"Quick and dirty numbers"? The reliability of a stated-preference technique for the measurement of preferences for resource allocation.

We investigated the reliability of an internet-based stated-preference survey to elicit preferences for priority setting using a conjoint study like approach. Preferences were elicited among members of an Internet survey panel in an experimental "allocation of points" task at two times. The main finding is that the survey showed good reliability and most participants consistently adjusted their allocations of points to differences in presented scenarios. At the repetition of the survey, respondents were more likely to prioritize between new treatment programs competing for funding and those that did prioritize discriminated stronger between programs. We found no evidence that respondents were making easy choices or arbitrarily "clicked" through the survey.