Analysis of one-best MCQs: the difficulty index, discrimination index and distractor efficiency.

OBJECTIVE To investigate the relationship of items having good difficulty and discrimination indices with their distractor efficiency to find how 'ideal questions' can be affected by non-functioning distractors (NF-Ds). METHODS The cross-sectional study was conducted at Fatima Jinnah Dental College, Karachi, during Jan-Jun 2009, with 102 First Year dental students (17-20 years). Physiology paper of the first semester, given after 22 weeks of teaching general topics of physiology, was analysed. The paper consisted of 50 one-best MCQs, having 5 options each. The MCQs were analysed for difficulty index (p-value), discrimination index (DI), and distractor efficiency (DE). Items having p-value between 30-70 and DI > or = 0.25 were considered as having good difficulty and discrimination indices respectively. Effective distractors were considered as the ones selected by at least 5% of the students. RESULTS The mean score was 27.31 +/- 5.75 (maximum 50 marks). Mean p-value and DI were 54.14 +/- 17.48 and 0.356 +/- 0.17, respectively. Seventy eight per cent items were of average (recommended) difficulty (mean p-value = 51.44 +/- 11.11) and having DE = 81.41%. Sixty two per cent items had excellent DI (0.465 +/- 0.083) with DE = 83.06%. Combining the two indices, 32 (64%) items could be called as 'ideal' (p-value = 30 to 70; DI > 0.24) and had DE = 85.15%. Overall 42% items had no Non-functioning Distractors (NF-D), while 12% had 3 NF-Ds. Excellent discrimination (DI = 0.427) was achieved with items having one NF-D, while items with 2 NF-D and no NF-D had nearly equal but lower DI (0.365 and 0.351 respectively). CONCLUSION One-best MCQs having average difficulty and high discrimination with three functioning distractors should be incorporated into future tests to improve the test score and properly discriminate among the students. Items with two NF-Ds, though easier, are better discriminators than items with no NF-D.