الملخص
Objective:To discuss the possibility of modifying nonfunctional distractors to improve the quality of multiple-choice questions in medical examination through analyzing the changes in the difficulty and discrimination of items and the selection rate of modified distractors.Methods:Thirty-two multiple-choice questions involving nonfunctional distractors from a medical examination were studied. According to the item-writing guidelines, experts modified nonfunctional distractors based on actual measurements. We analyzed the changes in the difficulty and discrimination of items and the selection rate of modified distractors before and after item modification using the paired Wilcoxon test, and investigated the correlation of item difficulty and discrimination before and after item modification by Spearman's rank correlation analysis.Results:Six questions were modified due to technical factors, and 26 questions were modified due to weak interference in knowledge content. Before and after nonfunctional distractor modification, the difficulty values of the questions were 0.676 (0.558, 0.893) and 0.637 (0.531, 0.839), respectively; the discrimination values (point-biserial correlation) were 0.261 (0.150, 0.316) and 0.262 (0.138, 0.358), respectively; and the discrimination index values were 0.215 (0.113, 0.352) and 0.259 (0.138, 0.346), respectively. There were significant differences in the difficulty and discrimination index of items before and after modification. The difficulty and discrimination of items before and after modification were both significantly correlated. The selection rates of modified distractors were 0.009 (0.003, 0.015) and 0.044 (0.021, 0.092) before and after modification, respectively, which were significantly different.Conclusions:Nonfunctional distractors in this study were mainly caused by reasons in the contents of the test questions, less in technical defects. Through the analysis and modification of nonfunctional distractors, the selection rate of nonfunctional distractors can be effectively increased, and item difficulty and discrimination can be improved.