Compromise and attraction effects under prevention and promotion focus

There is ample evidence that consumer preferences are context dependent. In particular, they are sensitive to the characteristics of the alternatives present in the choice set. The behavioral decision-making literature refers to the changes in share when another alternative is added to the choice set as 'context effects'. Two types of context effects, 'attraction' and 'compromise', are particularly important and have received considerable attention from decision researchers over the years. The attraction effect refers to the finding that adding an alternative which is inferior to another alternative in the choice set increases the share of the relatively superior alternative. The compromise effect, on the other hand, is observed when an alternative becomes more attractive when it is presented as a middle option in a choice set than when it is presented as an extreme option. Adopting a motivational approach to consumer decision-making, and building on regulatory focus theory, the present thesis hypothesizes that (1) consumers' susceptibility to the compromise effect is greater when the prevention system of self-regulation is activated than when the promotion system is activated; (2) consumers' susceptibility to the attraction effect is greater under promotion focus than under prevention focus; (3) consumers' need to justify their choices will increase their sensitivity to the compromise effect under prevention focus but will decrease it under promotion focus; (4) consumers' need to justify their choices will increase their sensitivity to the attraction effect under promotion focus but will decrease it under prevention focus; and (5) products that are associated with a prevention concern will be more attractive when positioned as compromise options than as asymmetrically dominant options, whereas the opposite is true for products that are associated with a promotion concern. Three experimental studies are conducted to test these hypotheses. The data are analyzed using hierarchical linear and non-linear modeling techniques. The results provide full support for hypotheses 1, 3, 4, and 5, and partial support for hypothesis 2. Theoretical and practical implications of these findings are discussed along with the study's limitations, and considerations for future research