This paper explores methodological issues surrounding the use of discrete choice experiments to elicit values for public goods. We develop an explicit game-theoretic model of individual decisions to a series of choice sets, providing general conditions under which surveys with repeated binary choices are incentive compatible. We complement the theory with a framed field experiment, with treatments that span the spectrum from incentive compatible, financially binding decisions to decisions with no direct financial consequences. The results suggest truthful preference revelation is possible in surveys, provided that respondents view their decisions as having more than a weak chance of influencing policy.