The discrete choice or ”referendum” contingent valuation technique has become a popular tool for assessing the value of non-market goods. Surveys used in these studies frequently suffer from large non-response which can lead to significant bias in parameter estimates and in the estimate of mean Willingness to Pay. We investigate the properties of tests for sample selection bias and the losses made by applying estimators assuming no sample selection. The effects of sample selection bias can be sizable but bivariate probit estimation give unbiased estimates. A computationally straightforward test for sample selection bias is found to perform well.
More papers in SSE/EFI Working Paper Series in Economics and Finance from Stockholm School of Economics Address: The Economic Research Institute, Stockholm School of Economics, P.O. Box 6501, 113 83 Stockholm, Sweden Contact information at EDIRC. Series data maintained by Helena Lundin ().