Research on Evaluation, Standards, and Student Testing (CRESST).