Brian Harrison and Melissa Michelson have written a compelling and accessible piece of social science that significantly enhances our understanding of how even strongly held attitudes can change. Building on existing research, they put forward a persuasive new theory of “Dissonant Identity Priming” and test that theory in a variety of experiments using different identities and contexts. The result is an excellent example of the way in which the resurgence of experimental methods has breathed new life into long-standing social science questions.

The project’s novel thesis builds on existing theory in political psychology, social identity, motivated reasoning, and the issue priming and framing literature. Dissonant Identity Priming contends that individuals can be persuaded to rethink strongly held attitudes when confronted with unexpected information or a message from a source with whom they share an in-group social identity. The authors contend that social identities are primed by well-known representatives of a group. If a well-known group member delivers a message, especially one that is unexpected or contrary to expectations, those with strong group identities will pause to consider this message. The pause will be enough motivation for some group members to reconsider their position on the relevant issue, even if their attitude is intense. For some group members, the priming of identity, a contrary message, and the consideration of the message will be enough to shift their attitude on the issue in the direction of the message. For example, if Charlton Heston, previous president of the National Rifle Association (NRA), had publically stated that he supported a ban on assault style rifles, Dissonant Identity Priming theory suggests that gun owners and NRA members would have given consideration to the message, and at least some might have shifted from strong opposition to a ban on assault style rifles to less opposition, or even supported for it.

Of course, the primary ideas behind the theory are not entirely new, and they have been used in politics before. An example that follows the logic of the theory is the political endorsement of a candidate by a popular figure from the opposite party. We see this in many recent high-profile elections, but until now we had little evidence that there might actually be an impact on voter preferences. The theory is also so compelling because it follows a giant wave of social science research that almost consistently demonstrates that our identities provide us with attitude predispositions that serve to filter incoming new information. Through a process described as motivated reasoning, most of us have the tools to reject contrary information, even when our existing beliefs can be shown to be objectively false. The process allows us to maintain consistency between our identities and our beliefs. Dissonant Identity Priming disrupts this process.

The book is also unique in that the authors partnered with advocacy organizations whose missions included efforts to influence public opinion, political behavior, and public policy on LGBT issues. All of the groups could be considered left-leaning social justice organizations. Of course, the lessons learned from the authors’ analyses could be applied by any type of advocacy organization, and there are plenty of examples of social scientists partnering with political campaigns to conduct field experiments. However, like concerns about who is funding a research project, these types of partner efforts should be entirely transparent. The authors have set a good example for what full transparency should look like with these research partnerships—what they call the practice of para-academics.

Harrison and Michelson test the theory by using experiments in a variety of conditions and across several issues. However, they focus their attention primarily on LGBT issues in four chapters. In these chapters they conduct analysis on 14 different experiments, mostly focused on same-sex marriage. In perhaps my favorite set of examples, the authors design five randomized experiments involving sports (fans and nonfans) in Chapter 3. The treatments were embedded in surveys and involved same-sex marriage equality or gay and lesbian antidiscrimination messages; three of the experiments were conducted online while two were conducted in person. Each experiment was designed to prime sports-fan identity, provide a pro-LGBT message (which would be a dissident message for professional sports), and produce an effect in some fans that was favorable to LGBT rights. The baseball and hockey fan experiments had minimal effects, but the National Football League football experiments were powerful. After the first such experiment, in which NFL Commissioner Roger Goodell endorsed nondiscrimination for gays and lesbians in the experimental treatment, support for sexual-orientation nondiscrimination laws increased 25% among NFL fans, but did not change among nonfans, relative to the control condition.

Likewise, the third experiment, which was conducted in person in the heart of Green Bay Packer country, Appleton, Wisconsin, also provided powerful support for the theory. Participant fan strength was assessed for all sports, the NFL, and the Packers. The primary subsequent treatment involved showing the participant a picture of Packers Hall of Fame player LeRoy Butler and telling the participant that Butler supports same-sex marriage. Participants were then asked about their support for same-sex marriage. Relative to the control condition, Packers fans support for marriage equality increased more than 14 percentage points. The influence of both experiments involving NFL fans held up in a multivariate statistical test controlling for factors such as age, education, and partisanship. The power of priming an NFL sports identity that might tend to have a conservative political predisposition with a dissident message of equality is impressive.

Perhaps just as importantly, Harrison and Michelson also establish that the psychological processes they theorize do actually function in the manner they suggest, by demonstrating how alternative specifications do not produce the same impact. For example, in one partisan-based experiment conducted in Virginia, the treatments involved cues from state party leaders on same-sex marriage. The partisan cues were consistent with stereotypical party positions on same-sex marriage. As the theory suggests, without a dissonant message, the treatments did not produce statistically significant changes among Democrats or Republicans. In short, the mechanism is not simply the message or priming of identity by the speaker; attitude change also requires an unexpected message.

Of course, Dissonant Identity Priming has its limits. The evidence suggests that shifts in attitudes might not produce changes in behavior, such as contributing money to campaigns. Additionally, context matters. One experiment that primed religious identity in the center of the Bible Belt did not produce expected changes in attitudes. Nevertheless, even these failed efforts highlight the utility of the theory through its falsifiability in particular contexts. Overall, I am encouraged by the utility of the theory and its potential for widespread application. Listen, We Need to Talk is an excellent addition to any course addressing experimental research methods, LGBT politics, American politics, and ethics in social science research.