Ripping Yarn: Experiments on Storytelling by Partisan Elites

Abstract

This article explores the role of personalized storytelling by partisan elites using a content analysis and two experiments. By personalized storytelling, I mean a political message in the form of a narrative that includes a specific reference to an individual affected by an issue. Using a content analysis of party convention speeches, this article shows that presidential candidates tell an increasing amount of stories, particularly from 1980 onward. Through randomized Internet experiments and a general population sample, I demonstrate that personalized stories have a unique influence on the public by parsing out the independent causal effect of the storyteller’s partisanship, the personalization of the story, and the content of the story. Not only can stories change attitudes about issues, but personalized stories can also change how individuals evaluate the candidate telling the story. However, an impersonal story that only references a generic group, rather than a singular individual, does not improve the partisan storyteller’s favorability. Results suggest that modern presidential candidates might be motivated to tell more stories because personalizing an issue may improve their standing with the public.

Keywords:

• Experiments
• narratives
• campaigns
• persuasion

Acknowledgments

I greatly appreciate helpful comments from Lynn Vavreck, John Zaller, Jeff Lewis, Michael Tesler, Chris Tausanovitch, Alan Gerber, Greg Huber, Kim Fridkin, John Henderson, Celia Paris, and participants at the Experiments Workshop at Yale University’s Institution for Social and Policy Studies. I also thank Robert Green and Adam Rosenblatt at Penn, Schoen & Berland for fielding my experiments. I thank my research assistants Katerina Bernasek, Megan Couch, Joseph Malhas, and Anaika Miller for help with the content analysis. Finally, I thank the three anonymous reviewers and two editors for helpful comments and criticisms during the review process.

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/10584609.2017.1336502

Notes

^1. By contrast, thematic frames involve more abstract and impersonal narratives that try to connect to a larger, ongoing theme without chronicling the life of an individual or group.

^2. Appeals to stories are mutually exclusive from appeals coded as arguments. A complete argument is built from premises and conclusions that work together to create a complete argument. An appeal is only coded as a premise or conclusion if it is part of a complete argument; otherwise, the appeal is coded as an enthymeme (a premise without a conclusion, or a conclusion without a premise). Both subcodes are considered “argumentation.” See supplemental Appendix for greater detail on the coding scheme.

^3. The three storytelling subcodes were identified qualitatively as appeals to stories before undertaking the content analysis, but each type of story was not formally coded and counted until the review process. As a result, I cannot report reliability statistics for specific storytelling subcodes.

^4. Sample size limitation potentially makes the Independent difference insignificant at a 95% level.

^5. Focusing solely on the partisan cue group, these results provide further evidence that simply associating a policy with a political elite acts as a “polarizing cue” (Nicholson, 2011), meaning that a partisan cue by itself can act to reduce support among out-partisans, but it does very little to increase support among in-partisans. This result suggests that candidates have to provide additional reasoning (like a story or an argument) to generate support.

Additional information

Notes on contributors

Andrew Gooch

Andrew Gooch is Postdoctoral Research Associate, Institution for Social and Policy Studies, and the Center for the Study of American Politics, Yale University.