Can citizens' opinions meaningfully guide government policy? Many political scientists fear the answer is no. While citizens voice opinions about public policies, these "opinions" often seem to be inconsistent with their underlying values and interests. These opinions change with minor differences in question wording or context. Moreover, the vast majority of citizens appear to know very little about the public policies they purport to support or oppose. Yet, determining whether citizens' policy opinions are "principled" — that is, consistent with their values and interests — is challenging, and existing approaches rely on questionable assumptions. This dissertation challenges the consensus that citizens' policy opinions are unprincipled. Using high-quality panel surveys, I show that opinion stability is a weak signal of how principled a policy opinion is. In a cross-sectional survey with a representative sample of American adults, I show that citizens associate values and policy opinions in sensible ways, regardless of their level of education and political interest. Finally, with a pre-registered experiment, I show that citizens can use policy information to successfully align their policy opinions with their values. The broad conclusion of this dissertation is that public opinion is more principled than existing research suggests. As such, we should be less concerned about whether citizens' policy opinions can meaningfully guide government policy.