A cross-sectional study on item analysis of prevalidated and nonvalidated anatomy multiple-choice questions

Abstract

Background: Multiple-choice questions (MCQs) are widely used tools to test knowledge. Pre- and postvalidation of the MCQs are prerequisites to maintaining the standard of the questions. Prevalidation is done by subject experts before the test. Postvalidation (item analysis) uses three indices – item difficulty index (P), item discrimination index (D), and distractor efficiency (DE). This study was done with the aim to assess the quality of MCQs used during the formative assessments using item analysis indices and to compare the indices in prevalidated and nonvalidated MCQs. Methodology: The study used two sets of MCQ tests, conducted as a part of formative assessments for phase I MBBS students in the department of anatomy. Set 1 comprised nonvalidated MCQs and set 2 comprised prevalidated MCQs. The three indices were calculated for all the questions in both sets. Results: The average P value for set 1 was 57.62 ± 21.90 and that for set 2 was 59.27 ± 20.32. Average D value for set 1 was found to be 0.23 ± 0.36 and that for set 2 was 0.29 ± 0.15. In set 1, 37.5% of the MCQs had 100% DE, whereas, in set 2, 60% of the MCQs had 100% DE. Thirty percent of the distractors were nonfunctional in set 1, but in set 2, only 16.7% of the distractors were nonfunctional. Conclusions: Study results favor prevalidation and postvalidation of MCQs to develop a good-quality MCQ bank. It is imperative to use the MCQs having moderate difficulty, good discrimination power, and with all functional distractors in any given assessment.