Despite what the media says, the 2020 presidential election is far from decided. According to the preliminary vote counts, Joe Biden is leading slightly in several states that are the key to the Electoral College.
If the vote counts are correct, then Joe Biden will become president in January. But are they? Some data anomalies suggest we need to take a closer look. Last night on "Washington Watch," I interviewed Dr. William "Matt" Briggs, a fellow at the Heartland Institute, former statistics professor at Cornell University, and author of Uncertainty: The Soul of Modeling, Probability, and Statistics.
Dr. Briggs said statistical information can help the Trump legal team's court challenges by identifying possible instances of fraud. For instance, mail-in ballots in Michigan saw an eight-fold increase from 2016 in people who voted only for president and filled in nothing else. In 2016, two percent of people voted only for president, but in 2020, it was 16 percent.
And remember, we're talking about mail-in ballots, where voters have all the time in the world to research other candidates and questions to fill in their ballot.
The discrepancy is "huge," said Briggs. Too big, in fact, to be "easily explainable just in terms of simple changes in behavior." He said that's an anomaly that must be explained.
And it's not just in Michigan. Briggs pointed out irregularities in Pennsylvania. "At one point in time, in the evening of the next day, three separate counties removed votes for Trump. That has to be explained," he said.
Someone in the media will object that these questions are cynical, partisan ploys to undermine democracy. "You're just mad that your candidate lost," they'll say. But Dr. Briggs made the case that answering these questions is important to instill trust in the election results.
"The explanation doesn't necessarily have to be fraud. But when we find these kinds of things, some kind of explanation is needed," he said. "It may be a perfectly innocent explanation. On the other hand, it may not be." In other words, it shouldn't be out-of-bounds to ask whether the emperor is wearing clothes.
Briggs continued, "this stuff has been formally researched before. It's not unusual, these types of questions." Rather, ballot marking where only the president is picked "conforms to other instances of known fraud," he said.
Rough estimates like this are the data science version of a "sniff check." If something smells funny, then we should "follow the science" and dig a little deeper. That's what President Trump and his legal team are doing. And they have the legal right to do it.
As Americans, we have a right to verify, as I quoted Senate Majority Leader Mitch McConnell yesterday, that "all legal ballots must be counted; any illegal ballots must not be; the process should be transparent and observable by all sides, and the courts are here to work through concerns."
The only real threat to democracy here is the alarming (but predictable) hostility towards openness and transparency among the self-proclaimed defenders of openness and transparency. Ignore the media's cries to "pay no attention to the man behind the curtain." Let's have a fair count and then award the White House to whoever really won.
