Implying Intentionality: Understanding Unsubstantiated Allegations Around Election Administration Mistakes
When errors in election procedures and materials surface, either by election officials via official statements or by voters who discover the errors, online discussion often includes speculative rumors of intentionality.
In this Election Integrity Partnership (EIP) analysis, we look at two case studies, in Colorado and Arizona, where unsubstantiated allegations of intentionality around mistakes or errors in election administration have spread rapidly online.
Tactics that facilitate this spread are often designed to invite engagement — and are difficult to moderate.
Election officials should strive to be rapid and transparent in reporting mistakes, and work to provide sufficient context to mitigate the spread of false rumors and disinformation.
Photo composite (from left): The Denver, Colorado skyline by bikesharedude via Flickr / Public Domain; The Pinal County Courthouse in Florence, Arizona by Larry & Teddy Page via Flickr / CC BY 2.0
This Election Integrity Partnership analysis was written by Taylor Agajanian, Joseph S. Schafer, Kayla Duskin, Rachel E. Moran, with contributions from Adiza Awwal, Damian Hodel, Stephen Prochaska, Kate Starbird, and Rebecca Gulbrandsen (University of Washington Center for an Informed Public).
Introduction
Mistakes happen in any large system, and demand attention. Identifying mistakes, bringing them to light, and acting on them thoughtfully helps refine a system, and makes future mistakes less likely. Voting and elections systems are no exception to this rule. However, mistakes can also be weaponized or assigned aspects of intentionality — often without evidence. In elections, honest human errors can be opportunistically exploited to imply intentionality and to support unfounded narratives of intentional, widespread fraud, undermining the legitimacy of electoral outcomes. However, as research shows, election fraud is exceedingly rare and such mistakes are unlikely to impact election outcomes.
In the following post, we examine two recent cases of election administration errors (in Colorado and Arizona) and explain how they were rhetorically framed as intentional attempts to influence the outcome of local elections. Our analysis highlights some of the rhetorical tactics employed on social media that make unsubstantiated allegations of intentionality effective in spreading election-delegitimizing messages.
Case Study One: Colorado Secretary of State’s Office Mistakenly Sends 30,000 Voter Registration Postcards to Non-citizens
In Colorado, the Secretary of State’s office mistakenly sent out 30,000 “eligible but unregistered” (EBU) voter registration postcards to non-citizens who reside in the state. The error occurred after the department office obtained a list of 102,000 names from the Electronic Registration Information Center, which was then compared to information within Colorado’s driver’s license records. Importantly, this error does not allow non-citizens to register to vote, as the online system Colorado uses would prevent applications from going through. After becoming aware of the mistake, the Secretary of State’s office shared their understanding of the situation and their intent to correct the situation. As discussed in a recent EIP tweet thread, this incident spurred skeptical discourse across social media as to whether the incident was truly a mistake, though the EBU postcard sent out clearly states that one must be a citizen in order to register to vote.
Misleading and misinformed narratives about non-citizens voting in U.S. elections have gained traction throughout 2022, which has been previously explored in an Election Integrity Partnership blog post. This narrative fits into a broader range of conspiracy theories regarding illegal voting in general, positioned as an attempt to sway election results toward one party or the other. The Colorado incident also contributes to a common pattern where sincere human error is falsely or mis-attributed to intentionality and often weaponized in service to claims of election fraud.
Incident Spread
The Colorado postcard incident gained initial attention on Twitter on October 7 when NPR Politics retweeted a Colorado Public Radio reporter’s article about the error. Engagement remained low until October 9, when a “micro” influencer with around 20,000 followers amplified this incident and it was subsequently engaged with in right-leaning online communities.
Image 1: A screenshot of an unverified influencer’s post about the Colorado postcard incident, which was the first post about this incident to gain major traction.
Mainstream news outlets reported on this story a short time later, including The Associated Press.
Figure 1: This graph shows a cumulative plot of tweets mentioning terms related to the Colorado incident. The x-axis represents the time of each included post in the UTC time zone, and the y-axis references the total number of collected posts about this topic up to that point. Points are colored and shaped based on what kind of tweet is represented, and are sized by the amount of engagement (retweets, replies, and quotes) that each post received. The graph is then annotated with arrows pointing to influential posts which helped to drive the spread of this incident, and the original post from the CPR Affiliate reporter about this incident.
Many Twitter influencers framed the error as intentional rather than an accident, often through the use of sarcasm. For example, Dinesh D’Souza, a right-leaning political commentator and author, posted about this incident claiming that election process accidents always favor one side, implying that these are intentional mistakes, and also linking this to his recent film which alleges there was an extensive effort by Democrats to steal the 2020 election — and which has been largely debunked.
Image 2: A screenshot of a Twitter post about the Colorado postcard incident from author and political commentator Dinesh D’Souza.
U.S. Representative Jim Jordan, a Republican from Ohio, was more explicit, expressing skepticism of the official statement from Colorado officials and asking his followers, “Anyone actually believe them?” This invitation to his audience to participate in the framing of the act as intentional was taken up in numerous quote tweets responding to Jim Jordan’s post. Examples of quote tweets range from answering the “Anyone actually believe them?” question to accusing treason and calling for arrests.
Image 3: A screenshot of a post about the Colorado postcard incident made by Representative Jim Jordan (R-Ohio).
Such implied allegations also surfaced on alternative social media platforms like Truth Social and Telegram. Many posts about this incident’s accidental nature employed sarcasm and skepticism as a rhetorical technique to sow or intensify doubt among readers. For example, on Truth Social, former president Donald Trump posted a link to an article about this incident from the Daily Wire, which puts the word “incorrectly” in intentional quotations, implying that this was not an accident. On Telegram, a post by former National Security Advisor Michael Flynn (who uses the name “General Flynn” to post on that platform), linking a Gateway Pundit article about the incident, has amassed approximately 200,000 views and states the Colorado error “was not a mistake,” instructing readers to “keep their eyes and ears open.”
Image 4: A screenshot of a Truth Social post made by former President Donald Trump about the Colorado postcard incident.
Image 5: A screenshot of a Telegram message by former National Security Advisor Michael Flynn linking to a Gateway Pundit article about the Colorado postcard incident.
Case Study Two: Primary Election Ballot Shortage in Pinal County, Arizona
Pinal County, Arizona’s third-largest county, faced controversy in the August primaries as election officials failed to print enough ballots to meet the demand for in-person voting. Pinal County Attorney Kent Volkmer explained that the mistake was the result of “human error,” that the county simply did not order enough ballots to meet the demand. Given that in-person turnout is never known ahead of time, ordering sufficient paper ballots is always an estimation, one that Pinal County officials miscalculated. Pinal County estimated approximately 750 voters could have been affected, but admitted that there was no way to reliably assess the precise impact. Jeffrey McClure, chair of the Pinal County Board of Supervisors, told an August 3 news conference about the incident that he had “not seen evidence of a nefarious act.”
Kari Lake, one of the Republican primary candidates and the candidate who has advanced to the general election, expressed concern over the mistake, amplifying a tweet that lay blame at the hands of her opposition and current Secretary of State Katie Hobbs (see Figure 1). In her primary victory speech, Lake argued that she had won in spite of problems at the polls, saying “we outvoted the fraud.”
Image 6: A tweet on August 4 from Arizona Republican gubernatorial candidate Kari Lake, blaming her opponent Katie Hobbs for the Pinal County ballots incident.
Incident Spread
Figure 2: This graph shows a cumulative plot of tweets mentioning terms related to the Arizona incident in its time of original spread, in August. The x-axis represents the time of each included post in the UTC time zone, and the y-axis references the total number of collected posts about this topic up to that point. Points are colored and shaped based on what kind of tweet is represented, and are sized by the amount of engagement (retweets, replies, and quotes) that each post received. The graph is then annotated with arrows pointing to influential posts which helped to drive the spread of this incident.
As the incident was happening in August, conversation drawing attention to the error spread on Twitter. The first major tweets talking about this incident and kicking off this spread, posted by Kelli Ward (chair of the Arizona Republican Party), describe the incident but do not link this to a frame of intentionality, instead just detailing the error in election administration. Later, Ronna McDaniel, the chair of the national GOP, and Charlie Kirk, the founder of Turning Point USA, bring the specificity of this incident to focus on David Frisk, the county elections officer, saying that this incident was his fault and that he should resign his position. Later, the official account for the Arizona Republican Party instead linked this blame to a mistake by Katie Hobbs, the statewide Secretary of State, who is currently running for the governorship as a Democrat. Kari Lake also criticized Katie Hobbs’ office for the mistake, though at this time she did not add a framing of intentionality. In the August spread of this incident, the first direct linking of this incident to blaming Hobbs was by Tyler Bowyer, the COO of Turning Point USA, who wrote that “Katie Hobbs’s office directly led to the disaster in Pinal County,” and subsequent tweets by Kari Lake in this immediate time frame echoed this framing.
Image 7: A tweet from Tyler Bowyer placing blame for the Pinal County ballots incident with Katie Hobbs’ office.
This incident gained further traction in October with much more salient use of allegations of intentionality. On October 10, 2022, Kari Lake raised this incident again, using intentional quotation marks to sarcastically comment she hoped her opponent “learned from her ‘mistakes,’” implying that this incident did not result from a mistake.
Understanding Framing Efficacy Via Replies to Influencer Tweets
Image 8: A tweet from Kari Lake which implies that the Pinal County incident was due to intentional actions by her opponent, Katie Hobbs.
Methods of Analysis
To better understand how the framing of added intentionality was received by audiences on social media, EIP researchers qualitatively examined audience replies to the key tweets associated with the stories. In the case of the Colorado postcard error, we randomly sampled 90 replies to the tweet by @Jim_Jordan (Image 3) and in the case of the Arizona ballot shortage we examined all 89 replies to the tweet by @KariLake (Image 8) as of the time of analysis.
In both of these tweets, the authors did not explicitly make accusations of cheating or fraud, instead insinuating intentionality through the use of quotation marks around the words accident and mistake respectively — and in the case of the @Jim_Jordan tweet, through the use of a rhetorical question. We coded (categorized through qualitatvie analysis) each reply to the tweet as either unrelated to the voting process error, claiming incompetence or sowing doubt in the legitimacy/competence of election officals, claiming fraud, disagreeing or correcting the original tweet, or connecting to other narratives of election illegitimacy.
Findings
The full results of coding are in Table 1, below. Most of the responses from audience members are unrelated to the voting process error highlighted in the original tweet, instead including general support or criticism of the tweet’s author. However, of those that do respond to the pertinent content of the original tweet, a portion of the replies continue with the insinuation of intentionality, explicitly claim fraud or connect the incident to other (unsubstantiated) past claims of fraud.
Table 1: Results of qualitative coding the Twitter replies to two Tweets insinuating that mistakes within the voting process were intentional.
In both of the above case studies, adding a framing of intentionality positions the mistake as part of a broader story of electoral fraud. Further, it moves the conversation to become explicitly partisan, with intentionality either directly assigned to political opponents or implied and subsequently supplied by audience participation. This is especially pertinent when the allegations of intent revolve around a political opponent, as was the case in Arizona. Though a different election official was responsible for the error, Republican gubernatorial candidate Kari Lake falsely directed blame toward her Democratic opponent Katie Hobbs, accusing her of deliberate sabotage.
Additionally, while these allegations of intent create a misleading framing of these incidents to spread, they do so in a way which is not easily addressable by platform moderation.
The tweet reporting the facts of the case by Colorado reporter Bente Birkeland, and the intention-alleging Truth Social post by Donald Trump, are extremely similar, the only real difference being the addition of two quotation marks around “incorrectly.” Moreover, the quotation marks here could be construed as indicating a direct quote, rather than as an indicator of sarcasm, further obscuring whether such content is misleading and thus violative of any platform community guidelines, even though the intended audiences — or at least some percentage of them — understand what is being implied. As a result, this rhetorical maneuver serves as both a moderation-avoidance technique and as a way to solicit engagement with a false frame, as strongly evidenced by the Jim Jordan post.
The spotlighting of real errors within the voting process is important in order to address vulnerabilities and rectify mistakes in future elections. However, the weaponization of these incidents of error as intentional and/or as evidence of fraudulent or untrustworthy election results is misleading. In the case studies highlighted, and additional incidents of election-related mistakes such as the recent case of misprinted ballots in Schuyler County, Illinois, conversation around the mistakes becomes a mix of factual coverage, legitimate criticism, and unsubstantiated accusations of voter fraud. Disentangling these conversations makes it difficult for citizens to understand the realities of the mistake, what is actually impacted, and how election officials are working to rectify it for future elections. This confusion similarly risks reducing the effectiveness of good faith critique and its impact in strengthening election integrity.
Recommendations
Reporters and News Organizations
When reporting on elections and election processes, it is important for legitimate actors in the news media sector to be clear and explicit about intentionality and what can be explained by sincere human error. Adding correct context in both headlines and the body of news publications is essential to reducing the spread of misleading claims of intentionality which often stoke conspiracy theories of election fraud. Also, news outlets should make sure this context is clearly visible on the article preview card that users share to their social media feeds to minimize the potential for articles to be taken out of context. While reporters cannot stop readers from making insinuations of intentionality by using quotation marks around their coverage, reporting can minimize the weight of these accusations by specifying the impact of the mistake and how election officials responded.
Platforms
Moderating this sort of content is difficult, and in most cases, unconfirmed claims of intentionality do not violate platforms’ terms of service. In these cases, crowdsourced context initiatives like Twitter’s Birdwatch, or amplifying authoritative fact checks alongside emergent hashtags and conversation, may be useful to provide clarity about the facts of a given incident where error is being misattributed to intentionality. Additionally, platform Trust & Safety teams should be cognizant of rhetorical techniques meant to sow doubt and confusion amongst users. This is especially important in cases where false claims of intentionality are combined with moderation-evading rhetorical techniques to foment targeted harassment aimed at another individual or user.
Election Officials
For elections officials, rapid transparency and honesty from the beginning is helpful in making clear what caused any election process errors. This is critical to avoid the appearance of intentional misconduct, as a delay in reporting can lead to conspiracy theorizing. This also helps reduce the co-optation of misattributions of intentionality in service to negative political campaigning. Further, clarifying the scope of impact of errors — specifying how many ballots were impacted (if applicable), and contextualizing where in the voting process the error occurred to ensure that mistakes are not used to undermine confidence in election results.
General Public
Finally, for the general public, applying an adapted version of Hanlon’s Razor is a useful heuristic for evaluating such errors as an outside observer. Hanlon’s Razor states “never attribute to malice that which is adequately explained by stupidity,” though is better applied here as “never attribute to conspiracy that which is adequately explained by sincere error.” When evaluating how to talk about and share stories related to election process errors, it is important not to add framings that are unsupported by the available evidence. Unless sufficient evidence exists to credibly claim intentionality, it is best to stick to the facts and avoid introducing additional, unsubstantiated claims within your discussion.