The effects of action, normality, and decision carefulness on anticipated regret: Evidence for a broad mediating role of decision justifiability

Abstract

Two distinct theoretical views explain the effects of action/inaction and social normality on anticipated regret. Norm theory (Kahneman & Miller, 1986) emphasises the role of decision mutability, the ease with which one can imagine having made a different choice. Decision justification theory (Connolly & Zeelenberg, 2002) highlights the role of decision justifiability, the perception that the choice was made on a defensible basis, supported by convincing arguments or using a thoughtful, comprehensive decision process. The present paper tests several contrasting predictions from the two theoretical approaches in a series of four studies. Study 1 replicated earlier findings showing greater anticipated regret when the chosen option was abnormal than when it was normal, and perceived justifiability mediated the effect. Study 2 showed that anticipated regret was higher for careless than for careful decisions. Study 3 replicated this finding for a sample holding a different social norm towards the focal decision. Finally, Study 4 found that, when decision carefulness, normality and action/inaction were all specified, only the former showed a significant effect on anticipated regret, and the effect was again mediated by perceived justifiability. Decision justification theory thus appears to provide a better account of anticipated regret intensity in this context than does norm theory.

Keywords:

• Anticipated regret
• Decision process carefulness
• Justifiability
• Normality
• Regret aversion

Notes

¹Anticipated regret rated for scenarios such as the one we used has been found to predict real choices. Wroe, Turner, and Salkovskis (2004) compared different potential predictors of actual immunisation decisions and found that “anticipated regret … was the strongest predictor of likelihood of immunizing the child” (p. 38), predicting 57% of the variance (demographic variables, in contrast, predicted only 1% of the variance). The vaccination decision has also been the context of a number of studies examining a possible omission bias, a tendency to prefer inaction over action, in health-related decisions, a bias possibly resulting from increased anticipated regret for action (e.g., Ritov & Baron, 1990). We are not concerned here with the debate about the existence of this bias (the interested reader can refer to the papers cited above as well as Connolly & Reb, 2003, and Baron & Ritov, 2004).

²We manipulated action only between-subjects because, based on previous findings showing a preference for, and general acceptance of, vaccination as the typical choice in the population (Connolly & Reb, 2003) we felt confident that the action/inaction manipulation would affect anticipated regret (and perceived justifiability) in a between-subjects design.

³Connolly and Reb (2003) provide evidence that participants find this a comprehensible scenario and are able to construct plausible explanations for the decisions they would make. Interestingly, a preference for action or inaction per se is very rarely mentioned in these explanations.

⁴Moderation exists when the sum of the repeated mediator measures predicts the difference of the repeated dependent variable measures. While we were not interested in a potential moderating role in the present context, Judd et al. (2001) recommend including both difference and sum predictors in order not to mis-specify the model. We have followed their advice in our analysis and found no significant moderating role of perceived justifiability.

⁵We recognise that there is an important and on-going debate on the broad issue of decision quality, coherence, outcome bias, cardinality, etc. We intend no novel contribution to this discussion here. We suggest only that, in a wide range of decision settings, a process described as careful, thorough, and drawing on a range of available information will be perceived as better justified than one that has none of these features. Our participants seem to agree with this judgement.