The role of 'reference goods' in contingent valuation: should we help respondents to 'construct' their willingness to pay?

A general population sample of 135 Australian respondents completed one of four contingent valuation surveys that asked them to value health benefits either in the absence of an explicit reference good or in the presence of one of three different forms of reference good. Results suggest that respondents have a 'ball-park' figure that is then challenged by the reference good. For values that appear far lower than, or similar to, this 'ball-park' figure, the reference good has little quantitative effect, but qualitatively appears to help respondents in their confidence in this value being their actual WTP. The implications for CV research in health care are outlined in the discussion.