PubAg

Main content area

Various paired preference tests: experimenter effect on "take home" choice

Author:
Weiss, B.H., O'Mahony, M., Wichchukit, S.
Source:
Journal of sensory studies 2010 v.25 no.5 pp. 778-790
ISSN:
0887-8250
Subject:
chocolate, consumer acceptance, prediction, product testing, purchasing
Abstract:
Three types of paired preference test were performed on a total of 213 consumers. In each test, consumers were required to choose between two types of chocolate. They performed the traditional preference test which measured relative degrees of liking. Also , they performed a "Choosing" preference tests where they were required to specify which of two chocolates they were more likely to choose when offered both or whether they might choose either or neither. They also performed a "Buying" preference tests where they specified which of two chocolates they were more likely to buy or whether they might buy either or neither. After testing, they were able to take away either one type of chocolate or two types of chocolate. For one group of consumers, this was done while the experimenter was not watching; for the other group, the consumers were being observed. Observing the consumers as they selected chocolates to take away had a biasing effect. Correspondence between the three types of test and what consumers took away with them was low. The results were discussed in term of "test" preferences and "operational" preferences. Accurate prediction of acceptance of a product in the marketplace is important. The goal of any test of acceptance is to predict consumer behavior. Therefore, an acceptance test is only valid in terms of how well it predicts the behavior of the consumers in the real world, once they have left the testing situation. It is costly if such tests do not predict this behavior accurately. Accordingly, validation studies should monitor the behavior of the consumers for a requisite amount of time after the test. This has been done for a few studies involving hedonic scales, but there are no published reports for paired preference tests. There is a reluctance to perform such studies because of the time and costs involved. One alternative is to offer consumers the chance to select the products being tested, to take away with them once the test has finished. This might give a clue to "real world" behavior. Since this approach is simple and cost-effective, it is worth investigating.
Agid:
410779