Distractors--can they be biased too?

Abstract

Numerous work has been done on item bias and differential item functioning. Although there is some research on distractor analysis, no detailed study has been attempted to examine the way distractors in an item function, with regards to comparing distractor performance. This paper examines how distractors function differentially and compares various methods for identifying this. The Pearson chi-square, likelihood ratio chi-square and Neyman weighted least squares chi-square tests are some of these methods. Possible causes of distractor bias are discussed with illustrations from a physics problem-solving scale.