A new study in PNAS Nexus suggests that public opinion about Donald Trump’s criminal prosecution was largely unmoved by messaging from either Trump or his federal prosecutor. In a survey experiment with Republicans and independents, only modest effects emerged: legal messaging reduced support for Trump among those already skeptical of him but triggered backlash against the prosecutor among his supporters.
Criminal prosecutions of political leaders have become increasingly common across democratic countries. Since 2000, more than 75 former heads of government have faced legal charges after leaving office. In the United States, Donald Trump’s multiple indictments and eventual conviction became central issues in the 2024 election.
The legal and political stakes raised pressing questions: How do different narratives about such prosecutions shape public opinion? Does attacking the prosecution mobilize partisan support and undermine democratic norms? Or can legal defenses sway public attitudes?
While media reports have speculated on these effects—citing everything from Trump’s increased fundraising to predictions of democratic backsliding—few studies have directly tested them. The new study aimed to fill that gap by measuring how voters respond to elite messaging about legal accountability.
“In democracies globally, including the United States, political leaders have faced criminal prosecution after leaving office. Yet we know little about how these prosecutions affect public opinion. We explored how prosecutions shape public opinion by studying the valuable case of Donald Trump’s prosecution during the 2024 elections,” said study author Andrew O’Donohue, the Carl J. Friedrich Fellow in Harvard University’s Department of Government and a non-resident scholar at the Carnegie Endowment for International Peace.
To investigate these questions, the researchers designed a preregistered survey experiment involving nearly 3,000 U.S. adults who identified as Republican or independent. The study was conducted online between October and December 2023, before the 2024 Republican primaries. Respondents were randomly assigned to one of three conditions: a video of Donald Trump denouncing his prosecution, a video of special counsel Jack Smith defending the prosecution, or a placebo video about artificial intelligence.
The Trump video featured familiar rhetoric. He called the charges “election interference” and portrayed the prosecution as an attack on democracy itself. The Smith video, by contrast, emphasized the impartiality of the rule of law, stating that legal standards apply to everyone equally. All videos were sourced from CNBC to avoid strong partisan cues from the media outlet itself.
The study focused on Trump’s federal prosecution for alleged mishandling of classified documents, chosen because legal experts regarded it as one of the strongest cases against him. The federal prosecutor in that case, Jack Smith, was also seen as a more neutral figure compared to elected Democratic officials.
The researchers measured four key outcomes: support for Trump, support for the prosecution and the prosecutor, support for democratic norms, and affective polarization (how warmly or coldly people feel about members of the opposing party).
The results challenge several assumptions about the power of elite messaging. Most notably, Trump’s rhetoric had no significant effect on any of the four outcomes. Among the Republicans and independents, watching Trump’s speech did not increase support for him, decrease support for the prosecution, or heighten hostility toward the opposing party. It also did not make respondents more likely to endorse retaliatory actions, such as prosecuting Democratic officials.
One possible explanation is that Trump’s message was already well-known. By the time of the survey, most respondents had been repeatedly exposed to his views. The researchers suggest that participants had likely internalized his rhetoric, making the additional messaging ineffective.
In contrast, the legal messaging from special counsel Jack Smith had more nuanced effects. Among those who did not already support Trump, watching the prosecutor’s video reduced their stated likelihood of voting for him in the Republican primary. The same group also became more likely to view the prosecution as politically neutral and consistent with democratic norms.
Yet these shifts came with trade-offs. Among Trump supporters, exposure to Smith’s message led to a sharp drop in favorability toward the prosecutor himself. This suggests a kind of backlash effect: even if the message persuades some voters about the legitimacy of the prosecution, it can also trigger hostility toward the messenger, especially among those loyal to the accused.
There were other subtleties in the findings. Trump’s rhetoric, while ineffective overall, did appear to slightly increase perceptions that the factual claims in the prosecution were true—but only among his own supporters. This may reflect a rhetorical strategy in which Trump concedes certain facts (such as possessing documents) while questioning the motives behind the prosecution.
Legal messaging also had modest effects on support for democratic norms. Among respondents who leaned Republican—but did not identify as strong partisans—the prosecutor’s message slightly increased opposition to retaliatory acts, such as firing the prosecutor or politicizing law enforcement. But these effects were not observed among strong Republicans or Trump loyalists.
Overall, both types of messaging had no measurable effect on affective polarization. That is, they did not make respondents more or less likely to dislike members of the opposing party.
“Our key finding is this: When citizens have strong attachments to a political leader, prosecutions have strikingly limited effects in reducing public support for those leaders,” O’Donohue told PsyPost.
While the study was well-designed and used a realistic format—videos from real press conferences—its findings may not apply to all contexts. The sample included only Republicans and independents, leaving open the question of how Democrats might respond to similar messages.
The research also focused on a particularly well-known figure: Donald Trump. Given the entrenched attitudes many people hold about him, it is possible that rhetoric might have stronger effects when directed at lesser-known political figures, such as state attorneys general or mayors facing legal trouble.
The researchers also note that their treatments were brief. A one-minute video may not be enough to shift deeply held beliefs. Yet this brevity mirrors how many people consume political information in real life, especially on social media.
Future studies could explore whether more sustained exposure to legal messaging, or message framing from different sources (such as non-governmental watchdog groups), might be more persuasive. The authors also encourage further research on other politicians and in other countries, noting that prosecutions of leaders are increasingly common worldwide but remain understudied in terms of public opinion.
The study, “The court of public opinion: The limited effects of elite rhetoric about prosecuting political leader,” was authored by Daniel B. Markovits and Andrew O’Donohue.
