Investigating the Effect of Missing Data on the Experimental Test of Mathematics for the Second-Secondary Students

Abstract

This study aimed to uncover the effect of the missing values using the Rasch model on an experimental test of mathematics for the second-secondary students. To achieve the goal of the study, a multiple-choice test was designed testing all basic skills found on the student’s book. The test consisted of 103 items. This study was conducted in September 2019. The results were analyzed using SPSS, Winstep, and MG3-BILOG. To determine the correlation coefficients, the results showed that the test had good psychometric properties and the test had high reliability, validity indications, acceptable difficulty, and discrimination coefficients. The findings also showed that the differentiation coefficients of items were equal. The results also showed that the premises of the Rasch model were fulfilled. Further, the results showed that there were statistically significant differences between the three models according to the percentage of data loss, standard error, difficulty, internal and external matching, and bias, and were in favor of the original model before missing the values.