Title: Learning from experience with good and bad brands: Evaluative shift or just signal learning?
Authors: Vanhouche, Wouter
Warlop, Luk
Baeyens, Frank
Issue Date: Oct-2001
Host Document: Proceedings of the Association for Consumer Research Conference vol:29 pages:284-286
Conference: Association for Consumer Research Conference location:Austin, TX (USA) date:October 2001
Abstract: In a series of seven experiments brands or scents of drinks were associated with a good or bad tasting substance, followed by taste tests of the branded or scented drinks without this substance. Prior research had shown that subjects can acquire likes or dislikes for food scents(Baeyens, Crombez, Hendrickx, & Eelen, 1995; Zellner, Rozin, Aron & Kulish, 1983), even without awareness of prior pairing with good or bad tastes. Our objective was to investigate whether mere prior association with positive or negative consumption experiences could also make brands start tasting better or worse. We contrasted the acquisition of acquired tastes with the learning of the predictive value of brands and scents before consumption (see van Osselaer and Alba 2000). If brand cues would not only predict brand quality before consumption, but also influence experienced brand quality during consumption, this might explain why in taste tests evaluations of branded products are more differentiated than those of unbranded products (Allison and Uhl 1965), and thereby increase our understanding of the origins of customer based brand equity (Keller 1998).
Subjects participated in 'a taste test of non-carbonated isotonic sports drinks’. They first received a number of learning trials, in which they consumed good (containing sugar) or bitter tasting (containing Polysorbate 20) drinks. Depending on the condition, two different brand names (e.g., Vigro vs. Fast-X) or two different neutral scents (e.g., Lemon vs. Apricot) were paired with the positive vs. negative experiences. In the test phase, transfer stimuli without the sugar or Polysorbate 20 were presented. Subjects either predicted the taste quality, or they evaluated the consumption of these drinks. They learned to predict quality with high accuracy for both types of cues, but only scents were evaluatively conditioned. That is, subjects learned to like or dislike the consumption of Lemon or Apricot scented drinks, but not of Vigro or Fat-X branded drinks.
In follow-up experiments, we systematically varied the learning and test conditions, in order to account for our failure to observe evaluative learning for brands. We tested the effects of adding a single ambiguous flavor (citric acid) to both test drinks. We induced nonsystematic scent variation or color variation to test samples or to learning samples. In each case we replicated our initial findings.
In two experiments we did find an evaluative effect for brands. In one experiment we asked subjects to indicate their liking or disliking for the brand names (brand attitude) without consumption and found an effect of prior experience. However, the absence of any dissociation between evaluations and predictions did not allow us to infer that both responses measured different constructs. In an other experiment we found that systematically combining a brand an a scent during learning produced larger differentiation in later taste evaluations than using scents only. However, we found an equivalent level of differentiation when the acquired valence of scents and brands should have been incongruous. The latter finding led to the current conclusion that brand names do not acquire evaluative valence; their function in taste evaluation may merely be to facilitate perceptual discrimination at the time of test, allowing for the behavioral manifestation of acquired valence of experience-intrinsic cues, such as scents (see Warlop, Ratneshwar and van Osselaer 2000, for a similar finding). Further research should examine on which basis we best distinguish between product cues that can acquire evaluative valence and those that merely discriminate.
Publication status: published
KU Leuven publication type: IMa
Appears in Collections:Research Centre for Marketing and Consumer Science, Leuven
Department of Marketing and Organisation Studies (MO), Leuven - miscellaneous

Files in This Item:

There are no files associated with this item.


All items in Lirias are protected by copyright, with all rights reserved.