CONSEQUENCES OF USING DIFFERENT QUESTION FORMATS IN CONTINGENT VALUATION - A MONTE-CARLO STUDY

Monte Carlo experiments were conducted to compare regression parameter and willingness-to-pay (WTP) estimates using the contingent valuation method's referendum and payment-card questioning formats. Strategic bias was controlled by making responses consistent with the true WTP. Two WTP models were constructed: linear and log-linear in explanatory variables. Based on the efficiency and mean-squared-error criteria, the payment-card model was superior to the referendum model. However, neither model outdid the other when the unbiasedness criterion was used. The performance of either model depends on the choice and distribution of the bids and payment-card categories.