ERRORS IN THE CONSTRUCTION OF MULTI-CHOICE QUESTIONS: AN ANALYSIS

Abstract

Background: This study investigated the frequency of errors in the construction of multiple- choice questions used for the assessment of para-clinical sciences in medical course at the University of Western Australia. Methods: In all there were two hundred and ninety (290) questions reviewed to examine if they adhere to the uniformly accepted guidelines for writing multiple - choice questions. Of these, one hundred and fifty-one questions were already administered in examination so performance of distracters in those questions was also reviewed in addition to the guidelines for the construction of the questions. Results: 38% of the reviewed questions did not adhere to general guidelines for their construction and 12% had more than one error. A placement bias was also observed in placing correct choices. Conclusion: A number of flaws were identified in the review process which raises doubt about the use of objective test items as a true indicator of student learning. This review has led to changes in the assessment policy at Faculty level and Pre-examination reviews have been introduced to identify any flawed items before their administration in the examination. Regular series of workshops on item writing and item analysis have been arranged and a Faculty Teaching and Learning Grant were provided to create online resources on assessment.

Keywords

• Educational Assessment
• Reliability of Results
• Student Performance
• Multiple Choice Questions