We empirically investigate the ef- fect of the payment mechanism on contingent val- ues by asking a willingness-to-pay question with one of three different payment mechanisms: indi- vidual contribution, contribution with provision point, and referendum. We e nd statistical evi- dence of more afe rmative responses in the refer- endum treatment relative to the individual contri- bution treatment, some weak statistical evidence of more afe rmative responses in the referendum treatment relative to the provision point treat- ment, and no statistical evidence of more afe r- mative responses in the provision point treatment relative to the individual contribution treatment. The relative credibility of the three payment mechanisms is also examined . (JEL H41, Q26)