Likert Items: Should(n’t) We Really Care?

Abstract

One of the controversial methodological topics in the social and behavioral sciences is the (ab)use of Likert Scale items, Likert-type items and ranked ordered response categories. The debate is whether parametric tests can be legitimately conducted on technically ordinal response categories that are represented with numbers. Participants answered survey questions on moral disengagement, where we changed the intervals of seven response categories and tested whether assigning numbers made any difference in two separate studies. The results showed that participants’ ratings were not significantly different with or without numbers. Participants tend to covertly superimpose numbers where none were provided. Also, there were no significant interactions between assignment of numbers and ‘intervalness’. However, ratings were significantly different between two key interval groups. Knowing the assumptions of respondents to these Likert items even without numbers could inform researchers especially if parametric tests are to be conducted.