Modeling the Growth of PSAT and SAT Mathematics Performance

• In Event: Roundtable Session 60
  In Roundtable Session: Approaches to Measuring Growth

Abstract

Identifying individual growth patterns and finding the factors affecting such growth processes have been major concerns of educators because understanding these patterns might help educators detect students who will eventually show lower academic achievement, thereby more accurately predicting the academic performance of matriculating college students. The current study examines, using multilevel modeling, individual growth trajectories of large scale mathematics scores and the effect of the demographic and academic background of the test takers on their growth parameters. Next, qualitatively different subpopulations in terms of individual growth patterns are examined using a growth mixture modeling approach. Finally, the predictive validity of the large scale assessment is evaluated by incorporating the outcome variable, academic performance in college into the growth mixture models.

Authors

• YoungKoung Kim, The College Board

• Judit Antal, The College Board

• Thanos Patelis, The College Board