Evidence About Contextual Drivers Of Liking and Preference
Date: February 8, 2007
Speaker’s Name and Affiliation: Debra Zellner, Montclair St. University
Title: “Contextual influences on liking and preference”
Do our likings and preferences stay the same, or are they influenced by the context in which they are presented? While we may think that the foods and beverages we like are constants, and our acceptance of them on a day-to- day basis does not vary, research from Dr. Zellner’s lab has demonstrated that these preferences may show a great deal of variation depending on the context in which they are presented. The data presented may have broad implications for the ways in which we study food preferences in a research setting.
To illustrate the point that context can affect our preferences, Dr. Zellner began with an exercise where half the room was shown pictures of attractive, colorful tropical birds, followed by a “less attractive” common North American variety. The other half of the group was shown less attractive birds, followed again by the same common North American variety presented to the first group. Results of this mini-experiment show that if you are shown attractive, colorful birds first, you will rate the common North American variety as less attractive than if you are shown the less attractive birds first.
Gustav Fechner originally proposed the idea of “hedonic contrast” as-that which gives pleasure gives more pleasure the more it enters into contrast with sources of displeasure or lesser pleasure; and a corresponding proposition holds for that which gives displeasure. This statement is qualified further by noting that the two factors being compared must bear some similarity to one another, or in other words, had to be in the same category of stimuli.
Is the old adage about comparing apples to oranges true? Two experiments were undertaken to determine if ratings of apples are influenced by oranges in some people who categorize them both as fruit, while for people who view apples and oranges as different things, ratings for one should not affect ratings of the other. This hypothesis was tested with coffee, as one can identify “good” (eg., coffee house, gourmet), and “less good” (eg. ordinary canned coffee), versions of coffee. As expected, subjects typically rate the coffee house versions of coffee as more pleasant and acceptable than the ordinary canned versions.
But, when asked whether they categorize these two drinks as the same or different, people who categorized them as the same rated the canned coffees as hedonically negative. However, subjects who thought of the two types of coffee as different beverages liked ordinary coffee, although somewhat less than the coffee house blends. From this, we see that Fechner’s definition of hedonic contrast exists when people use the same categories for foods or beverages, but it exists to a lesser extend when people use different categories.
To further support these findings, Dr. Zellner’s lab replicated this experiment with beers, (regular vs. specialty brews). The regular beers were rated poorly by subjects who categorize both regular and specialty brews as “beers,” in the same category. However, if subjects consider these “different” beverages, ratings of the regular beers were more positive.
Given the above results, what would happen if you instruct subjects to put good and less good versions of stimuli into one or two categories? Can you induce a hedonic contrast? In the next studies that were presented, categories of stimuli were experimentally manipulated to determine if 1) hedonic contrast would occur when subjects were told that “good” and “less good” stimuli were in the same category, and 2) if a hedonic contrast would be prevented by informing subjects that stimuli were in different categories.
When performing these experiments with Mistic brand juices (dilute and full strength version), and with pictures of tropical and North American birds, results showed that in fact you could induce a hedonic contrast if subjects were told that items were in the same category. However, in both studies, the hedonic contrast was not eliminated when subjects were told to categorize stimuli into different groups.
Though the contrast was not eliminated, the degree of liking in the group told to categorize was significantly less than those who were not told to categorize. Dr. Zellner hypothesizes that simply being told to categorize isn’t sufficient to completely eliminate all hedonic contrast, primarily because people naturally have categories for certain items, and thus experimental manipulations are not able to completely mask these.
To test this, further experiments were done in which novices and experts were asked to rate the attractiveness of flowers. The assumption guiding this work is that a flower expert should categorize orchids differently from irises, whereas a novice will see them all as “flowers.” When asked to rate the attractiveness of “less attractive” test orchids preceded by either attractive orchids, attractive irises, or nothing, experts only showed hedonic contrast when attractive orchids preceded the test orchids.
On the other hand, novices showed hedonic contrast regardless of whether attractive orchids or iris preceded the test orchids, as they considered them all simply “flowers.” So, Fechner’s statement that hedonic contrast occurs when the context and test stimuli “bear a certain resemblance to each other,” is further supported by these studies.
One final question that was addressed is whether hedonic contrasts can work in both a positive and a negative direction. Data using Goya paintings from the dark period and Goya “tapestry paintings” served as the stimuli for this study. Results suggest that Fechner was correct, and hedonic contrasts work in both directions.
Comparing mediocre stimuli to good ones make them pale in comparison (an example of negative hedonic contrast). Comparing mediocre stimuli to bad ones make them appear better (positive hedonic contrast).
How might these data influence food preference ratings in the lab? Certainly, the context in which foods and beverage stimuli are presented to subjects might drastically affect subjects’ ratings. Moreover, subjects may differ in their ratings of stimuli simply as a matter of their overall life experiences. These considerations might be factored in to food preference and sensory testing, to improve the ability to interpret experimental results.
Q. When you did your experiments with coffee, did you account for the fact that there are flavored coffees, and some people may have experience with them?
A. These studies were done before flavored coffees were as common as they are now.
Q. Are you sure that the visual analog scale you used was sensitive enough to pick up a 10 mm difference in ratings? You must have had a very large N.
A. Actually, we had a small subject number, but the results were highly significant.
Q. Are the individuals who only like the high-end beers most often the “foodies?”
A. Well, this whole area of research began because when I worked with Paul Rozin, he accused me of not being able to appreciate a regular Indian meal from a small family-owned restaurant near campus, whereas he could appreciate both really high-end foods and lower end burgers and fries and fast foods. Why are there such vast differences in hedonic reactions to different foods, and does some of this have to do with one’s experience with other foods?
Q. What year was this study done?
A. End of the 1990s, I believe.
Q. Have you interviewed people in their own houses, and do you think it would make a difference on your experimental paradigms?
A. We haven’t done this yet, but I don’t see why it would make a difference.
Q. What did you do about the fragrance of flowers when you were presenting them in your experiments?
A. We used pictures of flowers, so the fragrance did not make a difference.
Q. Did you ask them to describe what their personal definition of attractiveness was? It would seem important to know this, as people might show a lot of variation in how they judge the attractiveness of flowers.
A. I’m not sure if I’m following your point. How would we assess something like that?
Q. Have you done experiments the other way around, where you show the unattractive items (flowers, birds, etc), first, followed by the more attractive items?
A. Yes, and I will show you those data next.
Q. Do you use students in most of your experiments? Do they have a background in art?
A. We do use students for our tests, but we screen them so we can exclude anyone with a background in art.
Q. Do these experiments also work with sound?
A. Yes, if you give people a loud sound, they will judge sounds that follow as softer, due to a contrast effect.
Q. Do you think this might be a detection issue, and not a hedonic contrast issue? Could it be that subjects are unable to detect a difference in stimuli?
A. I don’t think it’s that, but it could be.
Q. Do you randomize presentations, and what is your recovery time between viewing stimuli?
A. There are only two different presentations, and the recovery time between samples is 1 minute. I’m not sure that a longer recovery time would matter.
Q. Have you done these experiments with chocolate?
A. Not yet, but that would be interesting to test.
Q. What if you use a less neutral painting as a test stimulus?
A. We’d probably end up with some sort of ceiling effect.
Q. Did you do any pre-testing to find out reactions that subjects have to test paintings before the study?
A. Yes, we did.
Q.Why do you think categorizing test stimuli into specific categories affects the preference that subjects have for it?
A.Probably because you are making a comparison between the category something is placed in (what your expectations are of items in that category), and your actual perception of the stimulus.
Q. Aren’t there some types of psychophysical tests you can perform, such as “same-difference” tests, that might help you screen people before your studies? In that way, you could verify that they can indeed detect differences in the test stimuli you are using.
A. I guess you could find people who have very large differences in preference. There are probably personality characteristics having to deal with that, but we haven’t been able to identify these yet.
Q. Are you saying that there are some people that are more likely to categorize items, and that the tendency to do this might be a hardwired trait? These people would categorize for multiple items (coffee, flowers, birds, art, etc).
A. I don’t know, we’ll have look across studies.
Q. Do you have any thoughts on what the brain mechanism that underlies these characteristics might be?
A. No, I don’t know what that might be.
Comment: I think there is some data on serotonin in autistics, who often aren’t able to categorize, but after treatment their ability to categorize improves.
Q.Do you think it could be as simple as broader life experience and level of exposure to a variety of experiences that are resulting in some of your contextual effects?
A.In response to that comment, in our lab, we have seen that college students that are living at home have less broad food preferences than students who live on campus. So, your point is well-taken.