A Comparative Study On Differences Of Three-option And Four-option Multiple Choice Items

This study aims to compare the differences of three-option and four-optionmultiple choice items in terms of item and test analysis, distractor performance,test-taker perceptions and completion time.There are two tests in this study. The same content is used for the correspondingitems of the two tests. The only difference is the number of options per item. Test Aadopts four-option multiple choice items and Test B adopts three-option multiplechoice items. Three-option multiple choice items are created by removing the leastchosen distractor per item from four-option multiple choice items. A total of120sophomores majoring in English from Hunan First Normal College, divided into twogroups according to their student numbers, participate in this study. The testadministrations are based on Bachman’s (1990) counterbalanced test design to controlthe ordering and practice effects. This study has adopted a questionnaire survey tocollect test takers’ opinions on test-taking motivation, test-taking anxiety and testfairness of the item used in the test. For data analysis, this study makes full use of thestatistical software SPSS19.0.The results of this study are as follows. First, there are no significant differencesin mean score, item difficulty, item discrimination and test reliability betweenthree-option multiple choice items and four-option multiple choice items. Second, thedistractor performance changes better when deleting the least chosen distractor peritem from four-option multiple choice items. Compared with four-option multiplechoice items, three-option multiple choice items have lower proportion of distractorswith low frequency (<5%) and poor discrimination (≥0); the proportion of itemshaving two or more functioning distractors are substantially higher for three-optionmultiple choice items; three-option multiple choice items have a greater mean numberof functioning distractors per item despite having fewer distractors per item; existingdistractors change more discriminating. Third, there are no significant differencesbetween three-option and four-option multiple choice items in terms of test-takingmotivation, test-taking anxiety and test fairness. Fourth, three-option multiple choiceitems will take17.6%less completion time than four-option multiple choice items.Considering that three-option multiple choice items are more efficient to write andadminister without sacrificing the test quality, three-option per item is concluded to bethe optimal number for a multiple choice item.