Friday, October 21, 2016

Things that Irrationally Scare Me About the Polling in this Election

After three debates, and a ton of horrible stories coming out about Trump, and his campaign staff starting to leave, it looks like Hillary Clinton has this election wrapped up. In fact, the question in everyone's mind is not whether she'll win, but by how much. And as a Democrat, I'm pretty happy about the state of affairs. But there are three things that sit in the back of my mind that concern me.

1) The Wilder Effect:

If you know polling, or if you know of Virginia politics in the 1990's, you'll know about the Wilder Effect - an effect where voters will purposefully lie to pollsters about a candidate or based on the voter's perception of the pollster's preferences. It is also known as the Bradley effect, but in either case, African American candidates had better polling numbers than election results because voters didn't want the pollsters to think they were racist by choosing to vote for the white guy. Now, in 2008 and 2012, there wasn't any Wilder/Bradley effect present for President Obama. His poll numbers and his election numbers were pretty much spot-on, and guys like Nate Silver were able to predict Obama victories within a hundredth of a percentage point.

Now, with the race between Clinton and Trump, one would assume that there wouldn't be a Wilder/Bradley Effect because a) it didn't exist for Obama; and b) both candidates are white. But, and this is what concerns me, Trump is currently really, really, toxic. He's a racist, a misogynist, and pretty much the stereotype every Democrat has for a Republican. His brand is so toxic, that it would not be at all surprising to learn that some Trump voters are simply ashamed to admit their preference for him, and the polls are, thus, slightly skewed.

Before we all freak out, the one thing that keeps me sane is that no discernible Bradley/Wilder Effect was present. Trump's wins were pretty much spot-on with the polling. The other thing that helps here, and which I just learned about 5 minutes ago, is that polling tends to underestimate the votes women candidates get. So, its entirely possible that both the Bradley/Wilder Effect and the Richards Effect are present in the polling, and that they'll cancel each other out. Its also possible that only one is present, and the vote will end up a lot closer, or more of a landslide, than we originally thought.

2) Third Party Candidates

What's unusual, but not entirely surprising about this election is the number of somewhat relevant third-party candidates present in the race. The Libertarian Party candidate, Gary Johnson, who's kind of an idiot (though his VP nominee is not), is polling around 7-9%; Jill Stein with the Green Party (a bigger idiot), is polling between 2-3%, and some random guy named Evan McMullin (not really random, I know) may actually win Utah. McMullin is actually up by 6 points over Trump and Clinton in Utah. And since no one has done any polling in Idaho (which has similar demographics to Utah), he might be up in Idaho too.

So, no doubt about it, the 3rd party candidates are more relevant in 2016 than they have been in any election since 2000, when Ralph Nader took away just enough votes to cost Gore the election. And there, he took less than 5% of the vote. Here, we have 3rd party candidates possibly taking states, and getting 10-15% of the overall vote. Those totals were last seen in 1992. But the people who say they're voting for a 3rd party, are probably amenable to swinging to either Trump or Clinton. So, will they stay out of it, or will they swing one to one of the mainstream party candidates? I don't know, and that concerns me. Slightly, but it does concern me.

3) Turnout

This is the most interesting question of the race - what will the turnout be like? If Clinton voters think this one is in the bag, will they bother to turnout? I'm guessing yes, because they want to see Trump beaten. Will Trump voters turnout to vote? And if they do, will they vote the full ballot? Will the Voter ID laws prevent people who are now classified as likely voters from voting?

What we do know is that Trump's advantage on likely voters is really, really small for a Republican - like around 1%. We also know that Trump has no ground game, and the guy who is supposed to run the ground campaign resigned today. That bodes ill for turnout being in Trump's favor. On the other hand, Trump has been encouraging his supporters to start poll watching* (i.e. discourage minority voters from voting). Of course, with early voting, that may not have any effect whatsoever. Plus, Clinton does have a ground game, and that may also overcome various obstacles. Honestly, I have no idea how this all goes (though it doesn't look good for Trump).

*The Republican National Committee is legally prohibited from poll watching due to efforts to prevent minority voting in the 1980's.


No comments:

Post a Comment