Title: Joint model-based clustering for ordinal survey data
Authors: Bettina Gruen - Wirtschaftsuniversität Wien (Austria) [presenting]
Sara Dolnicar - The University of Queensland (Australia)
Abstract: Survey data collected is typically ordinal in nature. As such, it is susceptible to response styles. Response styles are consistent tendencies displayed by survey respondents when responding which are not related to the specific item content. When clustering ordinal survey data, ignoring response styles can lead to clusters which do not differ in beliefs, but merely in how cluster members use survey answer options and which possibly occur in addition to the belief clusters. We propose a finite mixture model which simultaneously clusters and corrects for response styles, permits heterogeneity in both beliefs and response styles, accommodates a range of different response styles, does not impose a certain relationship between the response style and belief clusters, and is suitable for ordinal data. The performance of the model is tested using both artificial and empirical survey data.