Scale matters: unravelling the impact of Likert scales on political self-placement

Abstract

Worldwide, Likert scales are used to measure ideology of both public opinion and individual voters, with the utilization of a self-assessment numerical scale as the most widely accepted tool, despite the multifaceted nature of the concept. However, in contrast to this prevailing consensus, there is a lack of agreement regarding what particular scale should be used and about the benefits of the different scales. This paper seeks to provide insights into these questions by analysing Spanish public opinion ideology as measured by CIS surveys. In 1976, the Centre for Sociological Research (CIS), a leading public body for studies of Spanish society, defined the ideological self-placement scale as being from 1 to 10. Since then, a myriad of different scales has been applied by CIS in more than 1400 surveys (over 3.5 million interviews). In total, we identify 20 scales in studies with available microdata, with the number rising to 28 when all the CIS studies carried out are included. This research identifies which scale is used in each study and analyses whether this has an effect on the estimation of the ideological distribution of the population. After studying, from an empirical perspective, the relationship structures between the different scales, we conclude that the scale used does have an effect. We finish the paper by offering clear guidelines on how questionnaires related to ideology should be asked.