Comparison between Long-Menu and Open-Ended Questions in computerized medical assessments. A randomized controlled trial

BACKGROUND: Long-menu questions (LMQs) are viewed as an alternative method for answering open-ended questions (OEQs) in computerized assessment. So far this question type and its influence on examination scores have not been studied sufficiently. However, the increasing use of computerized assessments will also lead to an increasing use of this question type. Using a summative online key feature (KF) examination we evaluated whether LMQs can be compared with OEQs in regard to the level of difficulty, performance and response times. We also evaluated the content for its suitability for LMQs. METHODS: We randomized 146 fourth year medical students into two groups. For the purpose of this study we created 7 peer-reviewed KF-cases with a total of 25 questions. All questions had the same content in both groups, but nine questions had a different answer type. Group A answered 9 questions with an LM type, group B with an OE type. In addition to the LM answer, group A could give an OE answer if the appropriate answer was not included in the list. RESULTS: The average number of correct answers for LMQs and OEQs showed no significant difference (p = 0.93). Among all 630 LM answers only one correct term (0.32%) was not included in the list of answers. The response time for LMQs did not significantly differ from that of OEQs (p = 0.65). CONCLUSION: LMQs and OEQs do not differ significantly. Compared to standard multiple-choice questions (MCQs), the response time for LMQs and OEQs is longer. This is probably due to the fact that they require active problem solving skills and more practice. LMQs correspond more suitable to Short answer questions (SAQ) then to OEQ and should only be used when the answers can be clearly phrased, using only a few, precise synonyms. LMQs can decrease cueing effects and significantly simplify the scoring in computerized assessment.