The Effect of Option Differences on Psychometric Properties of Items in Likert-Type Scales

Abstract

Likert-type scales are often used in education and psychology. In Likert-type scales, response options/categories, like items, are expected not to direct individuals’ responses. Although the researchers themselves make decision on how to arrange categories during scale development, it is possible that different categories reveal different response behaviors. In the literature, it has been observed that differentiations in the number of categories of forms are studied more, yet there are a limited number of studies investigating the middle category in the forms with different labels. Furthermore, it has also been observed that there are limited number of empirical studies conducted based on polytomous Item Response Theory. This study, which was conducted to close this gap in the literature, was carried out with 377 students. The options of the attitude scale were denominated with different labels, and thus four different forms were generated. Only the middle category names were changed in the first three forms, and in the fourth form, the categories were graded. The data obtained from the forms were analyzed using the Graded Response Model and the Generalized Partial Credit Model depending on Item Response Theory. After the examination of reliability of the forms, the parameters in these forms, and the relationships between the parameters according to both models, inferences were made as to how the differences of the middle category in the forms had an effect on the perceptions of individuals.