Before a single ballot was counted in the 2024 election, researchers had already forecast the winner—based not on polls, but on how optimistically each candidate explained negative events in their speeches.
A new study published in The Journal of Positive Psychology suggests that Donald Trump’s victory in the 2024 presidential election may have been anticipated by analyzing how optimistically he explained negative events during the final weeks of the campaign. Although both Trump and Kamala Harris began the race with similar levels of optimism, Trump’s explanations for bad events became increasingly hopeful as the election approached. This shift, the researchers found, predicted not only the winner but also the size of the victory.
Past research has shown that the way presidential candidates talk about negative events—whether they see these problems as temporary or permanent, isolated or widespread—can predict electoral outcomes. In particular, an optimistic explanatory style, where problems are framed as specific and fixable, has been linked to electoral success. Earlier studies relied on nomination speeches, looking at a single moment in the campaign.
But modern campaigns are long, dynamic, and data-rich, offering more chances to understand how messaging changes over time. With the tools of positive psychology and political discourse analysis available, the researchers aimed to find out whether changes in optimism across the campaign—not just at the start—could help explain who wins.
The study was partly inspired by conversations between Abigail P. Blyler, the study’s lead author, and Martin Seligman, a founder of positive psychology and the Zellerbach Family Professor of Psychology at the University of Pennsylvania. Their discussion during a university course on the science of well-being coincided with the unfolding 2024 race.
“Last spring, I was a teaching assistant for a new, one-time course Dr. Seligman was offering for Penn undergraduates: ‘Science of Well-Being.’ The course walked students through the history of positive psychology,” Blyler told PsyPost.
“One pillar of that history includes work on optimism, pessimism, and political outcomes. Lightning didn’t necessarily strike, but rather the interest emerged in conversation with Dr. Seligman in the wake of the course. It just so happened we were standing before an important historical moment: a hyper-documented political race and we had a tool/technique (CAVE) to answer the question: who will win? In other words, the prior literature, the method, and the moment all lined up!”
To study changes in optimism, the researchers examined campaign speeches and interviews from both Trump and Harris, focusing only on how each candidate explained negative events. These included issues that were framed as threats, harms, or challenges for either the country or the candidate personally.
The research team collected verbatim transcripts from each candidate’s nomination speech, their debate on September 10, and 24 other speeches or interviews (12 from each candidate) spanning mid-September to October 27, just days before the November 5 election. From these materials, they identified 1,389 instances in which a candidate discussed a bad event and explained why it happened.
The team used a method called Content Analysis of Verbatim Explanations (CAVE) to analyze these statements. CAVE is a tool developed in psychology to measure explanatory style by scoring how people describe the causes of negative events. Trained raters assessed each explanation on two dimensions: stability (how lasting the problem appeared to be) and globality (how widespread or important the problem seemed).
For example, blaming a bad outcome on a single, short-term policy failure scored as more optimistic, while saying a problem stemmed from deep, ongoing societal rot scored as more pessimistic. Higher combined scores reflected a more hopeless outlook, while lower scores indicated greater optimism.
The researchers calculated mean “hopelessness” scores across four campaign periods: (1) the nomination and debate period, (2) September 12 to October 10, (3) October 15 to October 19, and (4) October 21 to October 27. They then compared these scores between the candidates and across time.
To ensure their results weren’t being driven by other speech characteristics, the team also used a widely accepted linguistic analysis tool (LIWC-22) to evaluate each candidate’s emotional tone, focus on past versus future, and use of agency-related language (words that reflect confidence and control). These additional analyses helped test whether optimism was truly a unique factor or just another way of measuring positive mood or assertiveness.
Finally, to prevent hindsight bias, the researchers encrypted their findings and shared them with third-party verifiers before the election results were known, only revealing the encryption key after voting had concluded.
The most striking finding was that Trump’s optimism increased significantly over the final weeks of the campaign, while Harris’s optimism remained mostly steady. At the start of the campaign, including the nomination speeches and first debate, Trump and Harris had nearly identical levels of optimism. But in the closing days, Trump’s language became much more hopeful. His explanations for negative events shifted from permanent and pervasive causes to ones that seemed more temporary and solvable.
By contrast, Harris did not show a comparable change. While her level of optimism fluctuated slightly, there was no meaningful upward trend.
Trump’s late-campaign optimism correlated with his eventual win. Not only did he win the election, but the size of his victory mirrored the size of his optimism shift. This was especially notable given that most public polls had forecast a tight race or slight edge for Harris.
Another important finding was the sheer volume of negative events mentioned by each candidate. Trump discussed over 1,000 bad events during the campaign, more than four times the number mentioned by Harris. Yet he paired many of these references with optimistic explanations. This combination of spotlighting problems while offering hope may have enhanced his appeal, activating voter concern while reassuring them that he had solutions.
“We’ve understood from prior studies that how an individual naturally describes the causes of a bad event can tell us whether the individual is relatively optimistic or pessimistic,” Blyler said. “In the simplest sense, if I tend to attribute the bad things in my life to something related to me (e.g., a trait, a behavior; i.e., it’s my fault) and I believe it will persist forever, and I believe it will affect everything else in my life—attributes we call: internal, stable, global, then I’m very likely a pessimistic person.”
“Yet, if my tendency is to attribute the bad things in my life to causes that are outside myself, temporary, and specific (i.e., only affecting a narrow slice of my life), then I’m probably on the optimistic side of the spectrum. What we saw overwhelmingly with Trump is that while he named many more negative events, he attributed the causes largely to things outside of himself; affirmed they would not last forever (i.e., he would fix them upon entering office); and in many cases, the causes were specific.”
“It was not as if Trump was always more optimistic, rather; it was in the final weeks (mid to late October) that Trump became significantly more optimistic,” Blyler noted.
To test whether these results could have been driven by general emotional tone or confidence rather than optimism specifically, the researchers looked at other linguistic features. Optimism was largely independent from the use of positive or negative emotion, time-focused language, or agency. While agency was moderately related to optimism, only optimism showed a sharp divergence between the two candidates at the end of the campaign. This suggests that the optimism scores measured something distinctive, not just how confident or emotional the candidates appeared.
Although the study found that Trump’s late surge in optimism predicted the election outcome, the researchers cautioned against assuming a cause-and-effect relationship. It’s possible that Trump’s shift reflected internal campaign data showing favorable trends, which in turn boosted his mood and shaped his messaging. Alternatively, the change in his messaging style could have influenced public perception and voting behavior. The study could not distinguish between these possibilities.
Still, the findings raise intriguing questions about how candidates talk about problems. The research suggests that optimism in political speech is not fixed, but can shift during a campaign and that this shift may matter more than previously recognized. It also challenges the assumption that focusing on negative events hurts a candidate. In this case, Trump’s frequent references to problems, when combined with increasingly hopeful explanations, may have enhanced his appeal.
The study, “Optimism predicted Trump’s victory: explanatory style during the 2024 presidential campaign,” was authored by Abigail P. Blyler, D. Cepeda, A. Michael, A. Shahi, S. Tousignant, R. Wainer, and M. Seligman.
