Comparing Performance of Pairs of Content-Parallel Items With Different Response Formats in a Licensure Exam

• In Event: Rating and Scoring Issues in Assessment

Abstract

This study compares the performance of standard multiple-choice items and pairwise matching items that are content parallel but differ in their response requirements (i.e., “fraternal twins”). The twin pairs were released simultaneously as evaluation items embedded in a high-stakes, computer-adaptive licensure exam. The distinguishing component of this work is the data collection design: parallel or “twin” items marked as enemies and simultaneously field tested on an authentic audience. Study results demonstrate that relative item information produced by well-written constructed response items is greater than that produced by strictly parallel multiple-choice items. The results of this study have important implications for practitioners of certification and licensure programs. This study should appeal to broad group of professions.

Authors

• Paul Edward Jones, National Association of Boards of Pharmacy

• Maria Boyle, National Association of Boards of Pharmacy

• William Finnerty, National Association of Boards of Pharmacy

• Andrew Dedes

• Joy L. Matthews-Lopez, National Association of Boards of Pharmacy