Saturday, June 18, 2016

Did A Stanford Study Prove Election Fraud? HELL, NO!

"A Stanford study proved it!" seems to be the latest battle-cry of the die-hards in the Bernie camp. They point to an article widely distributed on social media that demonstrates what the authors believe to be mathematical proof. One or two have mentioned a Harvard study as well, but that hasn't made the kind of headway the Stanford article has. So pervasive has this purported "study" been, that it's swayed even comedian Lee Camp, who cited it as a source for his news & comedy show Redacted Tonight. You can see the entire episode here.

"Yet more evidence has come to light that this election was stolen from Bernie Sanders," he opens. "And it's getting to the point where a lot of us are going, 'How much evidence do you need?!' It's like, we're trying to prove a guy's dead, and there's, like, a body lying on the ground for days, with no pulse and he stinks and there's like, flies, and like, a medic-alert bracelet that just says, 'This guy's gonna die for like, no reason!' And yet still a lot of Americans are like, 'Let's not jump to conclusions! All right? Maybe he's just inspecting the floors, for cracks, for like several days. Has anyone tried asking him whether or not he's dead?'
"The latest evidence is from a little out-of-the-way community college called Stanford."

He then goes on to share one item from that "study" which I will reference below.

Ah, Mr. Camp! I love you. You're one of us. But you are as hilarious as you are flat-out wrong!

I don't make this accusation lightly. When smart people get fooled by not-so-smart stuff, I tend to get pissed, but with the stakes as high as they are now, it seems even more obscene. Honestly, I haven't had to deal with this much information since I took up the odd hobby of debating creationists, and believe me, they come up with some serious mind-bending twists on what they think is "evidence" that evolution is fraudulent.

We humans are good at fooling ourselves.

But let's get down to brass tacks. You can read the "Stanford study" for yourself here. Except it's not from Stanford, and it's not much of a study, either.

The paper actually comes from two people: Rodolfo Cortes Barragan, a student (Possibly a masters student, but who knows? Could be an undergrad for all we know.) from Stanford University, and Axel Geijsel, another student (Grad? Undergrad?) from Tilburg University in The Netherlands. These two kids, operating without any endorsement from Stanford or any of its faculty members, assembled a small summary report and submitted it to a website called The Bern Report, along with a link to the complete study (as cited above). The Bern Report article can be seen here.

The paper is comprised of two basic parts. The first part shows that states which did not have a paper trail showed better results for Hillary Clinton than Bernie Sanders, while states that did have a paper trail showed the exact opposite. The second part claims to show that exit poll anomalies indicate that the final vote tallies don't jibe with the actual vote. Sound familiar?

Let's deal with the first part. A widely-shared pie chart that breaks down the argument into an easy-to-understand glimpse, and it's the item which was shown on Lee Camp's program. (See? I told you I'd reference it down here!) Here it is:

So the basic argument is this: It's easier to cheat in states that don't have a paper trail. Hillary won big in states that don't have a paper trail. Therefore, Hillary must have cheated, right?

The fallacy should be obvious. Potential for cheating does not equate to actual cheating, nor are lax rules any guarantee that rules will be broken. True, when the cat's away, the mice will play, but the absence of a cat does not automatically produce mice!

Imagine being a high school football player. You go out on a particular play, catch a pass, and score a touchdown. Right after the game, the state high school athletic association suspends you and your team for the use of steroids.
"But that's not fair!" you say. "I didn't use steroids! None of us did! I scored that touchdown fair and square!"
"Yes, but it's easier to cheat with steroids in high school sports than it is in college or the NFL. Clearly your success means you cheated."
"What?!" you shout. "That's ridiculous!"
Of course it's ridiculous. But it's exactly the same logic used in Geijsel and Barragan's report. Convicting someone of anything based on no more evidence than opportunity is a miscarriage of justice, and we should all be pissed off that anyone would assume such a thing.

But all that is just common sense. There are also solid numbers of my own which contradict the conclusion that these two students half-assedly slapped together. Conveniently provided with their report is a link to the spreadsheet containing their data. Big mistake, because now I get to apply my spreadsheet mastery, and I AM the master!

Geijsel and Barragan's data has a glaring error immediately visible to all who know anything about percentages. See if you can spot it:

In case you missed it, here it is: Each state is listed with its winning percentage. To obtain the percentages that Geijsel and Barragan put on their pie chart they break out the non-paper trail states (for example), add up all the percentages of those 10 states (for a total of 656.85%), then divide that by ten for a total of 65.685% This is where the pie chart gets 65% (rounded down for some reason).

BUT SOME STATES ARE BIGGER THAN OTHERS! Holy shit-fucking-DUH! A so-called Stanford student missed that?

Yes! If one is to do the calculation properly, one must give each state the actual weight based on the actual number of people who voted. Thus Texas has way more people, and had way more votes, than let's say Mississippi. So Hillary's huge win there would be tempered by the smaller margin of victory (but a clear victory) in Texas. The weighted percentage including the actual number of votes cast in each state looks like this:

Oh, shit! As you can see, the percentages shift a little bit when the actual population size of the vote is factored in. Clinton still shows a significant lead in states without a paper trail, but look at the states that DO have a voting paper trail! Hillary wins there too! Why? Because she won big-population states with paper trails like Ohio and New York!

Oops! Somebody needs to re-do their math homework.

Now let's see another pie chart that shows the percentage of actual votes cast in paper-trail states vs. non-paper-trail states.

As you can see, the total number of votes cast in paper trail states is significantly greater than the number of votes in non-paper trail states. Which means, even if Hillary won a vast majority in non-paper trail states, it wouldn't have mattered much. She still won by a fairly decent margin!

But even after all that, we are only half way done! The second part of this paper claims that exit polls show that the official vote does not jibe with the actual vote.

Oh really? I asked if that sounded familiar earlier, and it should. The reason why is because I dealt with exactly that topic in an earlier blog post. And wouldn't you know it, what do I find in the second half of their paper but this:

"Data Procurement: We obtained exit poll data from a database kept by an expert on the American elections."

Oh, wonderful! And just who is this wonderful expert whom they cite? Wisely, they do not name him. But comparing the table notes, the source is made clear. Why, it's none other than asshole extraordinaire Richard Charnin! The same moronic idiot I debunked in my earlier blog post!

Oh, you've got to be kidding!

When Snopes investigated the claims of the article, it rated it a mixture of truth and fiction. You can read that Snopes report here. But Snopes did a half-assed job of it. They correctly caught the fact that it was not a Stanford study, but the claims in the article and the math behind it they rated as true.

Never mind that the conclusions based upon it were false, and wishful thinking to boot. Never mind if the pie-chart was based on data apportioned by state and not weighted by votes on the ground. If one weighs that in (as Snopes should have) it would be 100% false.

Mark Twain said it best when he said that there are three descending orders of falsehood: Lies, damned lies, and statistics.

As if I needed anymore ammunition against this doltish duo, they added this to their paper:

"Statement on peer-review: We note that this article has not been officially peer-reviewed in a scientific journal yet. Doing so will take us several months."

No fucking shit this hasn't been officially peer-reviewed in a scientific journal! Does anybody think this shit would pass a peer-review? They say it would take several months. Try ever! Half their report is shoddy math, and the other half is borrowed from the proven-false work of a proven-false conspiracy theorist.

Consider your asses peer-reviewed, you shitheads!

Here's what Snopes had to say about Charnin, and keep in mind, this person had not researched the asshole as thoroughly as I had:

"The expert whose numbers were utilized for the paper wasn't expressly cited by name, but his moniker appeared on the linked spreadsheet: Richard Charnin. Charnin indeed lists some impressive statistical credentials on his personal blog, but he also appears to expend much of his focus on conspiracy theories related to the JFK assassination (which raises the question of whether his math skills outstrip his ability to apply skeptical reasoning to data)."

No fucking shit. Jesus, how did this shit ever get taken seriously?!

One more point, and then I'm done. The "researchers" put in their paper one phrase I feel is worth taking note of. They claim to have made a comparison analysis with the 2008 elections and concluded that, "no such similar irregularities occurred in the 2008 competitive election cycle between [Senator] Clinton and [Senator] Barack Obama. As such, we find that in states wherein voting fraud has the highest potential to occur, systematic efforts may have taken place to provide Secretary Clinton with an exaggerated margin of support."

No such irregularities in 2008, huh? Well their guru, Richard Charnin, disagrees. In his own blog post, dated 16 Sep 2012, he shows what he believes to be systemic election exit-poll-related fraud beginning in 1988 leading up to and including 2008.

Now, I don't by any means consider Richard Charnin a credible source. But Geijsel and Barragan do, and that source flatly disagrees about exit poll discrepancies not existing in 2008.

As if to excuse any mistakes they may have made, the "report" opens up with this statement:

"Important Note: This document is a living document. We are updating the numbers as they become available."

Yeah, how about you update the WHOLE THING?!

I find it an odd coincidence that the tone and attitude matches that of every creationist I've ever faced off against. One mindset comes from the right, and the other from the left, but the breathless and manic anger is hard to miss. Okay, I exhibit the same anger, but primarily because my friends, my freethinking family even, are being seduced by this crap! So yes, I am justified in saying that it reminds me of a creationism vs. evolution debate. No, the two are not the same! I'm not making an equivalency comparison. But I am making a conspiracy theory comparison. Fraud-creationists (because that's what they are), are essentially chasing UFO's (unsubstantiated fraud opinions) which aren't really there, but the die-hard Bern-outs are convinced they are, and solely because of them and their shoddy fact non-checking. Maybe they're mixed in with the uncounted California provisional ballots! Or hidden inside Area 51!

Geijsel and Barragan have convinced people of the existence of such phoney fraud.

Well, fuck these guys. FUCK these guys!



Addendum, July 3, 2016:
I based by calculation on actual vote totals to do the weighted percentages. But the calculation in this fake Stanford study used the percentages from the exit polls. I discovered my mistake accidentally when leaving the weighted percentage out on a different project and found that Hillary wins using that calculation too! Only unadjusted early exit polls can skew the vote so much that it appears like a Bernie win. So I wasn't quite perfect in my criticism. However, I have thoroughly dealt with the subject of exit polls elsewhere. I can therefore stand by my work, because my conclusion was fundamentally correct, and the authors of the Stanford paper were still thoroughly wrong.

1 comment:

Nancy Hall said...

Nice work. I knew this was bogus the minute I saw it, but I didn't do a thorough a job debunking it as you have.

A couple of observations: I question their categorization of states as having paper trails and not having paper trails. I checked Wisconsin, where I live and have been a poll worker for 14 years, and see that it's listed as a paper trail state. My county, Milwaukee, uses ES&S's DS200; which scans, tabulates, and stores paper ballots. Until recently, we offered an older optical scanner as well as a touch screen voting machine at each polling place. Many areas in Wisconsin still use the touch screen machines exclusively, so WI is not a 'paper trail' state. It's mixed.

I also question the validity of the assumption that exit polls are accurate. The whole basis for their argument is that the exit polls are at variance with the actual vote in states where there's no paper trail. They don't offer any scholarly study to back their assumption. I did a cursory Google search and found a lot of articles from people like Nate Silver declaring that exit polls are not reliable. So...their fundamental assumption seems to be flawed.