Sensory Evaluation Errors
So if we're not adapting or compressing, it's all good, right? Unfortunately our big juicy primate brains get us into trouble again. Adaptations derived from social hierarchy, or expectations driven by previous experience can push us into sensory errors, including the following: Stimulus error: when irrelevant criteria such as the style or colour or the container (or the product itself) influences the observer. E.g. Wine bottles with cork and screw top on the same product will get differently assessed because of the closure method. Expectation error: Information given with the sample may trigger preconceived ideas. What is expected is usually found. E.g., if five wines in a row have increasing levels of sweetness, the sixth sample will usually be rated as sweeter—even if it isn’t! Enhancement: the effect of the presence of one substance/stimuli increasing the perceived intensity of a second substance/stimuli. E.g., given two identical wine samples, if the acidity in one is increased, the taster will also perceive an increase in fruitiness. Error of habituation: After a while, things all taste the same—never taste more than eight similar samples in one session. Mutual suggestion: The response of a panellist/taster can often be influenced by other panellists. Vocalising an opinion should be prohibited (and the testing area should be free from noise and distraction). Capriciousness and timidity: Some people tend to use the extremes of any scale thereby exerting more than their share of influence over the panel’s results. Others tend to stick to the central part of the scale and to minimise the difference between scales. This effect can occur when you ask for an honest response in front of an authority figure, and can minimise objectivity and honesty. Contrast effect: Presentation of a sample of good quality just before one of poor quality may cause the second sample to receive a lower rating than if it has been rated alone. Group effect: One good sample presented in a group of poor samples will be rated lower than if presented on its own. This effect is the opposite of the contrast effect. Error of central tendency: Samples placed near the centre of a set tend to be preferred over those at the ends. In triangle tests, the odd sample is detected more often if it is in the middle position. Pattern effect: Panellists will use all available clues and are quick to detect any pattern in the order of presentation.
It's amazing we ever agree (or in the case of mutual suggestion, disagree) on anything, given the room for error. But a clear head, along with steps taken to eliminate bias can go a long way to decreasing the error rate of our tasting efforts. But what about our equipment, our noses and tastebuds? We'll look at the gustatory and olfactory systems tomorrow.