The Case of the Price:Perceived-Quality Relationship
by P. Bowbrick
Academic research on a subject may consist entirely of well designed experiments competently carried out, and may still fail to produce any useful results or any advance in our knowledge. One line of research on the price-perceived/quality relationship over the past thirty-five years has followed this pattern, combining poor scientific method with good experimental technique to produce little of value. In 1945 Scitovsky  commented on the commonly observed phenomenon that people frequently judge the quality of a product by its price, assuming that the more expensive item is better, and he discussed the reasons for this and its implications. Since then there have been more than 70 academic articles and theses, implying perhaps 90 man years of research, testing the hypothesis that some people sometimes judge quality by price , but we know no more than we did in 1945. In the typical experiment university students were given a set of cards, each card bearing a description of a product and its price, and were asked to choose which product they thought they would buy if they had the choice. Statistical analysis showed whether, other things being equal, they were more likely to buy the more expensive product. The result of this enormous research effort was to show that American university students and a few other populations sometimes do appear to judge quality by price . What is the value of this result? There can be few businessmen who do not believe that sometimes, but not always, some people think that the more expensive good is better. The research confirmed their belief. However, once one experiment had confirmed their belief the other experiments were superfluous, at best providing a little extra corroboration for a self-evident hypothesis. Even if all the subsequent experiments had failed to show the relationship it would have meant nothing: nobody expects that the relationship would apply always and in all markets. The tests were incapable of showing that nobody ever judged quality by price: all they could do was to show that nobody appeared to do so in the experimental situation. The series of experiments did not show how frequently people judge quality by price, or how strong the effect is. A statistical estimate of the frequency and strength of the relationship might perhaps be made if a probability sample of a carefully defined range of market situations was taken and an experiment was carried out in each of them. Samples would be drawn from different populations depending on whether one was interested in the average product, the average consumer, the average transaction or the average market. No attempt was made to sample in this way.
The research programme made no attempt to sample situations typical of the real world. In most of the studies reported, the experimental design demanded atypical products like carpet or curtain material which the consumers would not be able to judge objectively or be able to recognise by brand or other characteristics. In some studies, the consumers did not see the product: they sat in a classroom and made their choices from written descriptions . The consumers were nearly always groups like students and were not typical of the population. Researchers usually put a ritual caveat in their reports, saying that someone should see if the experimental results on any product have any application to the corresponding market in the real world, but, so far, no one has done so.
Until experiments or observational studies have been carried out to compare the results of these laboratory experiments with purchasing patterns in the real world, we have no reason to believe that they give any indication of actual purchase behaviour even in the markets they do examine.
In spite of the obvious weaknesses of these studies, many of them admitted by the authors, nearly all...