Political Partisanship Without the Politics

Polarization of American partisans continues to increase.  Liberals and conservatives alike have obvious contempt for opposing partisans — this is universally demonstrated by implicit, explicit and behavioral indicators.  Shanto Iyengar and Sean Westood of Stanford University and Princeton University, respectively, designed a set of four studies — titled “Fear and Loathing Across Party Lines: New Evidence on Group Polarization” to further investigate political dichotomy in America.

1. What They Did – Intervention Summary:

Study participants completed selection tasks.  Each participant was randomly assigned to one of two tasks that modeled existing scholarship assessments.

Participants in the first task had to choose to give a scholarship to either a Democrat or Republican high schooler.   Those in the second task had to choose between a European American and an African American candidate.  The academic and extracurricular achievements of each candidate were randomly varied, which allowed the study to measure the effects of partisan and racial bias without qualification confounds and compare the relative strength of in-group preference.

2. What They Found – Results:

Despite the lack of direct political connection, this study found that the party cue had the biggest impact on candidate selection.  Approximately 80% of participants, Democrats and Republicans alike, who participated in the partisan design selection chose the candidate who identified with their own party – even when the candidate from the opposing party was more highly qualified.  There was no evidence that those who participated in the partisan design took academic achievement into account.

Participants assigned to the race design selection showed relatively weak effects of in-group bias and tended to select candidates based on qualification instead of race.

3. Who Was Studied – Sample:


4. Study Name:

Iyengar and Westwood et al. 2014, Study 2

5. Citation:

Iyengar, Shanto & Westwood, Sean J. (2014).  Fear and loathing across party lines: New evidence on group polarization. http://pcl.stanford.edu/research/2014/iyengar-ajps-group-polarization.pdf

6. Link:


7. Intervention Categories:


8. Sample Size:


9. Central Reported Statistic:

“Democrats were more likely to select a fellow Democrat (b=1.04, p<.01) and Republicans were more likely to select a fellow Republican (b=1.60, p<.001).”

10. Effect Size:

The probability of a partisan selecting an out-party candidate never rose above .3.