With the increase in employment of Internet surveys, there is a need to address how this survey mode affects inferences. Internet and interview modes are compared in estimating the preservation value of Dokdo, South Korea. Estimates of willingness to pay in the Internet survey are smaller than those in the interview survey. This may be attributable to the interview respondents being conscious of the presence of the interviewers, given the social importance of Dokdo to Koreans. Internet surveys may be effective in reducing overestimation of the willingness to pay in situations where potential social desirability is present.