vol 24 no3 A, 2024

Examination of the effects of correct alternative position of multiple-choice test items on differential performance of items using the Mantel-Hansel method

Hits: 344

Examination of the effects of correct alternative position of multiple-choice test items on differential performance of items using the Mantel-Hansel method

Abstract:                                                                                                                                                                          

The study seeks to examine the impact of the position of the correct option in multiple-choice test items on differential item functioning (DIF) by employing the Mantel-Haenszel method. This method is one of the most commonly used approaches for identifying DIF and remains one of the most popular techniques for detecting bias in test items today. In order to achieve the objectives of the study, the study involved constructing an achievement test in educational measurement and evaluation, consisting of multiple-choice questions with four alternatives and a total of 34 items. The test included three forms with the same content, and the variable among these forms was the position of the correct alternative. The first form of test shows the location of the correct alternative on alternatives a, b, and the second form has the location of the correct alternative on alternatives c, d, and the third form is randomly distributing the correct alternative to the various four alternatives. The Mantel-Haenszel method was used through SPSS to detect the differential performance of the items for each of the three test models.The results of the study showed that the percentage of items that showed a differential performance in favor of the first test model was 13%, in favor of the second model, 23%, and in favor of the third model, 28% of the total number of items in the test. The results also showed that there were no statistically significant differences between the arithmetic averages of the ability parameter estimates for individuals due to the location of the correct alternative.The researcher recommended that when constructing a multiple-choice test, it is better to randomly distribute the correct alternative to the four different alternatives. It also recommended conducting more studies on the detection of differential performance in the light of other variables and in other ways, including the method that relies on the item response theory or the area between the curves.

Keywords: differential performance, Mantel-Hansel method, location of the right alternative, classical theory of measurement.

Download