Sono equidistanti le categorie di una scala Likert? Alcune risultanze di ricerca

One obvious, minimal pre-requisite in order to use Likert scores as cardinal numbers in data analysis is that they be perceived as more or less equidistant along the dimension of interest. Already 50 years ago, Johann Galtung among others raised serious doubts about the possibility that respondents actually do consider “agree” and “disagree” as equidistant from “uncertain”, given the well-known frequency of acquiescent response sets. In 1973, Benzecri maintained that the technique he had just launched on the academic market (correspondence analysis) supplied an excellent means to also control the assumption of equidistance of one of Likert’s short answers from the two adjoining ones. In the nineties, three young Italian students applied Benzecri’s technique to three different samples of respondents, and found converging results, all widely apart from the assumed equidistance. This essay replicates their survey on a larger sample of Likert scales coming from a dozen different surveys. Unlike the previous controls, the results diverge largely from one another. Yet, only a few offer a mild support to the assumption of perceived equidistance, while in their majority the results raise all sorts of further doubts about it.

Abstract licence: CC BY

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.