The predominant paradigm for marketing model of choice is the random utility maximization (RUM) paradigm, which assumes that a rational decision maker chooses an alternative so as to maximize the expected utility derived from the alternatives in the choice set. In contrast to the popular paradigm, Chorus (2010), in transportation science, has proposed a new discrete choice model of random regret minimization (RRM) based on regret theory (Bell, 1982; Loomes & Sugden, 1982). This model is an econometrically tractable, multi-attribute and multinomial reference-dependent model whereby the fundamental underlying decision mechanism is that individuals are assumed to make decisions so as to minimize the anticipated regret arising from the foregone alternatives performing better than the chosen alternative on attribute levels. Although there has been growing interest in the RRM, limited efforts have been made to utilize the RRM to investigate consumers’ choices, and little is known about the potential drivers for decision rules of utility-maximization and regret-minimization (Chorus, van Cranenburgh, & Dekker, 2014).
Filling in such research gap, the main objective of this thesis is to introduce the random regret minimization discrete choice model to the marketing domain, and to investigate expanding the applicability of RRM in marketing. The research question of whether segmentation based on the RRM can delineate different understanding on consumers’ choices and preferences compared to the conventional RUM is addressed via simulation and discrete choice experiments. Consumers’ unobserved response heterogeneity in RRM is examined with a latent class framework. The simulation study demonstrates the consequences of imposing an incorrect data generating process mechanism assumption on the estimated model of consumer choices. Both the model fit and parameter estimates are distorted when incorrect models are estimated. Furthermore, the empirical results from a discrete choice experiment for purchasing an energy-efficient energy appliance on US residents suggest that not only certain attributes but also the relative performance of alternatives at attribute levels are important. The relative segment sizes and important attributes for each segment, and thus, behavioral implications, are different between the RUM and RRM when unobserved heterogeneity is incorporated using a latent class modeling approach.
Furthermore, drawing upon consumer behavior literature, the consumer-specific and situation-specific drivers for decision rules between the RUM and RRM are identified through a series of online discrete choice experiments on US residents again for purchasing an energy-efficient energy appliance. Specifically, chronically promotion-focused consumers are expected more likely to be utility-maximizers while chronically prevention-focused consumers are more likely to be regret-minimizers. Moreover, when consumers are explicitly asked about the anticipated regret, or are expected to quickly find out the outcomes of their choices (both chosen and rejected), they are expected more likely to be regret-minimizers than utility-maximizers. Also, when the purchasing choice is irreversible, compared to when it is easily reversible, consumers are expected more likely to be regret-minimizers than utility-maximizers. The hypotheses and impacts of the consumer-specific and situation-specific drivers for decision rules are explored using a latent class modeling approach to incorporate structural differences of the RUM and RRM. The empirical results provide support for all the hypotheses. In sum, this thesis highlights the importance of incorporating response and structural heterogeneities in exploring consumers’ choices in marketing.