Appendix - Details of Empirical Studies and Statistical Analyses\

Summary

DESCRIPTION OF STUDIES

2012 Campaign Study. These data came from a module placed on the 2012 Cooperative Campaign Analysis Project. The module consisted of 1,000 panelists living in the United States recruited by YouGov and interviewed in December 2011, the week just before the 2012 elections, and then again in the four weeks after the election. YouGov maintains a Census-matched Internet panel that generates nationally representative samples. We measured need for affect, need for cognition, and Big Five personality traits on the second wave of the panel. Demographics were measured on the first wave. Partisan identities were measured in the first and second waves. Voting decisions were measured in the post-election wave.

Governor Evaluation Study. In the spring of 2013, we fielded a web-based survey using Amazon's Mechanical Turk (MTurk). MTurk is an opt-in Internet survey in which participants complete brief surveys in return for monetary compensation ($0.50 in our study). Reasonable variation in the compensation rates for MTurk does not meaningfully affect the quality of the data (Buhrmester, Kwang, and Gosling, 2011). Although we cannot claim that this sample is representative of the United States, MTurk samples have not been found to differ from nationally representative samples in terms of psychological or political characteristics (Clifford, Jewell, andWaggoner, 2015), and they tend to be more diverse than student-based convenience samples (Berinsky, Huber, and Lenz, 2012; Buhrmester, Kwang, and Gosling, 2011). Our sample consisted of 661 adults living in the United States who identified with the party of their state's governor. Participants were first asked to answer a series of demographic and background questions, and then they were asked to complete the full need for affect and need for cognition batteries presented in Chapter 3. Participants were then randomly assigned to one of the conditions in the belief perseverance experiment described in Chapter 5. Of the 661 participants, 336 were randomly assigned to the shirking treatment and 325 were randomly assigned to the loyal treatment.

Emotion Processing Study. In the fall of 2014, we recruited 105 Temple University students to participate in a “visual processing study” in the Behavioral Foundations Lab, which is housed on campus.