Voter Fraud?
I know, Christmas is right around the corner. You're expecting good cheer. Can I really ho-ho-ho my way through a newsletter on alleged voter fraud in the 2020 presidential election?
Well, I can least offer you a couple of presents.
The first is restraint. I won't scream, and I won't spend much time rehashing the usual arguments. There's been more than enough partisan screaming, and we're depressingly familiar with debates over the legitimacy of the 2020 election.
My second "present" is a demonstration of how simple statistics, plus common sense, can be used to call out disinformation.
My focus will be on a new poll that appears to show substantial mail-in voter fraud in 2020.
Some right-wing media outlets and leaders, including Donald Trump, are heralding the results of this poll as evidence of a stolen election. With a few exceptions, the liberal media mostly ignored the poll. I think the data merits a close look. 43% of voters in the 2020 election mailed in ballots, and a higher percentage is expected next year. Any data questioning the integrity of the process is worth considering.
My biases
Although I identify as liberal, I will try to be fair. I've discussed the poll with liberals as well as conservatives. I've interviewed the poll's creator. And although I draw a firm conclusion, I'll present all the data you need to decide for yourself. Who knows, you may end up disagreeing with me. (You'll be wrong though....ho-ho-ho.)
A semantic note
To simplify a bit, I'll be contrasting "election affirmers" vs. "election deniers". No judgment is intended by either term.
Election affirmers believe that the 2020 presidential election was legitimate. Any voter fraud that occurred was extremely small-scale and had no meaningful impact on the outcome.
Election deniers believe that the 2020 presidential election was illegitimate. I'll focus on the large subset of these individuals who blame voter fraud for the outcome.
(Roughly 6 in 10 Americans are election affirmers, according to a recent CNN/SSRS poll. Specifically, 96% Democrats, 60% of Independents, and 27% of Republicans considered Biden's win legitimate.)
A new poll
The poll, conducted by Rasmussen Reports on behalf of the Heartland Institute, was published on the Rasmussen and Heartland websites this December 12.
Rasmussen was a respected polling company that consistently yielded conservative-friendly data but has shifted further rightward in recent years. Liberal commentators disagree about how much that shift undermines Rasmussen's credibility, but the organization's reputation continues to decline among respected pollsters.
There's less disagreement about Heartland, thanks to its persistent efforts to discredit climate change, question the health risks of second-hand cigarette smoke, and otherwise promote a fairly extreme conservative/libertarian agenda.
I suspect that Heartland is the reason liberal media outlets mostly ignored the poll. This is a mistake, in my view. I don't think you should avoid data that burbles up from unsavory sources. Instead, have a look, especially when it concerns such an important topic. Whether you're an election affirmer or an election denier, you probably assume that the other side constitutes a threat to democracy.
Poll methodology
The poll, conducted between November 30 and December 6 of this year, included 1,085 likely voters.
The sample seems to capture a good mix of political ideologies: 36% of respondents described themselves as Democrats, 33% as Republicans, and 31% as other. Importantly, 45% described themselves as Biden voters, while 46% said they voted for Trump.
I have concerns about the sampling, but they're mostly speculative, so let's get to the data.
Main findings
21% of mail-in voters said "yes" to the question "During the 2020 election, did you fill out a ballot, in part or in full, on behalf of a friend or family member, such as a spouse or child?”
17% of mail-in voters said "yes" to the question "During the 2020 election, did you cast a mail-in ballot in a state where you were no longer a permanent resident?"
These were the most prevalent examples of fraud. On the whole, about 1 in 5 mail-in voters reported fraudulent voting behavior – a troubling figure, given that people typically underreport bad behavior on polls and surveys.
Heartland officials refer to that 1-in-5 statistic as evidence of "widespread" or "rampant" voter fraud. Donald Trump says that it proves that the election was "a giant scam". These kinds of statements are not helpful. As Heartland acknowledges, 43% of voters mailed in ballots in 2020. If 1 in 5 of those people committed voter fraud, we'd have an overall fraud rate of 8.6%.
I wouldn't call that "rampant", but it's still troubling. Just how troubled should we be?
Evaluation of the data
1. Voter fraud was not linked to voting behavior.
As I mentioned, most of the sample was evenly divided among Biden voters (45%) and Trump voters (46%).
What percentage of the Biden group engaged in fraud? What percentage of the Trump group engaged in fraud? The poll results don't say.
Mathematically, it's possible that the fraudsters could've all been Biden voters. They could've all been Trump voters. Or they could've been a mix.
Why on earth did Rasmussen/Heartland fail to link fraudulent behavior to voting choice? They have the data.
I reached out to Justin Haskins for clarification. Mr. Haskins is listed on the Heartland website as director of their Socialism Research Center and the primary author of the poll. Here's most of what he emailed back to me:
"Unfortunately, the overall sample size of the poll, 1,085, was not large enough to account for an accurate analysis of a subgroup (such as Trump voters) of a subgroup (those who said they committed fraud by mail) of another subgroup (those who said they voted by mail). Although the data does exist, it's not accurate."
Ho-ho-ho. If you're an election affirmer, Mr. Haskins just gave you a little present. He's saying here that the sample isn't large enough to link voter fraud to voting behavior.
If that's the case, the poll offers literally zero evidence that the election was stolen. We have no data on how many Biden vs. Trump voters committed fraud.
Trump himself, commentators at organizations such as Fox and Breitbart, and other Republicans have been openly claiming that the poll constitutes proof of a stolen election. Trump called it "the most important poll released in the last 20 years" (in all caps, naturally).
In fact, the polling data says nothing one way or the other about whether the election was stolen. It only says there was voter fraud. Even the author of the poll acknowledges this.
Still, the poll does seem to show that fraud was extensive.
For three years we've been assured (by liberal academics and journalists, for instance) that the 2020 presidential election was the most secure in American history, with a few isolated instances of fraud that had no meaningful impact on the outcome. The Rasmussen/Heartland poll strongly contradicts this narrative.
Should we be worried? I mean, data is data. Assuming that Rasmussen avoided sampling bias and didn't fudge the numbers, the voters themselves have just acknowledged fraud, right?
Well...
2. The statistics aren't credible.
As I mentioned, 17% of mail-in voters said "yes" to the question "During the 2020 election, did you cast a mail-in ballot in a state where you were no longer a permanent resident?"
Here's a place where a little data, plus common sense, tells you that you're looking at a false statistic.
One of the few prominent liberal organizations that covered the poll is the Washington Post. Columnist Philip Bump observed that the 17% statistic fails a smell test, on the grounds that about 6 in 10 Americans don't move out of state their entire lives.
Let's go beyond a smell test to what might be called a vision test. That 17% statistic turns out to be impossible when you look closely.
The U.S. Census Bureau reports that just over 7.5% of Americans relocated to a different state in 2020. Since there were around 168 million registered voters in the U.S. that year, a crude estimate is that about 12.6 million voters moved to a new state in 2020.
This is our starting point (which I'll revise shortly). Only about 12.6 million voters could've committed mail-in voter fraud.
In the 2020 presidential election, 66.5 million people voted by mail. 17% of that figure would be 11.3 million people.
Thus, the Rasmussen data could only be accurate if something like 90% of people who moved out of state in 2020 both voted and fraudulently claimed to be living in their former state of residence.
This is wildly implausible. Look even closer and you discover that the 17% statistic is impossible. Here's why:
That 12.6 million estimate actually errs on the side of being too high, because the Census Bureau includes everyone over the age of 1 when counting relocations. In 2020, 22.2% of the U.S. population consisted of minors. Thus, we can estimate that only about 5.85% of voting-age adults moved to a new state that year. 5.85% of 168 million registered voters equals 9.83 million mail-in voters.
What I'm saying now is that among Americans who moved to a new state in 2020, roughly 9.83 million of them voted by mail. Rasmussen's data implies that around 11.3 million of them voted by mail and did so fraudulently.
In other words, Rasmussen's data suggests over a million more fraudulent mail-in voters than there were actual mail-in voters. This is impossible.
I invited Justin Haskins to comment on these observations, but he hasn't responded.
Ho-ho-ho. Election affirmers are having a very Merry Christmas here. But what about that other key finding, where 21% of respondents said "yes" to the question "During the 2020 election, did you fill out a ballot, in part or in full, on behalf of a friend or family member, such as a spouse or child?” Doesn't that constitute quite a lot of voter fraud?
No.
Here the problem concerns measurement rather than statistics: The poll question is ambiguous. According to election laws, helping someone (e.g., a person with a disability) fill out a ballot is not fraudulent if you do it at an election site. It's only fraudulent when carried out off-site. Thus, we have no idea how many of that 21% group actually committed voter fraud.
3. The extent of true fraud is unknown.
We want voters to comply with election laws, but we also want election outcomes to represent the will of the people. These are closely related but not identical desires.
Suppose grandma is an eligible voter and wants to vote, but her arthritis is acting up. If you sit at your dining room table and fill out her ballot exactly as she requests, while she peers over your shoulder, you're breaking the law, but you're also supporting representative democracy. If you didn't break the law on grandma's behalf, the will of the people would be less accurately represented by one vote.
In theory, the Rasmussen/Heartland poll could've identified how many people filled out someone else's ballot differently from the way the other person would've wanted. I asked Justin Haskins about this; here's what he wrote:
"Based on the results of this survey, there is no way to know if a person misrepresented another person's intended vote. We didn't ask respondents who said "yes" to those questions whether they think they voted in a way that their friend or family member would have wanted. Although I understand your point, in a sense, it's actually irrelevant. It's illegal activity either way, and the ballots shouldn't have been cast. Part of voting involves taking the time to vote. If you're not willing to take the time to fill out your own ballot, your ballot shouldn't count, because it's not really your vote. It's someone else's vote, someone who probably already voted for himself or herself."
I think that both liberals and conservatives would disagree in part with those last few sentences. If you're elderly, sick, grieving, or temporarily bereft of normal vision or motor skills, the reason you don't want to fill in a ballot is not unwillingness to take the time.
In any case, as Mr. Haskin acknowledges, the poll doesn't reveal how many of the seemingly fraudulent voters, if any, completed a ballot in a way that misrepresented the other person's preferences. There's no evidence that anyone's vote failed to reflect their intentions.
Summary
1. The Rasmussen/Heartland poll seems to reveal mail-in voter fraud in the 2020 presidential election. But it doesn't link fraud to any particular group of people (e.g., Biden voters). Even the poll's creator acknowledges this. There's zero support here for the stolen-election claim.
2. The poll doesn't actually reveal mail-in voter fraud by anyone, because the main findings are either impossible or ambiguous. There's zero evidence here that fraud occurred during the election.
3. The poll reveals zero evidence that the will of the people was inaccurately represented in the 2020 election outcome.
Final thoughts
In this newsletter I tried to take a balanced, nonpartisan look at the new poll data.
Election deniers would probably still disagree with my conclusions, and they might find liberal biases in my analysis, but I suspect they'd also find at least some of my comments sensible.
I don't actually know how they'd respond, because I'm not sure if any of them read this newsletter. I mean, to be honest, I almost never read theirs.
(I know...siloing is risky...but Trump and associates have now lost every one of their 60+ lawsuits alleging voter fraud and other improprieties during the 2020 election. I don't read flat-earther forums much either.)
I almost never ask readers to forward my newsletters, but I encourage you to send this one along to anyone you know who believes voter fraud was widespread in the 2020 election.
If you're one of those election deniers, I won't challenge your denialism. I don't claim that this newsletter addresses all of your concerns. But I will share a more narrowly focused opinion: Everyone at the Heartland Institute involved in the reporting of this particular poll is lying to you. They are knowingly, deliberately lying about data on one of the most important and divisive issues Americans will be continuing to struggle with as we enter a new election year.
I say this with no fear of being sued for libel, because (a) the evidence of deceptiveness is overwhelming – I barely scratched the surface here – and (b) the folks at Heartland probably don't read these newsletters anyway.
Thanks to those of you who do read them! Happy holidays!