Data Processing by Fuzzy Methods in Social Sciences Researches. Example in Hospitality Industry

Abstract

Likert-type scales are a common technique used in social science. Plus, the Likert scale is among the most frequently used psychometric tools in social sciences and educational research. Despite its frequently used, the Likert scale raises up many questions mark. We can say that the use of the Likert scale in its classical form is too rigid and loses valuable information. Li (2013, p. 1613) calls on previous studies that "have claimed that fuzzy scales are more accurate than traditional scales due to the continuous nature of fuzzy sets". The aim of this research is to reduce the inaccuracy caused by the use of the Likert scale, by proposing a method of more appropriate processing of data collected in this way. As shown in this paper, fuzzy methods can be a good alternative. The research methodology consists of using the usual technique on the set of fuzzy numbers by considering the input data as linguistic variables, subsequently identified by triangular fuzzy numbers. The obtained scale is more elastic with respect to the input data, therefore it better captures the reality. The newly proposed method is applied in the concrete example of the competitors in the hotel field. The Importance-Performance Competitor Analysis is utilized. A weakness of the method is due to the use in its application of data collection with the Likert scale. The results conclude on the situation of the competitors regarding each attribute considered as in the crisp version of the method, but the identification and processing of data correspond better to the aspects of subjectivity and uncertainty specific to human thinking. A novelty is also the obtaining of a hierarchy within each category of attributes from the quadrants proposed by the Important-Performance Analysis in relation to the competition.