An Empirical Study of Response Category Effects: A Rasch Rating Scale Analysis

• In Event: Measurement Issues in Survey Research

Abstract

This study attempts to explore how changes in survey response categories affect data interpretations, response patterns, and scale psychometric properties from the aspect of reliability and validity: instrument dimensionality, item properties, and scale qualities through applications of Item Response Theory. Specifically, it aims to investigate the functioning of varied item response categories by utilizing the Rasch rating scale analysis to examine the two subsets of data collected from two survey forms composed of 3-, 4-, and 7-point rating scales. The results suggest that the form with three response categories provides better measurement quality and more reliable interpretations while the 7-point rating scales contained more apparent unequal-length intervals between the ordered response categories and the direct interpretations were misleading.

Author

• Zongmin Kang, DePaul University