Search

This is a pre-registration for an online field experiment. Users will begin with a survey, be matched with a partner in our custom-built chat environment, have a conversation about gun control, and then complete a post-survey. Some randomly selected individuals in the conversation will receive AI-generated prompts of alternative things they can choose to say throughout the conversation. We will simultaneously evaluate the technical capacity of the AI system to accomplish this task and the impact of the interventions on a number of measures of polarization and conversational quality.

Political scientists have been using individuals’ self-reported efforts to try to influence the votes of others as one indicator of political activism for more than a half-century. However, in spite of this widespread use, very little is known about the motivations of interpersonal persuasion. This dissertation examines why some people try to influence the votes of others during the course of their everyday political conversations, while others are content to discuss politics without trying to persuade. Although attempts to persuade are often treated as a form of campaign participation with a goal of influencing the outcome of the election, I find that the motivations for persuasion are more internal and interpersonal than are the motivations for other forms of campaign involvement. I argue that interpersonal persuasion should be treated as a form of discursive participation, with consequences for our understanding of public opinion and deliberation.I use three large-scale survey datasets to examine interpersonal persuasion as a distinct form of political participation: the 2012 American National Election Studies (Ch. 3), the Youth-Parent Political Socialization Panel (Ch. 4), and the 2008 National Annenberg Election Studies Phone Survey (Ch. 5). Each dataset has distinct advantages in design and content that allow for examination of particular features of the motivations of persuasion. Using a variety of statistical methods and data sources, the overall argument is that attempts at persuasion have relatively little to do with campaigns and elections, and much more to do with individual orientations towards politics and social norms about political discussion. In particular, I find that organizational membership and campaign mobilization efforts do not provide the driving force for attempts to persuade that characterize other forms of political activity. Additionally, persuasive behavior does not ebb and flow in expected ways relative to the campaign cycle or the competitiveness of races. Rather, internal (but not external) efficacy, political capital (e.g., political interest, attention to political news), and exposure to disagreement and attempts to persuade by other social contacts are highly predictive of persuasive behavior. Persuasive behavior also declines over the life cycle, unlike many other forms of participation, and it is not a stepping stone to other more costly or time-intensive forms of participation. Therefore, persuasive behavior is equally or better understood within a framework of discursive participation, where it is a mechanism for people to develop their political identities, form political opinions, and process political events within the context of their social networks.

There is mounting concern that social media sites contribute to political polarization by creating "echo chambers" that insulate people from opposing views about current events. We surveyed a large sample of Democrats and Republicans who visit Twitter at least three times each week about a range of social policy issues. One week later, we randomly assigned respondents to a treatment condition in which they were offered financial incentives to follow a Twitter bot for one month that exposed them to messages produced by elected officials, organizations, and other opinion leaders with opposing political ideologies. Respondents were re-surveyed at the end of the month to measure the effect of this treatment, and at regular intervals throughout the study period to monitor treatment compliance. We find that Republicans who followed a liberal Twitter bot became substantially more conservative post-treatment, and Democrats who followed a conservative Twitter bot became slightly more liberal post-treatment. These findings have important implications for the interdisciplinary literature on political polarization as well as the emerging field of computational social science.

© This open access article is distributed under Creative Commons Attribution-NonCommercialNoDerivatives License 4.0 (CC BY-NC-ND). How predictable are life trajectories? We investigated this question with a scientific mass collaboration using the common task method; 160 teams built predictive models for six life outcomes using data from the Fragile Families and Child Wellbeing Study, a high-quality birth cohort study. Despite using a rich dataset and applying machine-learning methods optimized for prediction, the best predictions were not very accurate and were only slightly better than those from a simple benchmark model. Within each outcome, prediction error was strongly associated with the family being predicted and weakly associated with the technique used to generate the prediction. Overall, these results suggest practical limits to the predictability of life outcomes in some settings and illustrate the value of mass collaborations in the social sciences.