Changing the Scale: The Effect of Modifying Response Scale Labels on the Measurement of Personality and Affect
Multivariate Behavioral Research
Much research in psychology is based on self-report questionnaire data using items with Likert-type response scales. Often the same items are administered with different response scale labels in different studies. Using measures of personality and affect, the effect of type of label (bipolar or unipolar) on the categorical item responses was investigated with the methods of item response theory (IRT). In two studies, the effect of type of label was examined in the context of all options labeled and only endpoint options labeled. In Study 1, we found that when every number of a response scale is labeled, the responses to the same items differ between bipolar (agree-disagree) and unipolar (not at all–very much) labels. Study 2 showed that these differences are not observed when only the endpoints are labeled. The findings are discussed in terms of their implications for measurement and research reporting of personality, clinical, health, social, and other psychological constructs. IRT methods offer a way to increase our understanding of the psychological processes underlying answering questions.
National Science Foundation
differential item functioning, item response theory, Personality measurement, rating scale format, response scale labels
Lynne Steinberg and Altovise Rogers. "Changing the Scale: The Effect of Modifying Response Scale Labels on the Measurement of Personality and Affect" Multivariate Behavioral Research (2022): 79-93. https://doi.org/10.1080/00273171.2020.1807305