Persuasion is at the heart of politics. Candidates persuade the public to vote for them. Politicians persuade each other to vote a certain way on legislation. Voters persuade their friends to share their preferences. Yet, as important as persuasion is in American politics, we know relatively little about whether or when it is possible and how it works.
In Persuasion in Parallel, Alexander Coppock tries to persuade readers that persuasion is possible. He introduces the persuasion in parallel hypothesis, which states that “the treatment effect of persuasive information on target policy attitudes is small, positive, and durable for everyone” (p. 31). In other words, people modestly update their policy preferences in the direction of relevant information to which they are exposed. He uses the metaphor of parallel lines to represent the idea that although two groups of people might start with different preferences (i.e., different Y-intercepts), persuasive information causes them to update their preferences in the same direction by about the same amount (i.e., the same slope). Coppock’s argument pushes back against the notion that people hold so steadfastly to their priors that they resist contradictory information. At a time when public discourse centers on information—its quality, veracity, bias, and influence—Coppock shows that information matters. I expect a rich intellectual exchange to follow from this important book.
Persuasion in Parallel is an excellent book with much to praise. The book is refreshingly clear, with a coherent argument reinforced by effective visualizations and empirical tests. Chapters 3 and 4 introduce the argument and research design and are a model of precision for scholars to emulate. The evidence presented in the empirical chapters is impressive, synthesizing results from 23 persuasion survey experiments, including original experiments, replications, and reanalyses (Table 4.2, p. 67). Coppock ambitiously uses a panel design to evaluate whether treatment effects are durable (chapter 6). His data strategy is precisely tailored to his argument, allowing for clean interpretations of the results. To analyze and present 23 experiments is an enormous undertaking, and the results are efficiently and effectively communicated (e.g., chapter 5).
Coppock provides convincing evidence that people update their attitudes in the direction of information. The straightforward argument speaks volumes in debates about voter competence, specifically challenging motivated reasoning accounts of information processing. The evidence tightly supports his argument, but there is room for more theoretical development around the persuasion in parallel hypothesis. Here I highlight four directions in which the persuasion in parallel hypothesis could be extended to enrich our understanding of persuasion.
First, Coppock repeatedly shows that persuasive information causes small changes in preferences in the direction of information, but there are few—if any—results where persuasive information causes people to cross the “midpoint,” such that they actually flip sides. Because voters are often faced with binary choices, nudging people’s attitudes in one direction or another might not be politically consequential if they do not ultimately vote differently. Coppock’s argument does not hinge on people completely flipping sides, but it will be important for future research to build on the foundation laid out in Persuasion in Parallel and examine the political implications of persuasive information. For instance, future work could extend the persuasion in parallel hypothesis to consider whether repeated exposure to persuasive information nudges people closer to flipping sides and whether people have a threshold for how much their opinions can change in response to persuasive information.
Second, one implication of the persuasion in parallel hypothesis is that understanding that people on the other side are capable of updating their preferences in the direction of information might soften the way we think about them (see pp. 1, 15, 16, 141). This potential implication is worth testing, particularly in the interpersonal contexts to which Coppock alludes. The rich literature on political discussion and persuasion within social networks might lead us to be skeptical that people would notice if they were able to slightly nudge their peers toward their side. Coppock is right that there could be important implications of the persuasion in parallel hypothesis for polarization, and future work should carefully engage with this idea.
Third, the persuasion in parallel hypothesis suggests that persuasive information changes minds for everyone (p. 49). Coppock defends this claim by analyzing the effect of persuasive information among many different groups of people. He primarily focuses on separately analyzing groups of people likely to support or oppose the target policy, either using their pretreatment preferences or partisanship. This distinction is theoretically important because if policy proponents and opponents responded to persuasive information differently, then persuasion would not be “in parallel.” Coppock repeatedly finds that people update their preferences in the direction of information, regardless of their priors and regardless of the groups to which they belong. The robustness of his core finding is remarkable and should give readers confidence in the persuasion in parallel hypothesis and perhaps some skepticism about motivated reasoning.
Although analyzing subgroups based on prior attitudes is theoretically motivated, there is room for more theoretical development around the other demographic divisions analyzed in the book. It is noteworthy that the pattern is “more or less universal” (p. 49) among the groups analyzed along the lines of gender, race, ethnicity, and education. However, there are important literatures on gender, race, and ethnicity that could deepen our understanding of the reach of the persuasion in parallel hypothesis. Similarly, Coppock notes that his results are limited to English-speaking Americans who take surveys (like many studies in American politics) but posits that he would expect the pattern to hold among those who do not speak English (p. 50). Recent work on language and public opinion opens the door to reconsider this point more carefully (see Efrén Pérez’s and Margit Tavits’s 2022 book, Voicing Politics). As the United States continues to diversify, it will be important to develop theories about the conditions under which the persuasion in parallel hypothesis holds for everyone. Bridging the persuasion in parallel hypothesis with the racial and ethnic politics literature, with particular attention to language, socialization, and social networks, will push the field forward.
Finally, Coppock is clear throughout the book about what he means by persuasive information and acknowledges that it is limited to a narrow type of information that explicitly excludes group cues and the messenger. Isolating the causal effect of persuasive information on target attitudes requires stripping the information down to remove group cues and messengers, which sacrifices some external validity. This is a nontrivial external validity limitation because in the real world people are rarely exposed to persuasive information, as defined in the book. Until we know how much the persuasion in parallel hypothesis stands against group cues and messenger characteristics, both of which are usually available in reality, it is hard to consider how much we can expect the public to change their minds. Future research will need to dig deeper to consider whether features of the person communicating the persuasive information disrupts the persuasion in parallel hypothesis.
Like many field-defining books, Persuasion in Parallel answers its own research question and raises more for the future. This bold, thought-provoking book will reignite research on persuasion and challenge us to reconsider voter competence.
