Sociological applications of fuzzy classification analysis

Abstract

Research methodology and data analysis in sociology are dominated by probability theory. However, probability theory cannot always deal with the imprecisions of the real world. In this paper, the basic concepts of fuzzy sets theory and fuzzy classification are discussed. An empirical example is presented. The fuzzy classifications of respondents based on their attitude of self protection and offense are highly related to the behavior of carrying weapons. By incorporating the classifications into our analytical model, we find that the predictive power of the model is enhanced and multicollearity among independent variables is avoided.