I invented a game I call “Find the Letter.” It’s pretty simple. You click the “start” button and see how long it takes you to find the letter that’s formed in the pattern of dots below. I ran this experiment with some colleagues; it took an average of about 8.2 seconds for them to find it.
Ready to play? Click the button.
How fast can you find the letter?
How’d you do? What letter did you find? Did you beat the average? If you didn’t, did you feel some pressure to find the shape once you passed the 8.2-second mark?
This game was an experiment, as you might have guessed, but not one about the perceived social pressure of that timer. Instead, it was an experiment centered on the human predilection to find patterns in randomness. As you may have guessed, there was no letter hidden in the dots. It was just random dots. Maybe they sometimes formed something letter-like; maybe, in the manner of the aphorism about animals bashing away at typewriters, they formed an explicitly obvious letter. But that wasn’t intentional. The size and location of each dot was always random and nothing more. (If you know JavaScript, feel free to check for yourself.)
The point is that humans are good at taking random assortments of information and finding patterns. This seems likely to be evolutionary; skimming the horizon and being able to pick out the shape of a bear had its uses. But we do it all the time in other contexts, too, spotting faces in patterns of brick or shapes in the clouds. There’s a term for it: pareidolia.
In the context of American politics at the moment, though, this tendency is problematic. What we’ve seen repeatedly since last year’s election is individuals sifting through enormous amounts of information related to the election and looking for patterns — patterns that they often find. After all, that was the task they had set for themselves: find the evidence of fraud in this set of documents or data. And, lo, they found the fraud, just as you might have found that letter. Another success for human pattern-seeking!
The problem, obviously, is that all of this fraud-seeking has yielded precisely no evidence of rampant fraud in the election. It’s safe to say that, thanks to the combination of the Internet and the prominence of the claims in the national conversation, there has never been an election subjected to as much scrutiny as last year’s. But even with that, even with actual experts looking and amateurs digging in, nothing has turned up besides the equivalent of, “Well, if you look at these dots running along here, you can see how it sort of forms an L.”
On Thursday morning, the House Oversight and Reform Committee held a hearing to consider some of these efforts. In particular, the committee focused on the “audit” of votes in Arizona’s Maricopa County, a review run by an active fraud conspiracy theorist and that served mostly to elevate various allegations that were quickly debunked. It was a very tangible demonstration of the find-a-letter effect, dropping scores of believers into a situation where they were tasked specifically with finding the fraud they believed existed.
Among the witnesses scheduled to participate was one invited by the Republican minority on the committee, an engineer and failed political candidate named Shiva Ayyadurai. If his name sounds familiar, it may be because he has long-claimed to have invented email, which he didn’t. He also ran in Massachusetts last year for the Republican nomination for a U.S. Senate seat, losing in the primary.
He blamed that loss on fraud, claiming that a million ballots were destroyed in a race in which about 265,000 people voted. This was not true. But it established Ayyadurai as a voice willing to bolster President Donald Trump’s claims about fraud, something that he parlayed into new prominence after the November election.
Ayyadurai has a very particular process. Clearly a smart man, he evaluates large sets of data and establishes patterns that he thinks can be used to define those data. Then, he identifies places where the data deviate from the pattern and builds up complicated explanations for why the pattern was important in the first place. And, presto, you have data that become suspicious by the standard he himself defined.
In the aftermath of the election in Michigan, for example, Ayyadurai created a complicated metric measuring expected support for Trump in state counties and compared it to vote totals. He created a lengthy YouTube video that included various asides about statistical patterns and was littered with complicated-looking Excel graphs showing dots scattered around various lines. With a solemn and professorial voice, he then slowly explained what we were looking at.
Imagine if you had taken the “find the letter” game very seriously and then been asked to explain in detail what letter you believed you had found and how you found it. Imagine that you viewed this presentation as deeply important and reinforcing your own identity as a savvy observer of dot-letters. Imagine the video that would result.
Ayyadurai was invited to the hearing Thursday because he was tapped by the Arizona “auditors” to similarly sort through the available information and find some patterns. And, sure enough! He found all sorts of letters statistical evidence of fraud (though no actual, direct evidence of fraud). Once again, he did so with alacrity. When the “auditors” presented their findings to the credulous Republican state senators who had ordered the review, Ayyadurai had a starring role. He raised questions about signature patterns and scans of votes, questions that were quickly answered. But still the same tactic: sift through lots of things, construct a pattern, elevate deviations from that pattern, make claims.
That Ayyadurai was invited to the Oversight Committee hearing at all is doubly embarrassing for the Republican minority. It’s embarrassing first because Ayyadurai is not credible, as reviews of his work have repeatedly shown. But it’s embarrassing, too, because it reveals that the Republican minority either doesn’t recognize that his allegations are ridiculous or doesn’t know how to evaluate the reliability of witnesses on such a critically important subject.
Some of this is obviously confirmation bias. Asked by Rep. Jamie B. Raskin (D-Md.) during the hearing who won the state of Arizona, Rep. Andy Biggs (R-Ariz.) said, “We don’t know” — which is obviously untrue. Biggs, a loyal Trump ally who has been tied to the protests on Jan. 6, is deeply invested in finding letters in those patterns of dots.
During the Oversight hearing Thursday, a witness from the Brennan Center for Justice pointed out that the Arizona “audit” was a function of “pressure” from Trump and the party. This is where the timer in our game came in: If you’ve been tasked with finding a letter urgently, it seems fair to assume you’re going to loosen your standards for what might apply.
That witness, Gowri Ramachandran, then pointed out Ayyadurai’s lack of credibility, calling his fraud claims “sad.” But in the end, Ayyadurai didn’t testify before the committee about his Arizona conspiracy theories. A staffer for the committee’s Republican minority cited a “scheduling conflict” that prevented his testimony. That this was the publicly offered reason for his failure to appear instead of the fact that he lacked any credibility on the subject is, again, a discredit to the elected officials who invited him.
