In many school systems (in the United States and elsewhere), student test scores are used to rank schools and, in some systems, they are used to estimate teacher effectiveness. In a recently published article in Journal of Applied Statistics, Alder Graduate School of Education Senior Research Scientist Dan Wright compares two statistical models that are used for this purpose. The student growth percentile (SGP) model has problems when there are variables that influence effectiveness and student scores, contends Wright, whereas a gain score model performs better in these situations, but has difficulties in other situations. The main plea from the paper is that if these (and other) models are to be used, then it is important to evaluate them with simulated data specific to the situation.

To learn more about Wright’s article “Estimating school effectiveness with student growth percentile and gain score models,” click here or email Dan directly at dwright@aldergse.org.