To date, much of the policy and research debate on contingent valuation mode effects has relied on experiences drawn from other research disciplines. This study provides the first contingent valuation phone-mail comparison that meets current standards for response rates, draws from a general population, is relevant to the valuation of general environmental goods, and allows comparisons with actual sign-ups. Consistent with previous research in other disciplines, social desirability bias is found in responses to subjective questions --thus leading to more environmentally favorable responses on the phone. However, this effect does not carry over to hypothetical participation decisions. Hypothetical bias is found in both modes. Yet, application of calibration methods using debriefing questions provided nearly identical values across modes. As such, neither mode appears to dominate from the perspective of providing more valid estimates of actual participation decisions. The selection of survey mode must be based on other criteria.