A recent article (Moskowitz 1996) presented ''evidence'' to demonstrate that consumer and expert panels are equally competent for rating products on their descriptive attributes. The purpose of this note is to show that the evidence presented by the Moskowitz article is of dubious value for the purpose of comparing panels. Some examples: the claim that the spread of the means indicates one panel's ability to discriminate better than the other is misleading - panel variability must be taken into account. Whereas the article's interpretation of high and low correlations favors consumers, there is another interpretation that favors experts, and neither opinion necessarily follows from the data, Also, a claim that both panels have equal predictive ability because principal component prediction models yield high R**2 values will be demonstrated to have minimal statistical substance by comparing the results to a similar analysis simulated with random data.
