Assessing School-Level Performance With Differing Data Points in Cohorts

• In Event: 29.036 - Marvelous Multilevel Modeling

Abstract

Reading achievement data from two longitudinally matched students cohorts were analyzed with hierarchical linear models to provide a systematic approach to evaluate school performance in a given year by using student level data from multiple cohorts with varying data points. The data came from 1,198 public schools, with a student population of 164,460. The initial results indicated that the results from the individual cohort models were largely preserved in the multiple cohort analysis. The findings of this study will help us to more accurately evaluate schools using the student level growth data on an annual basis.

Authors

• Nedim Yel, Arizona State University

• Roy Levy, Arizona State University

• Yixing Liu, Arizona State University

• Ann C Schulte, Arizona State University - Tempe