A new scale for marketing research? - A comparison between Likert scales and interval scales using the Interval Agreement Approach

Likert scales are widely used in market research and remain one of the most popular methods to elicit respondents’ attitudes and opinions through surveys (Albaum, 1997). However, many researchers advocate the reassessment of the Likert scale (Li, 2013; Blaikie, 2003; Jamieson, 2004). The current use of Likert scale varies significantly from the methodology originally introduced by Likert in 1932, and as a result, some of the very core assumptions around Likert scales are questioned. For example, it is widely disputed if Likert data is interval or ordinal in nature (Blaikie, 2003; Jamieson, 2004). This dissertation suggests the use of an interval scale. Built on fuzzy set theory, the Interval Agreement Approach (IAA) is introduced, which allows the modelling of the intervals as input data into aggregated and more easily interpretable IAA models. This method captures uncertainty and agreement among the respondents. In order to compare the scales in a practical marketing application, Thync, a neuro-enhancing consumer technology, was chosen as an example. 109 German students participated in the study asking about their attitudes towards Thync, answering the same questions for both scales. The extended Unified Theory of Acceptance and Use of Technology (UTAUT 2; Venkatesh et al., 2013) served as a theoretical basis and provides a widely accepted methodology for the analysis of the differing datasets. The aim of the dissertation is to evaluate data quality, results and respondent feedback. After statistical analyses, the Likert and interval data seem to lead to the very similar results, but the intervals are not as easy to statistically process as the Likert data. Especially for the hypothesis testing, there was no possibility within the scope of this dissertation to perform such statistical procedures in an easy and time-efficient way. However, the interval scale and IAA show significant advantages in some areas. The flexibility for respondents to select several values within one interval is not only appreciated from the respondents’ perspective, but also offers more granular data. The resulting IAA models are easier and more intuitive to interpret on a graphical level. Also, they show more clearly the differences in opinions between respondent groups. Because of the substantial criticism directed towards Likert scales, the dissertation concludes that interval scales and the IAA have huge potential for marketing research. At the moment, the IAA models would be a useful tool for descriptive analyses, for example in commercial market research or public opinion research. Going forward, methods should be developed in order to enable easy, quick and accessible statistical processing of intervals