You can never go back. I had an opportunity recently to take a sip of a wine that I used to happily drink some … ok, many ….years ago. A sip was more than enough! The shame of it was that the amount of pleasure I can get from an excellent wine now is probably no greater than the pleasure that Chateau Chunder (ok, not that bad) gave me all those years ago.
Pleasure is relative … not only to our current state of mind, but to recent experiences as well. In an excellent discussion of this phenomenon, Parducci [1] analysed the impact of recent experience on how happy we feel. How happy would you be to find $10 on the street? Let me ask you another question before you answer: are you just stepping out of your Ferrari or on your way to the unemployment office? Alternatively, imagine you find $20 on the street every day for a month ….. how do you feel if one day there’s only $10? In both cases, the context (your financial state or your recent experiences) provides the background that determines the degree of pleasure that you are likely to experience.
Our sensory experiences are not exempt from context effects either. Another balmy 19 C (66 F) day. Now, depending on where you live, you might agree with me or alternatively consider that it’s me who’s balmy. The fact is, I have given you a piece of information and an interpretation, but what is missing is to know where I am at present, and hence the season*. These extra details allow you to place both the information and my interpretation in context. Most of us at some time have eaten a dish that is just too hot (spicy). Part of this may be due to the physiological impact of recent chilli eating, or lack thereof. But mostly, it will be a mental comparison of the heat of the spicy dishes that you have had recently with what you are eating now [2].
Each of these examples show how any given sensory or emotional experiences occurs against a background of other similar recent experiences. From the perspective of wanting to measure sensory or hedonic experiences, context effects are not merely an interesting quirk – they are the main game. All human measurement is relative. Even a simple question of deciding whether or not you like a beverage is really a case of liking it “compared to what”. In evaluating consumer preferences for foods or beverages, the common scenario is that multiple samples – different brands or different versions of a product – are compared at each tasting session. In the less common situation, though, that a single product was rated for liking, we would expect that the point of comparison was the consumer’s prior, and especially recent, experiences with that class of product [3].
When multiple products are compared within the same evaluation session, the products are compared relative to one another. One consequence of the influence of context in ratings is that a given product can be given a low or a high liking rating entirely as a function of what other products are in the evaluation [4]. To take an obvious (and unrealistic) example, imagine that you are asked to rate your liking for different flavours of jelly-beans: orange, lime, blueberry, banana, and … tomato. The latter is obviously an odd sample (despite being a fruit) and is likely to be rated lower. Next, however, I ask you to rate your liking for the following jelly-bean flavours: spinach, brussel sprouts, potato, cabbage, and …. tomato. Here, I imagine that the tomato flavour would be a winner.
It is crucial to understand the implications of context effects if consumer data are not to be misused. Many food companies have in the past, and to a lesser extent still today, relied on fixed acceptability ratings as criteria for action. For example, on the standard 9-point hedonic scale, a mean rating of 7 (representing “like moderately”) in consumer testing may be required to launch a product. Since any given number on a rating scale can only really be understood in relation to the ratings given other products, there is no validity to adopting such a value. Ironically, dogmatic reliance on such cutoff values is a key reason why there is resistance in industry to adopting new approaches to measurement.
A key theoretical question about context effects is whether they reflect actual subjective experience or occur due to the way we use rating scales. One of the main explanations for context effects is Adaptation Level Theory which suggests that we adapt to the changing environment, establishing a reference level against which intensity/pleasure is judged [5]. This approach has many parallels to sensory adaptation (e.g, adapting to be able to see in a dark room) and leaves open the possibility that context alters experience. The Range-Frequency Theory on the other hand is a theory of how we use rating scales [6]. This theory suggests that when faced with the task of rating the intensity or pleasantness of a stimulus range (e.g., products varying along some dimension), we mentally divide the scale into equal intervals, and distribute the ratings over these intervals. Thus, context effects are due to the range of stimulus values and the frequency of each value.
There are certainly data to support range-frequency theory but clearly I do not need a rating scale to judge whether or not I am happy – to greater or lesser degrees - at finding $20 on the street. That’s an unusual, and positive event. Once this has happened 20 times though, it becomes the norm against which daily events are judged. Tomorrow’s $20 may not even evoke a smile, and $10 in its place may even leave me a little annoyed.
Food consumers, too, experience the same phenomenon. And this may give rise to an interesting paradox in which favourite foods – as reflected in recent, repeat consumption – do not evoke particularly strong positive emotions. A relatively new field in consumer research is the measurement of a wider range of emotions beyond liking. Understanding the impact of context effects in such measures is clearly an important step before their widespread application. Moreover, while sensory acceptability is crucial, there are other reasons why we consume. These include feeling of comfort or security. Perhaps these motivations are exempt from the effects of context, but as yet we don’t know.
[ * The missing bit is ‘Sydney in winter’ and, though making you envious is far from my mind, I need to point out that, yes, it’s often 19 C]
_______________________________________________________________________________________________________
1. Parducci, A., Happiness, Pleasure and Judgment, 1995, New Jersey: Lawrence Erlbaum.
2. Stevenson, R.J. and J. Prescott, The Effects of Prior Experience with Capsaicin on Ratings of Its Burn. Chem. Senses, 1994. 19(6): p. 651-656.
3. Walter, F. and R.A. Boakes, Long-term range effects in hedonic ratings. Food Qual. Pref., 2009. 20: p. 440-449.
4. Schifferstein, H.N.J., Contextual shifts in hedonic judgments. J. Sens. Stud., 1995. 10: p. 381-392.
5. Helson, H., Adaptation-level as frame of reference for prediction of psychophysical data. Am J Psychol, 1947. 60(1): p. 1-29.
6. Parducci, A., Contextual effects: A range-frequency analysis, in Handbook of Perception Vol. II. Psychophysical Judgment and Measurement, E.C. Carterette and M.P. Friedman, Editors. 1974, Academic Press: New York. p. 127-141.
