Dispelling Three Myths about Likert Scales in Communication Trait Research

Abstract

Three common myths about multi-item Likert scales are dispelled because (a) a multi-item Likert scale consists of interval quality data even when individual Likert response items are ordinal data, (b) parametric statistics are appropriately used to analyze data provided by a Likert scale, and (c) holistic Likert scale analyses for well-performing scales takes precedence over analyses at a subscale level. Validity of communication trait measures that perform well is best improved by demonstrating that altering scales leads to significant enhancement of criterion validity, not by using limited approaches that fail to account for the interconnectedness of Likert scale items in, ultimately, influencing participants' responses.