The chapter shows that Brazilians’ attitudes towards Lava Jato became increasingly divided and more negative over time. Attitudes towards the crusade are sensitive to partisan preferences, especially affect for the Workers’ Party. Results thus point to the precarity of optimism and the importance of voters’ priors in assessing prosecutorial zeal. The chapter also relies on an experiment to investigate whether putting crusades at the forefront of narratives of Brazilian corruption elicits optimism, compared to narratives that focus exclusively on corruption. The results show that when voters fixate on the crimes they are more likely to experience negative emotions and more likely to be dissatisfied with democracy. However,the crime-oriented narrative also increases respondents’ external efficacy, whereas the investigation-oriented one has the opposite effect. This suggests that the attitudinal impact of Lava Jato is far from being uniformly in line with the optimistic story. Under certain conditions, pessimists might be right in warning that crusaders’ anti-political message does more harm than good to the view that politics is redeemable.

Under what conditions are people most likely to discuss politics? Our focus in Chapter 5 is on the moment of decision itself (Stage 2). We use three novel approaches to answer this question. The True Counterfactual Study asked participants to reflect upon and describe either political discussions in which they had recently engaged or political discussions in which they could have engaged, but chose to avoid. Comparing these descriptions revealed that avoided discussions had larger groups with more disagreement. We then used vignette experiments to manipulate various features of a conversation, finding that individuals were more likely to avoid a discussion if they were in the political minority, less knowledgeable than the others, or conversing with weak social ties. The Name Your Price studies asked people to report how much they would need to be paid to discuss various topics with different groups. Individuals demand more compensation to discuss both political and nonpolitical topics with those who disagree, especially when that disagreement is defined in terms of partisan identity.

Chapter 7 continues our examination of the Discussion stage (Stage 3) of the 4D Framework, aiming to better understand what people actually verbally express in political conversations. We use a series of vignette experiments and a lab experiment to examine the extent to which people express their true opinions to the group, or engage in other expressive behaviors like self-censorship, silencing, or conformity. We find lab experimental evidence of conformity in real conversations and our vignette experiments revealed that individuals were less likely to reveal their real opinions when they were in the political minority and when they were less knowledgeable. We also find that individuals who were less knowledgeable were more likely to have affiliation concerns that explained their expression behavior. Finally, we analyzed the transcripts from the conversations in the Psychophysiological Experience Study to examine variation in what and how individuals discussed politics. We found that weak partisans were most likely to "hedge" their language when revealing their opinions and political identities.

Our exploration of the 4D Framework uses an eclectic set of methodological techniques; Chapter 3 is an overview of the methodological core of our inquiry. We explain the key operationalizations of the 4D Framework and provide context and details for the studies that appear in multiple chapters throughout the book. We specifically describe how we measure contextual features of a discussion, such as disagreement, as well as the scales we use to measure individual dispositions, such as social anxiety. We explain the utility of psychophysiological data for our purposes and describe the research design details for the studies that use psychophysiological data. We provide details on our survey samples from which several analyses throughout the book are derived.