The effect of presenting geometry items with and without shapes on the psychometric properties of the test and students' test scores

In this research, the effects of presenting geometry items with and without shapes on the psychometric properties of the test and students' test scores were examined. The study was conducted with a total of 480 eighth grade students in Mardin province in the 2023–2024 academic year. Within the scope of the study, two geometry tests were crafted, one with shapes and the other without shapes but only verbal expressions, and both tests consisted of 15 multiple-choice items. In the data collection process, a counterbalanced design was followed and the two tests were administered to the students three weeks apart. Analyses were performed on 405 students who participated in both applications and whose test forms could be matched. The results of the factor analysis revealed that the factor loadings of the items and the ratio of extracted variance were higher for the test with shapes compared to the shape-free form. The Cronbach's alpha internal consistency coefficient of the test containing shapes was found to be significantly higher than that calculated for the shape-free test. According to item difficulties, it was determined that the questions with shapes were easier for the students than the questions without shapes. In terms of discrimination indices based on the point-biserial correlation coefficient and 27% lower-upper group comparisons, a difference in favor of the shape-containing test was observed in almost all items. In parallel to this, the Ferguson delta statistic, which is a measure of discrimination for the entire test, was higher in the test with shapes. Correlation analysis denoted a strong, positive and statistically significant relationship between students' scores on the two tests. In addition, the paired sample t-test proved that there was a statistically significant difference between students' scores on tests with and without shapes. These results indicate that the geometry tests with and without shapes differ in terms of both psychometric properties and students' scores; therefore, these tests cannot be characterized as parallel forms even if they are intended to test the same objectives.