Many news reports today are suggesting that computer experts are worried that the 2016 U.S. presidential election was hacked, and are demanding a recount to see whether Hillary Clinton actually beat Donald Trump, but that's not true. What electronic-voting experts are suggesting instead is a vote audit of the kind that they think should be routine after every election.
This pervasive but mostly false story stems from a misleading report posted online yesterday (Nov. 22) by New York magazine entitled "Experts Urge Clinton Campaign to Challenge Election Results in 3 Swing States." It says that "a group of prominent computer scientists ... believes they've found persuasive evidence that results in Wisconsin, Michigan, and Pennsylvania may have been manipulated or hacked."
But in a blog posting today (Nov. 23), one of those experts cited, J. Alex Halderman of the University of Michigan, contested the accuracy of the New York magazine story. And Nate Silver of fivethirtyeight.com, one of the country's most prominent statisticians, said there's no evidence that the presidential vote in Michigan, Pennsylvania and Wisconsin was rigged.
The New York Magazine piece "incorrectly describes the reasons manually checking ballots is an essential security safeguard," Halderman wrote.
"Were this year's deviations from pre-election polls the results of a cyberattack? Probably not," he added. "I believe the most likely explanation is that the polls were systematically wrong, rather than that the election was hacked."
However, Halderman said the ballots should be audited nonetheless, because otherwise there would be no way to know for sure. And unfortunately, the only to do that is to demand a recount — not to give Clinton another chance to beat Trump, but to make sure that the various voting systems in use across multiple states actually work properly.
Halderman admits that the election certainly could have been hacked, because many of those voting systems are not very secure. He and other verified-voting experts have argued for years that electronic voting is very risky unless each recorded vote leaves a paper trail. (Indeed, the Netherlands has abandoned electronic voting and gone back to pencil and paper.)
But a stolen election would probably be easy to spot statistically. To that effect, Silver said on Twitter yesterday that "the claim ... of rigged results in Wisconsin is probably BS," and displayed several charts to demonstrate that. (Good luck understanding them.)
"Run a regression on Wisc. counties with >=50K people, and you find that Clinton improved more in counties with only paper ballots," Silver tweeted. "HOWEVER: the effect COMPLETELY DISAPPEARS once you control for race and education levels, the key factors in predicting vote shifts this year."
"Nothing in Pennsylvania, either, whether or not you control for demographics," he added. "And Michigan has paper ballots everywhere, so not even sure what claim is being made there."
Another prominent statistician, Nate Cohn, who was hired by The New York Times after Nate Silver left the paper following the 2012 presidential election, also dismissed the New York Magazine story.
"It's hard to stress how weak this is," Cohn tweeted yesterday.
That's not to say there aren't issues with the 2016 presidential election. Many computer-security and national-security experts believe that the Russian government influenced the election by releasing compromising material about Clinton through WikiLeaks — and got the result Russia desired. In other words, a hostile foreign power may have swayed a U.S. presidential election.
There's also the continuing vote count in California, which so far has added so much to the Democratic candidate's total that she now leads by more than two million votes nationwide. But both of these are policy and legislative issues that have little to do with voting machines, and the next president will have to deal with them.