I spent the Last month, with alternating apprehension and joy, President Trump’s cynical legal efforts to overthrow the presidential election were observed to have turned into absurdity. After dozens of lawsuits were thrown out of court and votes were certified in the contested states, I thought we had come to the end of the road. But it turns out there was one more whip to deliver, a bright red line that no scientist like me can bear to see crossed. It’s true, Donald Trump abused statistics.
The Texas Attorney General filed a trial Monday asking the US Supreme Court to intervene in the poll. Before your heartbeat changes too drastically, I must tell you that the legal experts are looking into the case “condemned. That’s not to say the trial can’t be dangerous. He introduced the odd but real “quadrillion” number into political discourse for a few news cycles and sowed a new set of digital conspiracy theories that could endure for years as so-called evidence of voter fraud. Tuesday, like 18 other states ready to support Texas trial, press secretary Kayleigh McEnany tweeted “The odds of Biden independently winning Pennsylvania, Michigan, Georgia, and Wisconsin after @ realDonaldTrump’s lead starts are less than one in a quadrillion.” She then entered the number with all of its 15 glorious zeros.
Since President-elect Biden has won all of these states, his odds of winning them are 100%. However. How this statistic was created, and then released in seemingly authoritative documents, is all too familiar to me as a doctor who relies on scientific literature. I want to suggest that the baseless lawsuits and medical research we use to guide treatments shouldn’t use the same statistical tricks.
Science is a challenge just like political polls. We are asked to explain how the whole world works when we can only see a small part of it. A pollster wants to know how the country will vote by calling a few people. Likewise, if we want to know if a treatment improves a medical condition, we can only afford to test it in hundreds or thousands of people – although it can ultimately be given to millions. Modern statistics have tools to deal with these situations.
the economist Charles Cicchetti, who calculated the electoral estimate of one in a quadrillion, used one of these tools, called the “null hypothesis significance test”. The idea is simple but insidious: can we use statistics to prove that a hypothesis about how the world works is compatible with what we actually observe? The insidious part is how you choose your hypothesis.
I guess the basic calculations behind Quadrilliongate are correct. If the voting pool counted on election night and the later voting pool was drawn at random from the same pot, with the same mix of Trump and Biden voters, then yes, of course you would expect that. the results are about the same. And of course -[math, math, math]—Perhaps the chances that an early Trump lead was reversed are very low, as one in a quadrillion small. But the problem is with the assumption and what the complainants seem to think it means. Cicchetti attempted to prove that “the votes totaled in the two periods could not be random samples of the same population of all votes cast.” Do you still see the problem? This is exactly what we were Told would happen months in advance: a “blue shift” stemming from Democrats favoring postal voting and Republicans leaning towards in-person voting. Almost proof of fraud. Cicchetti admits there is “some speculation that the ballots that have not yet been counted were probably mail-order ballots.” Of course, this is not speculation. The day after the election, for example, the Secretary of State of Georgia ad there were approximately 200,000 postal ballots left to count.