Strategic Heterogeneity in Policy-Level Positioning: Evidence from Congressional Campaigns

To what extent do politicians adopt consistent left-right policy positions, and under what electoral conditions does this consistency vary? These questions are central to the study of representation in American politics, yet current research provides limited insights into how elites position themselves on policy. We address this gap by combining an original corpus of campaign platforms with machine learning methods to estimate the left-right orientations of congressional candidates (n = 4,505) across six salient issue areas in U.S. House elections (2018-2022). Employing our validated measures, we uncover substantively meaningful within-candidate variation in positioning across issues and demonstrate that this variation is systematic. We show that (1) candidates’ policy positions closely align with district-level policy preferences, even after accounting for overall district ideology, and (2) policy-level positional inconsistency increases as a candidate’s district becomes less favorable for their party. These results challenge prevailing expectations about the extent and sources of heterogeneity in elite policy positioning, underscoring the need for policy-level measures to understand how candidates navigate complex electoral incentives.