Applying the random regret minimization model in consumer decisions

Abstract

Notwithstanding the increasing attention made to the random regret minimization, which is a new econometric multinomial and multi-attribute discrete choice model, most of light has been limited to certain fields. While quantitative choice modeling is also a prominent topic in marketing, there has been no empirical application of random regret model in consumer decision context. This research aims to test the validity of random regret minimization model and the conditions in which either model would be more appropriate in describing the data in two ways: (a) situational difference and (b) individual difference. First, individuals were explicitly asked to think about how they might feel if they have made a wrong choice, or explicitly informed to expect the outcomes of the alternatives (situational difference). Second, individual differences were investigated based on chronic self-regulatory focus. Online experiments are conducted on US-nationwide database on four electronic products. The results suggest that there may be individual difference for which the random utility model or the random regret model would work better; for promotion-focused individuals, the random utility model works better than the random regret model. On the other hand, although the random regret model does not necessarily work better than the random utility model, the difference in fit is not significant for prevention-focused individuals. When subjects are first asked how they might feel if they have made a wrong decision, their brand choices are not described significantly better by the random utility model than the random regret model. Lastly, the random regret model may be used when participants were expected to receive feedback on the outcomes of their choices. The empirical results suggest that there may be situational as well as individual difference in terms of the model fits of the random utility model and the random regret model.