Does the inclusion of a cost attribute in forced and unforced choices matter? Results from a web survey applying the discrete choice experiment

The cost attribute is of particular importance in discrete choice experiments, and this study is the first to explore the effect of a cost attribute on both forced and unforced choices. Patients' preferences for organisational characteristics in general practice in Denmark are elicited, and the cost attribute is operationalised as user fees for the consultation. A representative sample of 1435 respondents from the Danish population answered the discrete choice experiment in a web-based questionnaire with a random split including/excluding the cost attribute. The two groups were asked to make both forced and unforced choices in each choice set. Our results show that in the unforced choice utility and scale parameters were not affected and the rank order remained the same when a cost attribute was included. In the forced choice the test of equal utility parameters was rejected, and rank order, marginal rates of substitution, and variance was shown to differ between the two groups. We observed that the inclusion of a cost attribute tended to change underlying choice behaviour. Evidence of potential dominant preferences was found in all splits.