In modern democracies, political representatives often must use several different conflicting considerations when making policy decisions. I focus on two such considerations that are commonly thought to be in direct conflict: responsiveness to the wishes of the public and a commitment to do what is in the best interests of the public. Because the public is often mistaken about the likely outcomes of their preferred policies, if representatives commit themselves to being responsive to the wishes of the public then there will be many cases in which they cannot do what is in the best interests of the public (and vice versa). Additionally, the policy preferences of the public are often unstable or dependent upon the views of political elites. However, many theories of democratic governance (especially those in the American tradition) maintain that political responsiveness is key to political legitimacy. Therefore, it is not satisfactory to simply ignore the policy preferences of the public. I argue that this seeming dilemma can be reconciled if we model the wishes of the public in terms of preferences over possible policies and the interests of the public in terms of preferences over possible outcomes of those policies. Consequently, policy choices can be made taking into account both explicit policy preferences (wishes) and outcome preferences (interests). Although such a model substantially simplifies policy decisions, it requires that preferences over policies and outcomes be describable by interpersonally comparable cardinal utility measures. I hint at what such measures might be like and show that this has important implications for public opinion polling methodology.