Reducing Bias from Choice Experiments Estimates in the Demand for Recreation

In valuing the demand for recreation, the literature has grown from using revealed preference methods to applying stated preference methods, namely contingent valuation and choice modelling. Recent attempts have merged revealed and stated preference data to exploit the strengths of both sources of data. We use contingent behaviour and choice experiments data to show that, with choice experiments exercises, when respondents are asked to choose which improvement programme they prefer for a site with recreational opportunities, failing to consider the information explaining the number of visits that respondents intend to take to a recreational site under each hypothetical programme leads to biased coefficients estimates in the models for the choice experiments data.