Why You Should Care
Why should you care about all this?
Because vote reporting in the United States is vulnerable to unintentional error and deliberate fraud. You can help America Counts double-check the vote reporting using our Actual Vote system, and gain benefits!
Here are seven examples of real situations in which there may have been error or fraud in vote reporting.
America Counts’s Actual Vote system would have prevented each of these!
For more detail about how this page is in line with America Counts’s non-partisan mission, see this quick note:
America Counts is inviting everyone to use our Actual Vote app to help us audit vote reporting in the next election because:
- Vote reporting in the US is vulnerable to error and fraud, and is not sufficiently audited
- In a given election, when vote reporting was actually done correctly, an Actual Vote audit constitutes independent confirmation of this fact. This helps to defend against false accusations of error or fraud
- In a given election, when there actually was error or fraud in vote reporting that didn’t undergo an audit, or wasn’t caught or acted upon by the official audits that were done, an Actual Vote audit can uncover evidence that can help catch and rectify this error or fraud. This helps improve election accuracy
- Regardless of the outcome, an Actual Vote audit improves election transparency since it provides independent evidence that helps us form an opinion about the official vote reporting. Greater election transparency improves social well-being, which is great for everyone
- Therefore, it’s worth people’s time to use Actual Vote to help America Counts audit vote reporting in the next election
Points (2)-(5) will seem reasonable to many readers who are already convinced that point (1) is true.
But how do we know (1) is true? That is, how do we know that vote reporting in the US is actually vulnerable to error and fraud to enough of a degree that we need to do something about it? What evidence is there?
It turns out that there’s plenty of evidence, in the form of real-world stories involving accusations of error or fraud in vote reporting that the Actual Vote system could have helped to resolve. In each such story, if Actual Vote had been used to generate evidence about the accuracy of the vote reporting, it could have been used to help resolve the accusations of error or fraud.
Of course, in any such story, at least one political party will stand to gain unearned political advantage, depending on how the accusations of error or fraud end up being resolved. So you might think that America Counts, a staunchly non-partisan organization, would avoid citing such stories to prevent us appearing to portray a particular political party as less honest than another political party.
But America Counts is proudly featuring seven such stories on this page. We believe that giving this collection of stories is thoroughly consistent with our non-partisan mission. There are four big reasons why this is.
First, America Counts’s invitation to everyone to use Actual Vote in the next election is inherently non-partisan. But if they’re going to consider putting out the effort to use the app, we owe it to them to present the case that vote reporting in the US is vulnerable to error and fraud, and is not sufficiently audited (this is point (1) above). But the only possible way to make this case is to present real evidence, in the form of stories involving accusations of error or fraud in vote reporting, in which the audits that were done (if any) were not sufficient, and that Actual Vote could have helped to resolve. So, while individual stories may involve members of one party or another engaging in possibly questionable behavior, the entire collection of stories establishes point (1) above, and the rest of the points follow. The result is that more people will be motivated to use Actual Vote and contribute to our non-partisan vote reporting audits, and reap a great return in election transparency and social well-being for their efforts.
Second, due to the complexity around vote reporting and the difficulty of evaluating sources, it would be very hard for non-experts to independently arrive at the conclusion that vote reporting is vulnerable and insufficiently audited by themselves. Since America Counts are experts in vote reporting, part of our non-partisan mission is to synthesize known information about vote reporing, like we’re doing here, to give people helpful evidence for forming an opinion about the state of vote reporting in the US, and whether they’d like to help independently audit it. If America Counts were to decide to not feature a stories page like this on our website because individual stories may feature questionable behavior from one political party or another, it would detract from this non-partisan mission.
Third, we carefully selected the stories based on several criteria: they should clearly demonstrate the consequences of vulnerabilities in vote reporting and lack of sufficient audits, they should be easy to explain, they have easily-accessible supporting evidence, and we should try to end up with roughly the same number of stories that involve questionable behavior from one or the other political party. Of course, the only material we have to draw on that fits all these criteria are the cases that have been documented over roughly the past century of US politics. This means means that it’s unlikely that we could find a collection of stories that is theoretically perfectly balanced between how it portrays each party. Instead, we tried to achieve the best balance we could within these constraints, the way you’d expect from a non-partisan organization operating in a complex space such as this.
Finally, we’ve annotated each story to highlight the aspects that are most relevant for convincing people of the need to use Actual Vote. In these annotations, we are careful to present only known facts from each story, with citations. We also studiously withhold judgement about any relevant accusations about vote reporting that haven’t been definitively settled in a court of law. Such care further demonstrates our commitment to avoid even the appearance of partisanship on our part as a result of featuring a page like this one.
Together, these stories show that we have a choice in the next election: allow the vote reporting to remain vulnerable and under-audited, and potentially face situations like in the stories, or independently audit the vote reporting with Actual Vote and reap the corresponding benefits for election transparency and social well-being.
The Gaston County Courthouse
While the recount chaos in Florida was dominating the national spotlight, other stories like this one fell under the radar. Here, faulty machine cartridges caused a third of the precincts to incorrectly report zero votes. A different error caused the tabulated totals to cap out after 32,000 votes. These errors were detected and fixed by retrieving the correct vote data from the internal hard drive of the tabulating machines.
If an Actual Vote analysis had been done in this situation, the poll tapes would have printed with the correct vote totals, and it would have been clear that the there was an error reporting the vote totals.
The Baldwin County Courthouse
After the final numbers were tallied, the winner of a razor-thin Governor’s race was declared as Don Siegelman. Then, it was alleged that a computer glitch during the reporting had overcounted Siegelman’s total by about 7,000 votes in a certain precinct. When the allegedly overcounted votes were removed, his opponent, Bob Riley, became the new winner. But it was revealed that after closing, a county chair reentered the courthouse to “continue counting”. This caused the entire story about the 7,000 votes to be called into question.
If Actual Vote poll tape recordings had been available from the precinct in question, they could be compared against the officially reported results to definitively resolve this vote reporting controversy.
The Warren County Courthouse
Exit polls predicted John Kerry winning Ohio by a 4.2% margin, but then the officially reported results had Bush winning by 2.5%. Such a disparity is incredibly unlikely to be due to chance, which alerts us to be on the lookout for potential evidence of error or fraud.
Indeed, there were a great many questions raised in Ohio about this election.
In Warren County specifically, it was claimed that the election building had to be locked down due to an alleged terror threat (which meant that independent observers no longer could watch the reporting process) and that 80,000 votes were switched from Kerry to Bush during the reporting process through a convoluted scheme. If this actually happened, it would have meant that Kerry was actually the true winner of Ohio and the electoral college.
This means that, while we have no choice but to accept the official outcome now, the known body of evidence in favor of this outcome doesn’t fully meet the kind of high standard that democracy demands. This is because for claim above, upon which the election result depends, we simply can’t definitively mark it true or false with just the information we have.
But if Actual Vote had been run in Warren County for this election, it would have provided independent evidence of what the correct vote totals should have been, and could have been used to help clarify what really happened. But we have no such evidence, meaning that we will never know for sure.
Smith’s poll tape photo
Computer programmer Bennie Smith was wondering why Shelby County, which is a Democrat stronghold and predominantly African American, had had Republican sweeps of county-wide races in the past. It was strange enough that during the 2015 election he decided to snap a photo of the poll tape at the Unity Christian Church polling place, which showed that 548 people had cast ballots. Meanwhile, the officially reported results showed only 330 votes—40% short!
County technicians used Smith’s photo as evidence in an investigation that showed that these votes and more never made it into the election night count. All such votes were from precincts with large concentrations of black voters. Unfortunately, at the end of the day, local officials failed to get to the bottom of the problem.
But then in the 2018 election, the candidates who won were far more representative of the population of Shelby County. The election transparency boost from Smith’s poll tape photo seems to have initiated a process that resulted in an outcome in which there was far less reason to suspect problems with vote reporting.
The Prince William County Courthouse
This is a case of alleged corruption by the then-head of elections that resulted in the officially reported vote total for Joe Biden to be short by 4,000 votes. Charges were filed against the official in question but they were later dropped, making the extent to which this case was due to error or fraud unclear. Either way, the fact that it’s possible for a 4,000 vote discrepancy to occur means that it’s critical that we independently audit vote reporting.
An Actual Vote analysis here would have helped to clarify the situation, since if the story were true, it would have clearly shown that the vote totals on the poll tapes were higher than the official results by 4,000 votes.
The Monmouth County Courthouse
Human error caused the votes to be incorrectly reported to such an extent as to change the outcome of a tight school board race. The error came from an election worker accidentally running a certain set of votes through the counting process twice instead of once. A second source of human error, in which the software on the machine in question had been reinstalled incorrectly, prevented the error from being detected at the time. The error was later caught and fixed.
Actual Vote could have caught this error as soon as the official votes were reported, since the relevant official vote totals would have been twice the corresponding vote totals on the poll tapes.
If you ask me for a slice of pizza, and I take a bite out of one and hand it to you since “0.9 rounds up to 1”, that’d be pretty weird.
The Fraction Magic attack involves a widely-used election reporting software system called GEMS. It has been shown that the GEMS system could theoretically be configured to display votes as whole numbers, but to internally store them as fractions that round off to the corresponding whole numbers. This is bad, since each vote is supposed to count as 1, not as some fraction.
Imagine, for example, a bad actor who wants Sam Silver to win. When the official results are ready to be released, they show 95 votes for Silver and 100 votes for Gina Gold. They don’t want Gold to win, so before the official results are released, they uses the Fraction Magic attack to change the way each of Gold’s votes is internally stored from 1.0 to 0.9, while still displaying the rounded values of 1.0 on the screen. This makes the foul play impossible to detect by a simple visual inspection. Now Gold’s officially reported vote total will be 100*0.9 = 90, and Silver will be incorrectly (and “fraction magically”) reported as winning the race 95-90.
In such a situation, an Actual Vote analysis would show that the poll tapes clearly have Gold receiving 100 votes and winning the race, and can be used as evidence to help resolve the situation.
Clearly, any time software is involved in the election process, the result should be independently audited.
In each case above, there was no independent double-check of the vote reporting, meaning that they lack this critical component of election transparency. Without it, key questions can never be answered and people are forced to move on, despite a nagging sense that the official outcome isn’t supported by a high-enough standard of evidence.
In each case, if an independent double-check were done (such as an Actual Vote analysis), voters would have had easy-to-interpret evidence about the accuracy of the officially reported vote total, and in turn, about the likelihood that error or fraud may have taken place between tabulation and releasing the official results.
This evidence makes the election more transparent—that is, it increases the shared basis we all have to decide the extent to which we believe the results are accurate. This is critical for something as high-stakes as an election.
Actual Vote analyses with poll tape videos would have added much-needed clarity to each of the vote reporting incidents described above.