Incorporating respondents’ perceptions into ordinal scales. Measuring the confidence in political leaders

Abstract

Ordinal scales play a crucial role in the measurement of confidence in politicians. A common approach to investigate the status of public opinion involves conducting surveys that utilize ordered response categories. Evenly assigning scores to response categories is a common technique when analyzing survey results, whether done intentionally or implicitly. When doing so, it is usually assumed that respondents perceive equal proximities between the different response categories. Caution is essential when assigning scores, since this is most likely not true. To address this concern, this article proposes the application of a scoring function based on the concept of ordinal proximity measure. This function assigns scores to response categories of the ordinal scale considering respondents’ perceptions of closeness between response categories. These methods were applied to a survey on confidence in the Spanish Prime Minister by the Spanish Sociological Research Center (CIS). A small exploratory survey was conducted on perceived proximities between response categories in the CIS survey and aimed to replicate results using statistical regression models, highlighting potential biases when ignoring perceptions of response category proximities. While the findings offer insight into the method’s analytical implications, they are illustrative in nature and should not be overinterpreted as generalizable.