Psychometric Properties, Item Difficulty and Ability Parameters of a Multiple Choice Test Estimated According to the Number of Options Using the 3-Parameter Model

Abstract

This study aimed at investigating the effect of using a number of alternatives in multiple choice tests on the students’ ability, and the estimation of the item difficulity. To achieve the aim of the study, three forms of a multiple choice test consisting of 32 items were constructed to measure students’ achievement in mathematics. The forms were different in the number of options. Each item was constructed to have four, three or two alternatives. The sample of the study consisted of 1500 male students in the 8th grade in schools in Jordan during the 2016/2017 academic year. Each of the three test forms was administered to 500 students. The three test forms were administered and scored objectively. The statistical Package for Social Sciences )SPSS( and the BILOG-MG3 programs were used to analyze the data, provide estimates of the individual's abilities, item parameters, standard error of measurement, and test reliability for each test form. The findings of the study indicated statistically significant differences (α= 0.05( between the discrimination parameter mean scores due to the test form in favor of the four-option alternative. Also, the findings showed statistically significant differences between the values of the reliability coefficents in favor of the four- and three-option test forms. As for the results related to the accuracy of the ability parameter estimate, the findings of the study indicated statistically significant differences in favor of the four-, followed by the twoand three option forms of the test.