The test construction literature has emphasized the randomization of answers within multiple-choice items as a way to eliminate test construction bias. This study examines the answer placement bias of the multiple-choice portion of the 1987 - 1992 CIA, CMA, and CPA exams. The results indicate a consistent bias toward the "A" position in the CIA exam and against the "A" and "E" positions on the CMA exam. Also, all three exams were found to have significantly fewer correct answer runs (e.g., four "A's" in a row) than would be expected from the random placement of answers. A discussion of test item construction and answer placement techniques is also presented.

Document Type


Publication Date

Fall 1997

Publisher Statement

Copyright © 1997 Marshall A. Geiger. First publication rights: Accounting Educators' Journal. This article first appeared in Accounting Educators' Journal IX, no. 2 (Fall 1997): 89-108.

Please note that downloads of the article are for private/personal use only.