Unconscious influence of decisions – the case of preferences

In principle, decisions may exert unconscious influence on virtually any psychological process. We will here restrict our analysis on the influence of decisions on preferences. This influence has been extensively replicated (e.g., Harmon-Jones & Mills Reference Harmon-Jones and Mills1999), although the exact nature of the mechanisms at play has recently been debated (e.g., Chen & Risen Reference Chen and Risen2010). Here, we focus on a different aspect: the level of processing (unconscious vs. conscious) that is required for the modulation that decisions can exert on preferences, which is controversial.

The assumption of awareness

The influence of decisions on preferences can be summarized as follows: After a decision between two similarly liked alternatives (and only in those instances), the chosen one is rated as more pleasant and the rejected one as less pleasant (Brehm Reference Brehm1956). Classical authors (have argued that the impact of decisions on preferences requires awareness (Festinger & Walster Reference Festinger, Walster and Festinger1964; Wicklund & Brehm Reference Wicklund and Brehm1976). Thus, one is assumed to be aware that one's choice is in conflict with the desirable aspects of the rejected option and with the undesirable aspects of the chosen one, both options being similarly liked. The awareness of this conflict is assumed to lead to the need for cognitive dissonance reduction (Festinger Reference Festinger1957), that is, in this case, readjusting one's preferences to be more in line with one's choice – by devaluing the rejected option and/or valuing the chosen one more positively.

Distraction from awareness

Lyubomirsky and Ross (Reference Lyubomirsky and Ross1999, study 3) have shown that distracting participants reduced postchoice preference modulation. In the same vein, Allen (Reference Allen1965) showed that the modulation of preferences following a difficult choice was less pronounced when postchoice time was filled with an extraneous cognitive task. As put by McGregor et al. (Reference McGregor, Newby-Clark, Zanna, Harmon-Jones and Mills1999), “individuals seem to have a remarkable capacity for avoiding awareness of inconsistencies unless their noses are quite vigorously rubbed in them” (p. 331). In these experimental settings, decisions do not impact preferences as much when awareness of the dissonance (which is assumed to drive this impact) can be avoided. In other words, awareness appears to be required in postchoice preference modulation.

How can this evidence be related to N&S's discussion? Although the design of these experiments is different from the deliberation-without-attention paradigm, the two criteriaFootnote ¹ that need to be met in this paradigm can be argued to also apply here. Criterion 1 was met in both experiments: Distracting participants led to significantly lower postchoice preference modulation than when participants were not distracted. However, just as in the case of the deliberation-without-attention paradigm, criterion 2 was not so well accounted for. In Lyubomirsky and Ross's (Reference Lyubomirsky and Ross1999) experiment, criterion 2 could not be assessed: The second measurement was made 8 minutes after the decision, but no immediate condition was run. In contrast, Allen's (Reference Allen1965) experiment includes four conditions: 2 and 8 minutes of cognitive rest, as well as 2 and 8 minutes of cognitive activity. Preference modulation at 8 minutes in the cognitive activity condition was lower than the one at 2 minutes (independently of the condition), which supports the idea that awareness is a necessary requirement for preference modulation by choice.

Recent evidence against the necessity of awareness

Recent empirical data have, however, questioned the assumption of the necessity of conscious processes in the impact that decisions can exert on preferences. These experiments have investigated this impact in three ways. First, patients with anterograde amnesia were asked to make decisions, which they were, due to their condition, unable to consciously recall (Lieberman et al. Reference Lieberman, Ochsner, Gilbert and Schacter2001). Second, healthy participants were asked to make several decisions. Their conscious memory for these decisions was measured in an incidental memory task at the end of the experiment. This makes it possible to dissociate the impact of consciously forgotten decisions from the impact of consciously remembered decisions on preferences (Coppin et al. Reference Coppin, Delplanque, Cayeux, Porcherot and Sander2010). Third, participants were presented with decisions that they had not in fact made themselves. This third approach has been applied to both rhesus monkeys (Egan et al. Reference Egan, Bloom and Santos2010) and human participants (Sharot et al. Reference Sharot, Velasquez and Dolan2010). In all of these three experimental contexts, participants' lack of decision memory makes it highly unlikely that experimental demand effects might be at play. Unconscious influences of decision on subsequent preferences were nonetheless still reliably demonstrated.

These experimental settings control the potential pitfalls of experimental demand or social desirability effects (i.e., reliability criteria) by targeting the conscious memory of the decision. Thus, although participants' attention was not, per se, diverted away from the experiment's hypothesis, participants could not be consciously acting in a way that fits the experimenter's hypothesis. As recommended by N&S, the most recent literature on the (potentially unconscious) impact of decisions on preferences has moved away from the deliberation-without-attention framework by replacing it by this type of experimental setting. The similarity between the literature of the unconscious influences on decisions and the literature on the unconscious impact of decisions is striking in this respect.