Customer Voice Retaliation (CVR) Construct Verification: A Rasch Analysis Approach

Abstract

The study on customer complaining behaviour has received significant attention over the past decades. Generally, customers expect a fair treatment for the investments made in a relationship with the organisation. Hence, perceived unfairness will create an impression that they have been betrayed, and thus motivate them to complain as a way of expressing dissatisfaction. In extreme cases, they might opt for aggressive complaining to compensate unfairness. In this study, the term aggressive complaining will be replaced by customer voice retaliation (CVR). Based on previous literatures, a framework was developed to measure CVR which consist of three constructs. The aim of this study is to verify the construct for a new CVR framework. Rasch analysis was used to examine reliability for both respondents and items. It should give us the list of items that should be included in measuring the CVR constructs. Sample used for the study were 27 for pre-testing, and 66 for the pilot study. Respondents consisted of subscribers who had experienced dissatisfaction, and had to some extent performed complaining behaviour. From the pre-test analysis, item polarity indicates that all items were measuring in the same direction. Similarly, the summary statistic from the pilot study indicated that item reliability and item separation was 0.95 and 4.39 respectively, while for person reliability and person separation was 0.95 and 4.6 respectively. However, the result from the pilot test for model Unidimensionality suggested the existence of a secondary dimension which was a possible contributor to multicollinearity problem. This indicates that items for measuring CVR needs to be reviewed and instrument construct validity call for further refinement.