Against the (Campaign) Grain: The Cross-Cutting Effects of Authoritarianism in Political Campaigns

Abstract

For decades politicians focused their campaigns on issues that primed feelings of normative or existential threat among voters. Whether national security, increasing crime rates, or civil unrest, these campaign appeals led voters to consider external political threats. We examined one particular issue, national security, and showed that Republican candidates in particular benefitted from campaigning on this issue. When the campaign context featured national security issues, Republican candidates benefitted as highly authoritarian Democrats increased their support of Republican candidates. Interestingly, similar appeals did not benefit Democratic candidates. We found that the political context raised perceptions of threat and increased the influence of authoritarianism among voters, but these perceptions were asymmetrical across partizan and psychological divides. This asymmetry led to drastically different candidate evaluations and political behavior.

Keywords:

• authoritarianism
• campaign strategy
• national security
• threat

ACKNOWLEDGMENTS

The authors thank Paul Goren, Joanne Miller, Howie Lavine, Gene Borgida, Geoff Sheagley, the University of Minnesota American Politics Proseminar, the University of Minnesota Political Psychology Proseminar, and the discussant and panelists from the 2013 meeting of the Midwest Political Science Association for helpful comments throughout this project. In addition, they are thankful to Pierce Ekstrom, Hannah Kim, Christina Farhart, Brianna Smith, Jake Appleby, Allison Williams, and Aaron Rosenthal for access to their data from the 2012 University of Minnesota Center for the Study of Political Psychology Multi-Investigator Panel Study.

Notes

1 While authoritarianism is connected with needs for order and security for all individuals, we are particularly concerned with Democrats in this study. Although we present results across the partisan spectrum, the interest of readers should focus on Democratic voters and their tendency to shift right-ward and support Republican candidates. The nature of the US party system is such that right-ward shifts in preferences among Republican voters would simply solidify support for Republican candidates, making the results generally uninteresting for Republicans.

2 Data from all experiments are available from the corresponding author.

3 The dependent variables are a summary of branched questions. Respondents were asked “Do you approve or disapprove of the way [Bill Clinton/George W. Bush] is handling [his job as President/our relations with foreign countries]?” and this was followed by the question “Do you [approve/disapprove] strongly or not strongly?” These two questions were combined to form a four-point index which is scaled to run from 0 to 1, with 1 representing strong approval and 0 representing strong disapproval.

4 The child-rearing questions ask respondents “Please tell me which one you think is more important for a child to have:” and this stem is followed by four paired attributes. The four attributes are “Independence or Respect for Elders,” “Obedience or Self-Reliance,” “Curiosity or Good Manners,” and “Being Considerate or Well Behaved.” The attributes of Respect for Elders, Obedience, Good Manners, and Well Behaved are scored as authoritarian responses.

5 Replication code can be found in the online appendix.

6 The surveys and manipulations for both experiments are available from the authors by request.

7 While individuals who participate in Mechanical Turk surveys are clearly not representative of the national population, the system is ideal for fielding experiments quickly and affordably. Scholars show that Mechanical Turk samples are more representative than in-person convenience samples (Berinsky, Huber, and Lenz 2012; Buhrmester, Kwang, and Gosling 2011; Mason and Suri 2012) and recent research demonstrates that the platform is appropriate for experimental research that does not demand a representative sample, as the respondents appear psychologically consistent with the general population (Paolacci and Chandler 2014). Additionally, previous studies have utilized Mechanical Turk samples to successfully replicate experimental results concerning psychological mechanisms that closely resemble authoritarianism (Craig and Richeson 2014; Crawford et al. 2013; Crawford and Pilanski 2014). These results support our claim that psychological tendencies operate similarly among Mechanical Turk workers as they do in the general population. Additionally, because of the quickly changing nature of national security issues, studies need to be fielded quickly, which is possible with Mechanical Turk but significantly more difficult with more representative samples.

8 These advertisements were created by examining the current discussion by American politicians about the situation in Syria. We decided to use the Syria case because the parties had not formed crystallized positions on intervention, allowing us to credibly present both candidates with the same positions. The advertisements were created by a local political communication specialist.

9 Unfortunately, while a partisan identity manipulation check was included, we did not include a manipulation check about threat activation in this experiment. The advertisement used, however, was developed after a pre-test of issues which showed the highest levels of expressed fear for advertisements concerning nuclear and chemical weapons, as opposed to those dealing with crime, immigration, or instability in the Middle East. Mean levels of fear in the pre-test were significantly higher at the p < .01 level. Pre-tests also showed that fear, more than anger, was aroused by these advertisements.

10 This question asked respondents the following question: “We’d like to get your feelings towards some of our political leaders and groups who are in the news these days. Please rate each of these people or groups on something we call the feeling thermometer. Ratings between 50 degrees and 100 degrees mean that you feel favorable and warm toward the person or group. Ratings between 0 degrees and 50 degrees mean that you don’t feel favorable toward the person or group and that you don’t care too much for that person or group. You would rate them at the 50-degree mark if you don’t feel particularly warm or cold toward the person or group.” Respondents then rated the “Barack Obama,” “Mitt Romney,” “The Republican Party,” “The Democratic Party,” “Liberals,” “Conservatives,” “Alex Johnson,” and “John Sanders.”

11 The 12 conditions occur because each candidate could be assigned one of four speeches (Security, Health Care, Leadership, or Family) and these were crossed such that a respondent could not see the same speech from both candidates.

12 The national security speech attributed to Romney/Obama was not based off of actual text. The speech was written by the researchers to approximate a political speech that could credibly be given by either major party candidate.

13 The three collapsed condition for Romney were Romney Security Speech/Obama Health Care Speech, Romney Security Speech/Obama Leadership Speech, and Romney Security Speech/Obama Family Speech. The Obama conditions reflect this same pattern.

14 The online appendix includes the political knowledge questions used in the two experiments.

15 To preserve some sample size, we do not break down partisanship in these analyses by strong, weak, and leaning Democrats. Instead we group all three categories together.

Additional information

Funding

The authors graciously thank the Center for the Study of Political Psychology at the University of Minnesota for two research grants that enabled the completion of this project.