What qualifies as good or bad representation has plagued scholars and officeholders for years. These studies often focus on outputs of two general types: the distribution of particularistic goods, allocative responsiveness, and the member’s position on roll call votes, policy responsiveness. My dissertation unites these disparate literatures on representation, asking a simple but fundamental question: what are elites doing, and how do constituents respond? Early work on electoral incentives contends that a record of narrow distributive accomplishments is essential to winning reelection. Broader policy achievements, however, are believed to be either too difficult to take credit for, or of little import to the member’s constituency. My dissertation challenges the notion that member strategies and voter responses continue to operate along these lines. Compared to the early 1970s, actors in the contemporary political environment are better sorted along policy lines at every level, from elites in the beltway to activists in the electorate. This means that an officeholder’s core supporters are more interested in their record on policy. Moreover, by taking credit for centrist policy achievements, officeholders can avoid alienating moderates. For a first cut at constituent preferences, I examine what voters like about their representative, and how these trends have changed over time. Since the 1970s, the percent of respondents who identify distributive goods as a reason for liking their member of Congress has remained fairly flat, whereas justifications based on specific policies has risen steadily. Digging deeper, these data suggest that reliable voters, activists, and donors are most likely to provide policy justifications. Distributive answers are not correlated with these measures of electoral intensity. If distributive and policy strategies appeal to different electorates, then the latter may produce greater returns by motivating those groups most vital to reelection. To evaluate these claims further, I deployed a series of experiments to test the conditions under which voters are receptive to representatives who focus on broad policy achievements. I find that policy-based credit claims can be just as effective as pork. Building on these initial results, I fielded a second set of experiments using a simulated news story about two anonymous incumbents running for a redistricted seat. In the end, I find that constituents often prefer policy-based records, but this choice is driven by the issue area and ideological position of the policy. After examining representation from the voter’s perspective, I then examine novel data on the advertising strategies of members of Congress. If campaign messaging reflects a member’s perception of the respective benefits from distributive and policy work, then ad buys afford a unique way of studying electoral strategies under a budget constraint. Looking at the 2008 elections, I find that broader legislative appeals are actually more common than distributive credit claims, both in terms of the number of airings and in dollars spent. Finally, I synthesize many of the earlier findings by combining the advertising data set with survey data on voter behavior and legislative effectiveness. By merging the member-level campaign variables with individual-level data, I show that policy appeals are not only common in congressional campaigns, but also potentially more beneficial at the ballot box. Using a series of multi-level models, I find that spending on broader policy achievements is consistently correlated with vote choice, whereas the benefits of other forms of campaign advertising are less apparent. What is more, when the sample is restricted to incumbents, I find that those who prioritize policy ads are relatively more productive in office. This finding demonstrates the value of advertising data. These results suggest that the tradeoff between local distributive goods and broader policy is significantly misunderstood with respect to voters and representatives alike. The implications of these findings extend beyond the reelection prospects of a single incumbent. Indeed, the collective outputs of Congress depend upon whether individual members spend their time working on local goods or national policy. If officeholders overindulge on narrow particularistic goods, then the chamber fails to pass large policy accomplishments. Representation, I contend, is driven less by a motivation to allocate, and more from an incentive to run on policy. While the end goal may differ, individual incentives continue to undercut the outputs of Congress as a whole. If individual members are unwilling to compromise on policy issues, then Congress runs the risk of producing little more than gridlock and empty position-taking.