Policy Repositioning and Affective Polarization: A Ten Arm Randomized Experiment

• In Event: Foundations of Affective Polarization

Abstract

This project investigates the connection between policy repositioning (aka flip-flopping at the elite-level) and affective polarization. I argue that policy repositioning can increase distrust of the in-party and out-party. In addition, candidate repositioning causes partisans to increase their social distance from the out-party (using the four canonical survey questions about talking with, being friends of, neighbors of, and children marrying the out-party). Previous research has not investigated how repositioning might influence affective or social polarization, but these hypotheses build on evidence that affective polarization is rooted in attitudes toward the elite-level (Druckman and Levendusky 2019).

Using a randomized survey experiment with ten treatment arms and a sample of 4,226 respondents, I evaluate a hypothetical candidate (without partisan or ideological labels) who repositions on immigration policy. Unlike past research, my design uses a campaign control group for a liberal or conservative candidate, and in-office positioning as treatment comparison groups. The novelty of this design allows for all possible positions in the liberal and conservative direction to be accounted for, which has not been tested in previous repositioning experiments. That is, a liberal candidate during the campaign (control group) is compared with a candidate who held that same position during the campaign and now in-office is 1) consistent on policy, 2) repositioned to the left, 3) repositioned to the right, or 4) repositioned to the status quo policy. This design is mirrored for a conservative candidate too, resulting in ten randomized treatment arms.

I use several affective polarization measures. I asked questions that assess feelings toward individuals and politicians in the in-party, out-party, and four measures of social polarization. These social polarization questions were also asked among pure Independents with the word “partisans” inserted instead of the out-party.

Results show that candidate repositioning does alter views of in-group and out-group partisans, but not every variation of repositioning is significantly different from the control groups. Several results are briefly discussed here. For example, when liberal candidates reposition to the right or status quo, Democratic respondents have significantly lower opinions of Democratic people and politicians. This demonstrates that repositioning can also alter views of the in-party. Both Democratic and Republican respondents increased their perceived importance of their own party when candidates reposition on immigration. Repositioning caused Republican respondents more than Democrats to say that they are less likely to talk to/be friends with/have neighbors/approve of children to marry the out-party. Lastly, results show that repositioning can also cause Independents to increase their social distance from partisans.

Author

• Andrew Gooch, Rowan University