We hesitate to approach something that smells unpleasant. One study tried to capture this everyday feeling through the speed of an on-screen action. After smelling synthetic body odor, participants were slower to initiate an action that brought them closer to a perfume bottle on a screen than after an odorless stimulus or a pleasant fragrance.
This 2021 study did not measure how far people actually walked toward something. Through separate experiments with sounds and odors, it investigated whether a task combining screen movement and keypresses could be adapted into an assessment that did not directly ask about likes and dislikes.
Creating a sense of approaching and retreating on screen
The researchers modified an experimental task called VAAST. Participants pressed a key to approach or move away from an object in a room shown on screen. Each correct press enlarged or reduced the entire scene by 10%, making it appear as though they were moving forward or backward.
A single movement required three keypresses, but the analysis concerned the time taken to make the first press and its accuracy. The researchers predicted that pleasant stimuli would lead to faster approach actions, while unpleasant stimuli would lead to faster retreat actions.
Another aim was to extend earlier tasks using only words on a screen to a sounding loudspeaker or a scented bottle. With backgrounds resembling everyday settings, they tested whether stimulus properties would be reflected in actions without asking participants to report their preferences.
First, testing the method with sound quality
The first experiment involved 26 university students, with 25 included in the analysis after one was excluded because of technical problems. A living room and loudspeaker appeared on screen while French words played through headphones. The words named living beings or inanimate objects, and participants approached or retreated according to that category.
The sounds were undistorted, slightly distorted, or moderately distorted. The words had neutral meanings, and poorer sound quality was assumed to be more unpleasant. Participants judged the word category, not whether they liked the sound.
The first approach action took an average of 886 milliseconds for undistorted sounds, 915 milliseconds for slightly distorted sounds, and 925 milliseconds for moderately distorted sounds. Responses were faster for undistorted sounds than for either distorted condition. By contrast, sound quality produced no significant difference in retreat actions, and there was no clear difference in accuracy between conditions. The predicted response difference appeared on the approach side.
In the odor experiment, judging only the bottle’s tilt
The next experiment involved 26 participants recruited from different departments at Firmenich: 18 women and eight men, with a mean age of 45.1 years.
Three stimuli were selected through preliminary testing: a pleasant fragrance, synthetic body odor, and an odorless solvent. The pleasant fragrance was a deodorant-product scent with floral, green, geranium, citrus, and other characteristics. The fragrance and synthetic body odor were selected to differ greatly in pleasantness while differing relatively little in intensity and familiarity.
Participants in the main experiment also initially rated the odors’ pleasantness, intensity, and familiarity. Pleasantness was highest for the fragrance, followed by the odorless stimulus and then synthetic body odor. The fragrance and body odor did not differ significantly in intensity or familiarity, although this does not guarantee that they were completely equivalent.
The screen showed a bathroom and a perfume bottle, and participants were told to imagine that the odor came from the bottle. After they smelled an odor, the bottle tilted left or right, and they pressed the approach or retreat key according to that direction. The correspondence between tilt and movement direction was reversed across participants.
Each odor presentation was followed by four tilt judgments, for a total of 96 responses. Here, implicit assessment means that participants were not asked to state their likes or dislikes during the action task. They knew that odors were present and had rated them beforehand.
Synthetic body odor delayed approach actions
Once again, the odor effect appeared on the approach side. The mean time to the first approach keypress was 661 milliseconds for synthetic body odor, 626 milliseconds for the odorless stimulus, and 636 milliseconds for the pleasant fragrance. Responses to synthetic body odor were significantly slower than those to either the odorless stimulus or the fragrance.
The difference between the pleasant fragrance and the odorless stimulus was not significant. Thus, the central finding was not so much that a pleasant scent accelerated approach as that unpleasant synthetic body odor delayed the approach action. Retreat actions did not differ significantly across the three odor conditions.
Accuracy was slightly higher for retreat than for approach actions, but it did not change significantly with odor. The odor effect was found as a difference in response time involving movement direction, rather than as a general reduction in task accuracy.
How far can this go as a tool for measuring preferences?
In both the sound and odor experiments, stimulus quality affected the speed of approach actions. In the odor task in particular, participants did not even have to classify the odor itself; they judged only the bottle’s orientation. The method’s promise lies in the finding that a previously smelled odor affected a task that did not directly refer to emotion.
However, neither experiment supported the prediction that unpleasant stimuli would speed up retreat. The authors suggest that the stimuli and settings may not have been perceived as frightening dangers, and discuss future comparisons between approach and no-movement conditions. This explanation is a hypothesis for understanding the different approach and retreat results, not a finding from a direct measurement of fear.
The odor experiment also used markedly different stimuli: a pleasant fragrance and unpleasant synthetic body odor. Whether the method can distinguish subtle preferences between similar perfumes, or whether on-screen responses relate to actual product choices or purchases, remains to be tested. Given the small in-house sample as well, the method is best viewed as a possible complement to consumer assessment.
Summary
After smelling synthetic body odor, participants were slower to make the first action toward an on-screen perfume bottle than after an odorless stimulus or pleasant fragrance. In the sound experiment, approach actions were also slower for distorted than for undistorted sounds. Neither experiment, however, found a clear difference in retreat actions between stimulus conditions.
The study showed that properties of sounds and odors can be reflected in the timing of simulated approach actions without asking about preferences at that moment. Its contribution is to developing a method for combining behavioral indicators with verbal ratings.
The research received support from the French National Research Agency, ANR. The paper discloses that two authors were employed by Firmenich at the time of the research, and that other authors were affiliated with Orange Labs, PSA Groupe, and Silliker/Mérieux NutriSciences.
Reference
Cereghetti, D., Faye, P., Gros, L., Mahé, L., Diaz, E., Cayeux, I., Heritier, T., & Versace, R. (2021). Validation of New Methods of Using Simulated Whole-Body Movements as Implicit Indicators of Sound and Odor Preferences. Frontiers in Psychology, 12, 659269. Original research article
