ClearFactAll topics

Is election fraud in the U.S. significant enough to impact results?

2 peer-reviewed studies for, 7 against

Key takeaways

Stronger evidence: Against

Verdict

The scientific consensus regarding U.S. election fraud is that while isolated instances occur, they are not significant enough to impact election results. Landmark research published in the American Political Science Review (2020) utilized administrative records to estimate the prevalence of double voting at approximately 0.025%, a rate so low that it is largely indistinguishable from clerical error. Conversely, studies that claim significant fraud often rely on survey data where small percentages of respondents provide incorrect citizenship information; for instance, the 2014 Richman study was largely discredited when a 2015 rebuttal in Electoral Studies showed that its findings were statistically indistinguishable from measurement error. Furthermore, post-election audits and hand recounts in high-stakes contests—including the 2020 presidential election—have repeatedly confirmed the accuracy of machine counts and found no evidence of systemic malfeasance. While theoretical vulnerabilities in voting machines and voter rolls are legitimate subjects for security improvements, the empirical reality across millions of ballots cast shows that fraud is vanishingly rare. Therefore, the weight of evidence strongly supports the conclusion that election fraud is not a significant factor in determining the outcome of U.S. elections.

Strongest study on each side

Ranked by citation count.

Strongest study for

Strongest study against

No evidence for systematic voter fraud: A guide to statistical claims about the 2020 election (PNAS, 2021)

74 citations

The study reviews widely circulated statistical claims about fraud in the 2020 U.S. election and finds that none provide credible evidence of systematic fraud. It shows that alleged “anomalies” (e.g., turnout patterns, voting machines, or county results) are either based on incorrect facts or are consistent with normal election behavior. Overall, the authors conclude that statistical evidence does not support claims that the election was stolen.

Open the interactive page →

The evidence

For (2)

Against (7)

No evidence for systematic voter fraud: A guide to statistical claims about the 2020 election (PNAS, 2021)

74 citations

The study reviews widely circulated statistical claims about fraud in the 2020 U.S. election and finds that none provide credible evidence of systematic fraud. It shows that alleged “anomalies” (e.g., turnout patterns, voting machines, or county results) are either based on incorrect facts or are consistent with normal election behavior. Overall, the authors conclude that statistical evidence does not support claims that the election was stolen.

Statistical detection of systematic election irregularities (Proceedings of the National Academy of Sciences, 2012)

Peter Klimek, Yuri Yegorov et al. · 28 citations

Klimek et al. developed parametric statistical methods to detect systematic election fraud by analyzing vote distributions across electoral districts. The technique identifies statistical anomalies (high kurtosis) that indicate ballot stuffing and fraud, showing it is detectable through vote pattern analysis in elections. Applications to US elections would presumably reveal such patterns if systematic fraud were significant.

Securing Elections, Securing Confidence? Perceptions of Election Security Policies, Election Related Fraud Beliefs, and Voter Confidence in the United States (Election Law Journal: Rules, Politics, and Policy, 2024)

Joseph A. Coll

Study examining how voter perceptions of election security policies designed to prevent and monitor fraud affect fraud beliefs and voter confidence. Results show voters believe there is less election-related fraud and are more confident when they perceive their state employs effective security policies.

American Views About Election Fraud in 2024 (Frontiers in Political Science, 2024)

Mitchell Linegar and R. Michael Alvarez

Survey of 2,211 U.S. registered voters shows substantial proportions believe election fraud is common, but partisanship and conspiratorial thinking strongly predict fraud beliefs rather than actual evidence. Findings support that voter fraud concerns don't reflect systematic electoral problems but rather reflect partisan and conspiratorial narratives.

The debate

Mark: David, we can't just dismiss concerns about election integrity as "conspiracy theories." There is peer-reviewed research, like the 2014 Richman study in Electoral Studies, suggesting that non-citizen voting could be significant enough to flip Senate seats or even the presidency. If the system has holes, they'll be exploited.

David: Mark, that same journal published a devastating rebuttal just a year later. The "non-citizens" in that study were mostly citizens who accidentally checked the wrong box on a survey. When you account for that simple measurement error, the evidence for non-citizen voting virtually disappears. We’re talking about an infinitesimal fraction of the electorate.

Mark: But it's not just survey data. John Lott's 2022 study in Public Choice pointed to statistically significant anomalies in the 2020 election results in Georgia and Pennsylvania. He argued that the shifts in vote totals didn't align with traditional voting patterns and could only be explained by irregularities. Why is it "irresponsible" to look at the math?

David: Because the "math" often ignores context. Those shifts are better explained by changes in voting methods, like the surge in mail-in ballots during the pandemic, and shifts in demographic leanings. Every time these claims are brought to court or subjected to a hand recount—like we saw in Arizona and Georgia—the results come back the same. If there was a massive fraud machine, it would leave a paper trail, but we only find clerical errors.

Mark: A "clerical error" is exactly what people worry about when they see names of dead people on voter rolls or people registered in two states. Even the study in the American Political Science Review admitted that double voting happens. If we don't have strict ID laws and clean rolls, we're inviting trouble.

David: The study you’re referencing, the 2020 Goel paper, actually found that for every double vote prevented by purging rolls, you’d likely disenfranchise 300 legitimate voters. The "fraud" people panic about is almost always a guy named John Smith being confused with a different John Smith. When we actually investigate, we find that the threat of fraud is tiny compared to the threat of making it harder for citizens to vote.

Mark: I'm not saying we should block voters, I'm saying we need a system that people can trust. When half the country doubts the results, that’s a crisis of legitimacy. Significant or not, if the vulnerabilities are there, the public perception will remain negative.

David: The crisis of legitimacy isn't caused by the system; it’s caused by the rhetoric. The empirical evidence is clear: U.S. elections are among the most scrutinized and secure in the world. Fraud is rare, localized, and ultimately insignificant to the final tally.

Mark: I hope you're right, David, but as tech advances, the risks only grow. We need to be vigilant.

David: Vigilance is good. Misinformation isn't. Let’s stick to the data.

Frequently asked questions

Did voter fraud change the outcome of the 2020 presidential election?

The weight of peer-reviewed evidence says no. A 2021 review found that widely circulated statistical claims of fraud were not credible, and audits and hand recounts confirmed machine counts. One 2022 study reported anomalies in Georgia and Pennsylvania that it said were consistent with fraud.

Do non-citizens vote in U.S. elections in large numbers?

A 2014 study using survey data argued non-citizen voting could affect Electoral College and congressional races. A 2015 rebuttal showed those findings were statistically indistinguishable from measurement error, such as respondents giving incorrect citizenship information.

Does mail-in voting lead to election fraud?

A study of three states from 1996 to 2018 found that universal vote-by-mail did not change either party's turnout share or vote share. The authors read this as evidence that mail voting does not create meaningful opportunities for fraud to alter outcomes.

Why do so many Americans believe election fraud is common?

A 2024 survey of 2,211 registered voters found that partisanship and conspiratorial thinking strongly predicted fraud beliefs, rather than actual evidence. Another 2024 study found that voters who see their state's security policies as effective believe there is less fraud and feel more confident in elections.

Related questions