By Kate Starbird, Mert Can Bayar, Ashlyn B. Aske, Danielle Lee Tomson, Emma S. Spiro, Nina Lutz, and Michael Grass
This article is the second in a series of “What to Expect in Election Rumoring” posts about the 2024 U.S. general election, drawn from insights gained through previous research, contextual knowledge, and consultation with community partners. Among other outputs, this series from the Center for an Informed Public at the University of Washington features articles introducing conceptual tools for understanding election rumoring (both generally and online) and for anticipating the types of rumors and tactics that we expect to see during different phases of the upcoming election. We’d like to acknowledge former CIP research scientist Mike Caulfield for previous work around election rumors that helped lead to the analysis in this piece.
Key Takeaways
We’ve identified five major ways that election rumors mislead the public: false evidence, misinterpreted evidence, exaggerated impact, overlooked remedies, and falsely attributed intent.
Election rumors rarely feature entirely false or fabricated evidence, though we may find more examples of this as generative AI becomes ubiquitous.
The complexity of U.S. election procedures, which vary by state and local jurisdiction, makes them susceptible to misinterpretation. These misunderstandings are often amplified and framed in misleading ways, contributing to the spread of false narratives.
A hallmark of truly democratic elections are their unknown outcomes, accompanied by attendant uncertainty and anxiety — transfers of power and surprise victories are part of the democratic process. Minorities can become majorities, and majorities can become minorities peacefully via the ballot box. Rumors are a natural outcome of humans coming together to try to make sense of and meaning out of uncertain events and outcomes through speculative stories of what is or could be going on. As such, rumors are a common feature of democratic elections as sincere activists, savvy political operatives, and concerned citizens share, frame, and amplify political and electoral information to shape public opinion.
For better or worse, rumors about candidates and policies are a core — and arguably a natural and predictable — part of the democratic process. False rumors and conspiracy theories about election administration and election integrity, however, present a potential threat to democracy. False claims about when and where to vote can disenfranchise (causing people to lose their opportunity to vote), false reports of violence at the polls can suppress votes (causing people to stay home to stay safe), repeated false allegations about election fraud can diminish trust in election results and even motivate disruptions to the peaceful transfer of power, eroding the foundations of democracy.
We acknowledge that rumors can turn out to be true, false, or somewhere in between. Even if false, rumors can reflect a degree of concern among a population and should be thoughtfully addressed by election officials and journalists to ensure the integrity of elections. Our concern is when and how rumors about election administration are not a simple check on electoral processes but part of a larger strategy to deteriorate democracy itself. Given rumors are a natural human feature in a democracy, discerning the functions, roles, and nature between different types of election administration rumors is important when triaging them.
At the University of Washington’s Center for an Informed Public, our team has been studying election rumors — specifically, rumors related to election administration and results — since 2020. In addition to “rapid” research analyzing rumors in real-time, which we plan to continue for the 2024 U.S. general election, we have also conducted extensive post-hoc research to more thoroughly investigate election rumoring in 2020 and 2022 (e.g., publishing peer-reviewed work on repeat spreaders and mobilization around participatory disinformation) and map out similar rumors from past elections. Four years of intense research have provided perspective into some of the deeper patterns of election rumors — including persistent narratives (such as non-citizen voting) and repeated tropes (such as mysterious white vans at ballot boxes or polling places).
The 5 Moves of Election Rumors
Across the hundreds of rumors we have studied, we’ve found that there are five primary ways in which rumors about election processes and results mislead:
False Evidence
Election rumors may mislead the public by introducing false — i.e., either wholly fabricated or substantively doctored — evidence. Colloquially, this is often what people understand “misinformation” to be when invoked: a false or misleading representation about the “facts” of an event, individual, organization, or process. In the election context, false evidence can include false witness statements from voters, election observers, or election officials. It can also take the form of false messages about election processes misattributed to election officials.
Historically, this category has included intentional deceptions like forgeries and false news articles. But in our modern media environment, and especially with the introduction of generative AI, this can include synthetic or deceptively edited audio and/or visual material. Though some aspects of its production may be inadvertent, false evidence is most often intentionally created, edited, and/or introduced to generate a false perception of events in the world.
For example, during the 2024 presidential primaries, a political operative used generative AI to impersonate the voice of President Biden in a robocall that provided false information about the election and encouraged people not to vote in the New Hampshire primary. The individual who sent the message — which featured wholly fabricated evidence — was charged with voter suppression in May 2024.
Significantly, though false evidence is often what we think about first when someone invokes the term “misinformation,” it is not the most common way false rumors about election administration mislead. Having studied hundreds of rumors about election processes, procedures, and results, we could only come up with a few examples that rested wholly on false evidence. However, the “false evidence” rumor may become more common with the increased ubiquity of and advances in generative AI.
Fabricated and manipulated evidence can also work in concert with other dimensions — such as misinterpreted/misrepresented evidence — to mislead. For example, in a highly publicized case of two Georgia election workers who were falsely accused of fraud, selective edits to surveillance video reinforced the false characterization of the election workers’ activity as outside normal processes and therefore “fraudulent.” The workers have been cleared of wrongdoing and won a major court case for defamation.
Misinterpreted Evidence
A much more common type of misleading rumor emerges from misinterpreted or mischaracterized evidence. In these cases, the evidence itself is true, but its meaning is distorted, often unintentionally, creating a false perception.
A clear example of this is the “SharpieGate” case from 2020, where rumors claimed that Sharpie pens were bleeding through ballots and systematically disenfranchising election-day voters. In truth, the pens were indeed bleeding through ballots (and many social media users posted “evidence” of exactly this), but the ballots were designed so that the bleed-through would not affect whether (or how) they were counted.
Another example, which occurs across locations and election cycles, involves misinterpretations of “vote dumps,” when overall vote counts are updated to reflect a large batch of votes that may often significantly shift the relative proportions. Vote dumps are often (falsely) alleged to reflect problems with vote counting or the introduction of fake ballots into the counting process. But they almost always have a mundane explanation — e.g., a large area that leans heavily towards one party releasing their results all at once.
In 2020, rumors about Sharpie pens and vote dumps primarily spread among conservatives and Trump supporters. But on the other side of the political spectrum, in August 2020, photos of thousands of mail collection boxes at a facility led to false interpretations and then accusations that the U.S. Postal Service — run by an appointee of President Trump — had nefariously taken the boxes to “a dump” to sabotage mail-in voting. In reality, the boxes were at a facility where mailboxes are regularly refurbished — their presence there had nothing to do with the election.
In these and other cases, the problem is not in what the evidence is (Sharpies do sometimes bleed through ballots, vote counts do get updated in ways that significantly change the proportions, and mail collection boxes often appear in seemingly strange places), but what that evidence means. In our work, we see this type of rumor emerge both from sincere misinterpretation and strategic mischaracterization.
In 2020, our research team recorded dozens of rumors stemming from people sincerely misinterpreting their own and others’ experiences around voting and vote counting — from the holes in ballot envelopes at many locations (perceived as enabling fraud, but actually there for increased accessibility for blind and low-vision voters) to a red wagon with a suitcase outside a polling location in Detroit (perceived as a “suitcase of ballots” enabling fraud, but actually the camera equipment of a photographer from a local media outlet).
A significant contributor to the misinterpretation/mischaracterization rumors is the fact that modern American election procedures have multiple layers of complexity that exist to ensure accessibility, scalability, accuracy, and security. Unfortunately, few people fully understand how each of those layers works, which leaves us vulnerable to misinterpreting information about different elements of the voting and vote-counting process. Additionally, in the U.S., voting administration is distributed across the state and local levels, with varying laws and procedures, meaning that the rules in one location likely differ from those in another. This creates opportunities for misinterpretation and intentional mischaracterization.
Real Evidence of Election Issues, Strategically Amplified and Falsely Framed
The first two categories of election administration rumors rest on false or misinterpreted evidence. The following three categories feature cases where something has indeed gone wrong in the election process — e.g. a problem with voting machines, human error, or even malfeasance — but where that evidence is selectively amplified and/or framed in misleading ways to create a false impression.
Mistakes happen. And in a general election where millions of people are voting across thousands of different locations where rules and procedures vary by state and county, we are likely to see errors. Mail-in ballots get lost while in transit or fail to arrive on time. Administrators send registration materials to the wrong people. Ballots arrive for people who are no longer living at that address. Machines have glitches on election day. Reports of these issues will often turn out to be superficially true. However, they also become fodder for false rumors and misleading narratives that twist the meaning of this “evidence” to fit the frames of untrustworthy or rigged elections.
Overlooked or Obscured Remedies
Modern elections in the U.S. have multiple layers of verification, security, and resilience. There are backup plans for cases where voters don’t receive mail-in ballots and when machines go down. Tracking systems allow voters to double-check to ensure their ballots are received and counted. There are numerous mechanisms to prevent the counting of two ballots from the same voter. People who arrive for election day voting and find an error with their registration can file a provisional ballot and work to resolve the issue. The vast majority of voting issues — from system and human errors to (rare) attempts to commit voter fraud — have remedies. However, those remedies are often overlooked or intentionally obscured in election related rumors that function to sow distrust in the process and/or the results.
For example, on election day in Maricopa County, Arizona, in 2022, issues with voting machines caused a significant disruption for several hours, affecting thousands of voters. Those issues had remedies — voters could place their paper ballot in a box (“Box 3”) to be counted later or could go to a different location to vote — but rumors about the issues ignored or actively cast suspicion upon the potential remedies, contributing to distrust in the process.
Exaggerated Impact
Another way that election administration rumors mislead is by exaggerating the impact of real issues. One dimension of this is making isolated issues seem widespread. For example, in 2020, we saw numerous reports of mail-in ballots that were not safely delivered during the postal process — and were instead lost, damaged, discarded, or stolen. Unfortunately, our postal system is imperfect, and mail (including registration materials and ballots) does not always reach its intended destination. These issues are not common, but with millions of ballots in transit, it is common to find a few cases of them not reaching their intended destinations. Throughout the 2020 election cycle, motivated amplification of these and other isolated issues — by political actors working to push the “rigged election” frame, influencers seeking engagement, and journalists looking for readers — created a perception that issues were widespread, sowing doubt in the integrity of the election and reducing trust in the results. We can expect similar dynamics in 2024, though perhaps with less focus on mail-in ballots and more on other parts of the process.
A second dimension of exaggerated impact involves falsely implying that a small but real issue with election administration will change the outcome of an election. Though election administration errors or intentional voter fraud have in the past — and could in the future — impact the results of a close election (typically on a local level, if it happens), these impacts are rare and unlikely to affect state or presidential elections. However, we often see cases where reports of real issues with election administration are misleadingly framed to imply an impact on election outcomes.
In a recent example from a primary election this August, an administration error related to a new district map left the candidates in one race off the ballot in Summit, Wisconsin, causing 188 people to lose their ability to vote for that office in that race. The outcome of the race was not impacted by the error; the victor won by a margin larger than the number of impacted voters. That said, the error and the disenfranchisement of voters fed existing rhetoric questioning the integrity of elections and the accuracy of the results, reinvigorating skepticism of procedures, like the use of voting machines that had nothing to do with the error by the county clerk.
Falsely Attributed Intent
The final category is falsely attributed intent. These are cases when there are real issues with voting and/or election administration — for example, issues with the delivery of mail-in ballots in 2020 or malfunctioning voting machines on election day in Maricopa County, Arizona — but where those issues are falsely alleged to be intentional, rather than accidental. Voter fraud exists, but it’s rare. Similarly, election officials do make mistakes, but those are rarely intentional. (When they are intentional, those have been primarily due to misguided attempts to find or demonstrate vulnerabilities in the system.)
Many election rumors incorporate falsely attributed intent in combination with other misleading moves (e.g. exaggerating impact and overlooking remedies). These include rumors about mistakes by voters — e.g. cases where an individual mails ballots in for friends or family members or when someone moves without canceling a previous registration and ends up on the voting rolls in two different places. However, falsely attributed intent rumors most often feature errors with election administration — from poorly designed ballots to malfunctioning machines to misapplication of election laws. For example, in 2022 the Colorado Secretary of State’s office accidentally sent out registration forms to thousands of non-citizens, feeding existing rumors that Democrats facilitate fraud by allowing non-citizens to vote. Similarly, our team has repeatedly seen false or unfounded claims — across the political spectrum — that issues with mail-in ballots result from intentional sabotage by USPS officials (left-coded rumors) or USPS workers (right-coded rumors).
Conclusion
This simple “five moves” framework emerged through years of research on election rumoring in the U.S. context and as part of an effort to develop conceptual tools to support our own “rapid research” team in quickly categorizing different rumors. We hope that this framework can be helpful to others in determining how a specific rumor misleads. When we see a rumor about voting, we often get distracted by questions about the underlying evidence: Did this happen? Is this real? These questions are necessary to answer in order to determine the veracity of a rumor, but don’t tell the whole story about the meaning of the rumor or what it is doing.
As we have tried to explain here, rumors often mislead not just about the facts of the matter, but about what those facts mean. Such false or misinterpreted facts may have real consequences, eroding people’s faith in the system and damaging the functionality of a democratic society and its institutions. This framework could support researchers, fact-checkers, and election officials in quickly diagnosing a rumor — i.e., identifying the different ways that an election rumor misleads to determine the missing context. We believe that this approach can also help journalists who cover election related news recognize where their reporting might feed false rumors and take action — such as highlighting important context about the remedies and actual impact of issues — to reduce the exploitation of their work for the spread of misleading rumors about election integrity.
Kate Starbird is a University of Washington Center for an Informed Public co-founder and UW Human Centered Design & Engineering professor.
Mert Can Bayar is a CIP postdoctoral scholar based in the UW Department of Human Centered Design & Engineering.
Ashlyn B. Aske is a CIP graduate research assistant and a Master of Jurisprudence student at the UW School of Law.
Danielle Lee Tomson is the CIP’s research manager.
Emma S. Spiro is a CIP co-founder and associate professor in the UW Information School.
Nina Lutz is a CIP graduate research assistant and HCDE PhD student.
Michael Grass is the CIP’s assistant director for communications.