After choosing one of two odors you like equally well, might you find the chosen odor more pleasant than before? In a 2012 study, pleasantness ratings increased for odors selected from difficult pairs, and that change was still observed one week later.
The researchers wanted to examine three questions: how long preference changes after a choice last, whether people need to remember their choices explicitly, and whether choosing can truly be said to change preferences.
Choosing between similarly liked odors, then smelling them again
The first experiment involved 41 University of Geneva students. They initially smelled eight odors and rated their pleasantness, intensity, and familiarity. Those ratings were used to create two difficult pairs with similar pleasantness and two easy pairs with large differences in pleasantness.
Participants chose their preferred odor from each pair and then rated all eight odors again. One week later, they actually smelled and rated the same odors and chose again from the same pairs. Rather than simply recalling that they must have liked something, they experienced the odors again at the time of rating.
The results were clear for the difficult pairs. Chosen odors became more pleasant than at the initial rating, while rejected odors became less pleasant. Both changes were significant immediately after choosing and one week later. Comparable changes were not confirmed for the easy pairs, whose pleasantness had differed substantially from the outset.
Ratings diverged even on trials where the choice was not correctly remembered
In the first experiment, participants remembered their choices on too many trials to allow an adequate comparison of remembered and unremembered choices. An experiment with another 35 students therefore increased the number of rated odors to 16 and the number of choice pairs to eight. These included odors evoking detergent, shampoo, flowers, fruit, and sandalwood.
At the end of the initial session, participants completed an unexpected memory task that also included previously unpresented odors. They first reported whether they had smelled an odor before and, for odors they said they had encountered, whether they had chosen or rejected it.
In this experiment too, pleasantness increased for odors chosen from difficult pairs both immediately after the choice and one week later. By contrast, the decrease for rejected odors was not significant at either time. The difference between changes for chosen and rejected odors was also found when examining only trials on which the initial memory task did not yield a correct answer about the choice.
The timing of the memory assessment matters. Choice memory was tested at the end of the initial session, not retested one week later. The results therefore should not be read as confirmation that participants had completely forgotten their choices a week later.
Across the two experiments, participants chose the same odor again one week later approximately 80% of the time. Choices were relatively consistent regardless of initial memory performance, but that alone does not establish that the act of choosing caused the effect. Pre-existing preference differences may have persisted.
Liked because it was chosen, or chosen because it was liked?
Identical initial ratings do not necessarily mean that underlying preferences are exactly equal. Ratings fluctuate, so choosing may reveal small pre-existing differences that then appear to widen at the next measurement. The researchers conducted a third experiment to address this methodological problem.
They recruited 26 participants and analyzed 23 after excluding three with equipment problems. Participants first rated 20 odors. On each subsequent trial, without knowing which odor would come next, they chose whether to pay 5 Swiss francs from experimental money they had received, and then smelled the odor. They were instructed to pay on half of the 20 trials.
The odors themselves had already been rated once, but participants did not know which would appear next when deciding whether to pay. This was designed to prevent the payment choice from directly revealing a pre-existing liking for that odor.
Less pleasant odors changed more favorably when participants had paid
The analysis was restricted to the 10 relatively less pleasant odors among the 20 initially rated by each person. This corresponded to the hypothesis that paying for an unpleasant experience might lead people to adjust their evaluation to make it consistent with their choice.
Within this subset, pleasantness ratings changed more favorably for odors that had cost 5 francs than for those experienced without payment. The difference was significant both immediately afterward and one week later. Memory performance for payment decisions did not significantly exceed chance, and neither familiarity nor intensity ratings differed significantly between payment conditions.
However, the original paper is inconsistent about how the effect changed over time. It describes the time-by-payment-condition interaction as nonsignificant but gives p<.001 for the test in the same passage. The reported condition differences at each time point can be interpreted, but they do not establish that the size of the difference remained unchanged for the entire week.
Summary
Choosing between similarly liked odors was followed by increased pleasantness for the selected odor, a pattern observed not only immediately afterward but also one week later. Decreased pleasantness for rejected odors was clear in the first experiment but not significant in the second.
Ratings diverged even on trials where the initial choice was not correctly remembered. In the experiment involving payment without knowing the next odor, more favorable changes in ratings of the less pleasant odors also persisted. Odor preferences may therefore involve the experience of choosing rather than being entirely fixed. Explanations involving memory and justification of one’s own choices remain hypotheses examined in light of these findings.
The study received funding from the Swiss National Science Foundation’s affective-sciences research center and a research grant from Firmenich. The paper disclosed that two authors were employed by Firmenich at the time and stated that the funders had no role in study design, analysis, the decision to publish, or related activities.
Reference
Coppin G, et al. (2012). When Flexibility Is Stable: Implicit Long-Term Shaping of Olfactory Preferences. PLOS ONE, 7(6), e37857. Original research article and open full text
