How to Construct Individual Preference Scales

The present work is a methodological study to investigate the possibly construction of individual preference scales based on paired comparisons. The data material is consumer preference of pork in Denmark.

A set of 56 consumers tasted four samples of meat as paired comparison. All of the six possible pairs of four samples of meat were served to each consumer. In each comparison of two meat samples, the preference was scored on a 15 cm visual analog scale. The consumer was asked to mark on a line, which sample he or she preferred. A mark right between the two meat samples means that they equally well liked or disliked the samples. The closer the mark was to one of the sides the more was this sample preferred. The meat was served to the consumers in a sensory laboratory (ISO 8589).

After a 10 minutes break the consumer scored the same four samples of meat on a 15 cm visual analog scale. This makes it possible to compare the results appearing from the use of hedonic scales to the results appearing from the application of paired comparisons. Using the two methodologies results at the aggregated level basically in the same differences in preferences, but it seems that the consumers differentiates the samples more individual differences visible. This shows a heterogeneity among consumers, which questions the concept of "mean consumer".

For the consumers to rate a meat sample into a hedonic scale often requires use of preliminary sorting. The preliminary sorting may be related to family values or to cultural values - sometimes named "meta preferences". The choice between the use of hedonic scale or paired comparison therefore involves the question of which kind of preferences is of interest in a specific situation.