Why Seeing a Candidate Lose Late Makes Us Suspect Fraud

When Donald Trump tweeted “STOP THE COUNT!” during the 2020 U.S. presidential election, he wasn’t just appealing to partisanship. A new study suggests that many people—even those across the political aisle—may have genuinely perceived fraud because of how the results were reported.

According to seven new experiments published in Psychological Science, the way partial vote counts are released can distort public perception. If a candidate starts out in the lead but later loses, people are more likely to think something went wrong—even if nothing did.

The Psychology Behind the Suspicion

The culprit is a cognitive glitch known as the cumulative redundancy bias (CRB). This bias causes people to place too much importance on early leads, forming impressions that are hard to shake even after all the votes are in. “The winner was viewed more poorly if they took the lead late in the count,” said study coauthor Moritz Ingendahl of Ruhr University Bochum.

Researchers ran seven studies with participants from the United States and the United Kingdom. Some experiments used fictional school elections, while others mirrored real-world scenarios—including Georgia’s 2020 presidential vote count. In all cases, impressions of the candidates and beliefs in election legitimacy changed depending solely on the order in which results were presented.

What the Studies Found

Across the seven studies, the researchers found:

• Participants rated early leaders more favorably, even if they ultimately lost (Study 1)
• People were more likely to suspect fraud when the winner took a late lead (Studies 2–4)
• This effect persisted even if the count was still in progress (Study 5)
• Providing rational explanations for vote order only slightly reduced the bias (Study 6)
• Partisan identity did not eliminate the effect—both Democrats and Republicans were influenced (Study 7)

In one of the most revealing tests, participants viewed actual data from Georgia’s 2020 presidential vote. When the late-lead sequence was shown as it happened—Trump leading early, Biden pulling ahead late—participants suspected fraud favoring Biden. When the exact same data was shown in reverse order, suspicions flipped.

The Bias That Ignores Context

The researchers emphasize that this distortion persists even when participants are given logical explanations. “Although the explanation reduced the effect, it did not eliminate it,” they wrote. Even participants who knew and supported the real candidates (Biden or Trump) were not immune. The CRB appears to be a deeply rooted way our brains process competitions: we favor whoever seems to be winning early on, and we struggle to update our mental models even when final results contradict that trend.

Should We Stop Reporting Partial Results?

So what can be done? The study suggests one bold idea: avoid reporting partial vote counts at all. While delaying results could introduce its own risks, such as fueling conspiracy theories about lack of transparency, the authors argue that there are better ways to communicate elections. Improved forecasting models and more public education about how and why vote tallies change over time could help.

In a polarized media environment, even the simple act of tallying votes in a different order might have changed public perception. The team behind the research hopes their findings can guide election officials and news outlets in how results are shared—and ultimately, help prevent false claims from taking root.

Study Details

Journal: Psychological Science
Publication Date: July 24, 2025
DOI: 10.1177/09567976251355594
Authors: André Vaz, Moritz Ingendahl, André Mata, Hans Alves
Funding: Deutsche Forschungsgemeinschaft (DFG Grant No. 538466518)