How Good is Good Enough?: A Comparison of Three Methods for Establishing Cut Scores on Placement Tests

Abstract

The purpose of this study was to determine whether three standard-setting methods would produce sufficiently consistent results for placement decisions. The cut scores were derived from modified Angoff method, borderline group method, and cluster analysis. The results indicate that the cut scores derived from the three standard setting methods did not entirely agree with each other in assigning students into different levels, suggesting that the choice of standard setting method influenced the resulting cut scores. Specifically, the borderline-group method tended to produce lower cut scores than the other two methods. The cluster analysis yielded two cut scores, which were very similar to the cut scores derived from the modified Angoff method at the corresponding levels. The implications of the findings are discussed and avenues for further study are suggested.