The Stability of Cut Score Estimates from the Beuk Standard Setting Method

• In Event: Performance Levels and Standard Setting

Abstract

We examined the stability of cut scores from the Beuk method. We studied the effects of panelist agreement and score distributions, and we introduce a bootstrapping procedure to estimate standard errors. We simulated varying panel disagreement, sizes, and number of bootstrap replications. The results provide guidance about achieving optimal stability.

Authors

• Joe Grochowalski, College Board

• Lauren Molin, College Board

• Lei Wan, College Board

• Amy H. Hendrickson, COLLEGE BOARD