Paired Comparison (Preference) Test
In this test the assessor is presented with two coded products and asked to indicate whether there is a preference between them. The test design should ensure that each sample is assessed equally often in first and second position.
The panel size should be at least 50. If the panel is drawn from staff on site, care should be taken to exclude people who may have particular knowledge of the nature of the work or knowledge of the objectives for carrying out the work. Bear in mind that such a panel is
unlikely to be representative of the target consumer population.
At the simplest level the assessor is asked to state which sample is preferred and to offer reasons for preference. No - preference discussions are usually allowed and although, they are excluded from analysis, they are usually reported. This is a two - tailed test, as it is not known in advance which product is preferred, and both directions are of equal interest. The basic statistical analysis is by reference to two - tailed binomial tables, and the reasons for preference are tabulated.
Comparative assessments of acceptability or preference can be undertaken using the paired (preference) method or by the ranking test.