Differences in performance between experts and novices in many fields have been consistently noted in psychological literature, with an aim to try and find underlying causes of the "expert" phenomena. Wine tasting, describing and matching are no exception. However, this field poses some interesting challenges for the researcher. It is harder to test wine expertise in completely objective terms, such as chess or dinosaur naming (Chase & Simon, 1973, Chi et. al, 1989, both cited Solomon, 1997). The wine tasting community has it’s own sub-class of language and rituals (Lehrer 1975). The question that is most often asked by the researchers in this field is; Are the differences between novice and expert performance a genuine difference in perception, or is it a different cognitive organization of perceptions?

Early research centered on perceptual differences amongst groups by administering test that where based on "expert" guidelines (Solomon 1990,). It is no surprise then that experts performed better, given they where working in a milieu in which they had more experience. This type of argument tends to favors perceptual learning as the determinant of expert behavior (Lawless, 1985).

Work on absolute detection of odors was conducted by Bende & Nordin (1997). It was found that there was no difference in absolute detection of novel odors between experts and novices (Exp. 1). Odor is closely related to flavor, as both these perceptions are mediated by the olfactory organs (Lawless, 1985), and are thus relevant in wine perceptions. Bende & Nordin did find that experts outperformed novices in a recognition task, when scents common to wines where used. A cognitive arrangements for odors is proposed, given the transferability of the task.

Other research shows that differences in dimensions of description, and group categorizing might be caused by the features considered salient by novices. A novice is unlikely to be familiar with grape varieties, but will tend to group wines on single dimension salient to them, such as "sweetness" (Solomon 1997). Therefore it maybe that the disparate descriptions of wine by novices only reflect individual difference in considered salience, while experts are taught to think homogeneously on certain cognitive schemas. Since experts are "taught" to classify wines by common properties found in grape type, it is more natural for them to make descriptions that are homogeneous amongst themselves.

Lawless (1984) proposes that everyone holds a "library of odor terms, depending on our real life experience"(pg 102). These odor terms can serve as cues when tasting wine. Since experts have a more systemized library of odor terms, classified by grape variety, would giving a similar cognitive arrangement to a novice facilitate recognition and classification of wine? Lawless experimented by getting novices to match expert descriptions in a force choice situation, with no significant findings. The question that he poses will try and be answered here; would novices match their own descriptions better? (pg. 107). Also, given a prototype schema of wine varieties, would novices be able to match wine to grape variety?

Subjects consisted of 35 Sydney University students, all from the first year psychology course and who participated as credit for their course. The experiment utilized 5 different wine varieties (Sauvignon Blanc, Semillon, Riesling, Chardonnay and Verdehlo), whose characteristics where quite distinct. The wines where in the moderately-cheap price range. Subjects where given two sets of five wine glasses, placed randomly in two lines and numbered A-B and 1-5 respectively. The glasses where painted on the outside, to avoid visual cues. Subjects where provided with three answer sheets, and where given water to cleanse their palettes between trials.

Subjects where shown how to smell wine, and where told to smell the first row of five wines carefully. They where then told to match the odors of the wine to the five prototypes. These prototypes consisted of three descriptors, generated from known properties of the grape variety. Subjects where also instructed to mark their confidence on a five point scale.

Subjects where randomly placed into two groups of 11 and 24. The former group was given a short list of wine descriptors. Five separate grouping of descriptors where given, each to correspond to a particular wine. The terms corresponded to possible properties of the grape variety being tasted. The second group acted as a control, and was not given any list.

Both groups where told to taste the first row of wines. Instructions on how to taste the wine where given. Subjects where told to write down descriptions of the wines A to E, with the experimental group using the short list provided. Without coming back to wines A-E, subjects where told to match the previous descriptions with wines 1-5. They where again given a five point scale to mark their confidence in picking the right wine. Finally, both groups where then given a written questionnaire, with 8 general wine questions (to asses level of expertise), and 4 personal questions on their wine consumption.

A two-tailed t-test on the data from Experiment 1 gave the results (t = 3.42, p<0.01). This suggests that novices can match wines to prototype descriptions after olfactory analysis at above chance levels. Lehrer (1975, Experiment 2, session 4) performed a similar experiment, but instead of having novices match the prototype, they had a forced choice between expert and bogus descriptions. Lehrer did not achieve significant results. It does however support research by Rabin and Cain (1984), who found that cueing of odors codes greatly helps recognizablity. Giving subjects a short list of odor terms encoded their expected perception of these flavors. It could therefore be speculated that while novices cannot pick absolute descriptions of wine, they are able to recognize characteristics of a wine comparatively. Perhaps novices will only measure desired dimensions of a wine when they are cued for this detection, and when they are given a point of reference, i.e. the other wines.

Experiment 2 also showed that overall, the 35 participants could match wines from their own descriptions at above chance levels (t=2.95, p < 0.001). Similar experiments have been unable to show this effect (Solomon 1990, Lawless 1984, Lehrer 1975). Solomon’s experiment is nearly identical, where novices are asked to match expert opinions (Exp. 1), with not significant results.

A surprising result showed that testing the experimental group against the control group turned up no significant findings. This would conclude that a short list lexicon has no effect on matching wines from descriptions. These findings are inconsistent with Hughson and Boakes (in preparation), who showed that novices provided with a short-list performed significantly better than controls. These discrepancies could be caused by the wine selection. The wines in this study where quite different in their flavor. Perhaps the level of the task was set too low, meaning that Short term Working Memory was more a factor in the matching than where the descriptions (Ericson & Delaney, 1998).

Analysis on the descriptions used turned out some interesting results. This data supported Solomon’s (1990) observation that both experts and novices use more terms to describe wines when given a lexicon to work from (see table 1). That this did not prove to show a difference in performance between groups shows that number of terms used is also irrelevant, suggesting that the description task in this case was substituted for a short term sensory retrieval.

A within subject contrast for the confidence variable gave a (F=5.458, p>0.05) on the interaction of correct answers and given task (Table 2). Though participants confidence levels in the prototype task match their scores, it they correlate negatively with the matching task. High levels of confidence suggest that subjects found the matching task easier than their performance shows, though they showed above chance performance on both. These anomalies could be further studied. One of the problems would be the homogeneity in age and wine experience of the group. Perhaps research into confidence levels can show that university students are generally more confident in their perceptory skills, regardless of true outcomes.

Past studies suggest that giving novices odor cues improves their performance (Rabin & Cain, 1984), as does the aid of a schematic organization (Hughson & Boakes, in preparation; short –list, Solomon, 1997; wine wheel). This is highly suggestive of a difference in cognitive approaches between novice and experts. As Lawless (1984) postulates, experts have only the advantage of learning these different patterns, commonly in biologically existent schemas, such as grape variety and region. Though this study could not show differences in short-list vs. controls, it has been shown otherwise that with the help of such schemas, novices can perform with greater accuracy.

Bende M & Nordin S. (1997). Perceptual learning in olfaction: Professional wine tasters versus controls. Physiology and Behavior, 62, 1065-70

Ericson K & Delaney P (1998) Working Memory and Thinking: Current issues in thinking and reasoning Ch. 6, Psychology Press, England

Hughson A & Boakes R (in preparation) The role of long term memory in wine expertise. University of Sydney, Australia.

Lawless H (1985). Psychological perspectives on wine tasting and recognition of volatile flavors. In Birch G and Lindley M (eds.) Alcoholic Beverages, London: Elsevier Applied Science

Lehrer A (1975). Talking about wine. Language, 51, 901-23

Rabin M & Cain W. (1984). Odor recognition: Familiarity, Identifiably and encoding consistency. Journal of Experimental Psychology: Learning, Memory and Cognition, 10, 316-25.

Solomon G. (1990). Psychology of novice and expert wine talk. American Journal of Psychology, 105, 495-517.

Solomon G (1997). Conceptual change and wine expertise. The Journal of the Learning Sciences, 6, 41-60