Weak Connection between Policy Repositioning and Affective Polarization

Abstract

This project investigates the connection between policy repositioning (aka flip-flopping) at the elite level and affective polarization. Although usually conceptualized as a mass-level phenomenon, affective polarization might be influenced by elite-level cues. Using a randomized survey experiment with ten candidate profiles, I evaluated a hypothetical candidate who repositions on immigration policy. Unlike past research, my design uses a campaign control group for a liberal and conservative candidate and in-office positioning as treatment comparison groups. Several affective polarization outcome measures were asked: the importance of one’s own party, feeling thermometers of partisan people and politicians, and four social polarization questions (talk with, live near, be friends with, and have a child marry the out-party). Results show that repositioning has minimal effects on affective polarization. Liberal candidates repositioning to a more liberal position (e.g., moving in the polarized direction) caused an increase in the importance of the party for both Democrats and Republicans. Several other polarized repositions also changed affective polarization. However, I find mostly null and inconsistent results for views of the out-party and questions about social polarization. Alternative model specifications and subgroup analysis did not change the null results. Alternative explanations are also considered. This suggests that candidates who break campaign promises on policy are not causally related to changes in affective polarization.