The Effect of Item Sequence on Bar Examination Scores.

摘要

Large scale testing program can reduce the likelihood of one examine copying another's answers by: having sufficient distance between seats, having adequate proctoring, varying answer sheet format, and using multiple test forms. The use of multiple forms usually involves having one form contain one set of items and the other forms contain different sets of items. Although this strategy may be sound in terms of psychometric standards, it may be inconsistent with the policies of the organization sponsoring the testing program. For instance, the National Conference of Bar Examiners requires that all examines taking the Multistate Bar Examination (MBE) answer the same set of questions. One solution to the forgoing problem is to use multiple test forms that differ in terms of the order in which the items appear. This strategy is consistent with the policy of having all examines answer the same items and it would substantially reduce the opportunity for cheating. There are two major concerns with this approach: (1) some sequences may be easier than others thereby giving some examinees an unfair advantage and (2) it might change the characteristics of items used for equating tests across administrations. The present study was conducted to determine whether varying the sequence in which blocks of items were presented to examinees would affect tests and/or item characteristics.