Monday, November 7, 2016

The Only Election Guide For 2016 You'll Ever Need (Aside From All The Others)

Edit: Since I posted this yesterday, the polls have edged ever so slightly towards Clinton. Most of what I say below still holds, since we're still within margins of errors in a lot of places, but if it makes anyone feel better/worse, Clinton's chances are probably better than what I indicated below.

Tomorrow is Election Day, so let's take a look at what could happen.

As of right now, it looks like the states really too close to call are Nevada, Ohio*, North Carolina, Florida, and New Hampshire. Michigan, Iowa, Arizona, and Georgia are also unusually close, and there's an outside chance that Pennsylvania, Colorado, Wisconsin, and Minnesota are also close to the margin of error, although not likely to change. Finally, both Maine and Nebraska have congressional electoral votes that may vote differently than the rest of the state.

But this being as close as it is, looking at those five closest states are probably what will be the most important.

If we subtract out those five states, that leaves Clinton with 269 and Trump with 197. 270 are needed to win.

Why Republicans Should Be Panicking
Well, you should be panicking because you nominated Donald Freaking Trump four months ago. But ignoring that for now, the polls have had you behind the entire race. The few bright spots that you see are blips on the pollster's radar, barely-nuanced margin-of-error leads in states you absolutely, positively have to win. You have to win all of the contested states, and hope that Arizona and Georgia don't fall apart on you--which is possible, because they're within the margin of error on her side.

In addition, early voting has already happened in many key states, and those votes were cast when you were well down in the polls. The only upshot to this is that early voters are probably not undecided voters, 

You should be panicking because it's almost impossible for you to win. Unless...
Why Democrats Should Be Panicking 
Polling errors!

Right now, you're sitting on a sweet 269 electoral votes. There are a few shaky states that make up some of that, but not that bad. The polls have favored you there for a long time.

However, those last five states might give you a headache. Ohio, for example, has been pretty solid in Trump's column, albeit barely. Florida and North Carolina have also favored Trump, but not nearly as much. Your support in Nevada and New Hampshire has been much, much better--and remember, you only need to win one of these. You literally only need one your electoral vote to win, and that includes Nebraska's extra vote.

However, this year has (of course) been unusual...there are an unusually high number of undecided and third party voters. Your throne of 269 electoral votes assumes that these voters will break more or less in your favor...which, historically, they break for the challenger. There's also a collection of other factors--is there a Shy Trump factor, voters who are afraid to say they support Trump to pollsters but eventually will? Is there some slice of demographic that the pollsters are missing that favor Trump? Is there a reason why these undecided voters may break heavily for Trump? Most of these will probably have little to no effect--"Shy Trump" voters seems highly unlikely--but even a .5% edge towards Trump will cause trouble in many states. Leads in states like Colorado and Michigan are almost within the margin of error, and both of those states in particular have been trending down. Remember--the polls were off by almost 3 percentage points in 2012 for the nationwide vote.

The good news is that the polling errors aren't inherently against you--these could be pointing towards a pro-Clinton stance as well. But there's no way to know for sure until the votes are cast. Polling errors aren't your enemy--uncertainty is.

What You Shouldn't Have Been Doing These Past 12 Months
Curating your echo chamber. I get it, people on social media piss you off, but if the only people you hear are the people you already agree with, you have no means to gauge exactly how popular someone is. I've heard a lot of variations of "Everyone I know is voting for Clinton/Trump, so if they don't win it's obviously rigged!" Well, no, because you've spent twelve months slicing out the people who would let you know how much in trouble your candidate is in. Add in to this the fact that most people cluster in real life with people whom they agree with or at least are in the same socioeconomic strata, and the average person literally has no idea how the nation at large might feel.

The First Sign Of The Apocalypse
Any political science geeks out there probably already blanched at the one number I listed above: 269. If Trump manages to sweep the five states I listed above, that means a 269-269 split. Go buy as much popcorn as you can, because this crazy dumpster fire train is going to go off the rails. 

But What About The Senate?
Who knows? It's always going to have been a bad year for Republicans, since they had to defend nearly every single seat. Illinois is a lost cause, and Wisconsin nearly so (although there's been a surprising resurgence in the last week, but probably not enough to win); however, Republicans were poised to retain control with a safety valve in Nevada (which would flip from Democrat to Republican). They could afford to lose a few seats. But as it stands now, they will be struggling--there are five (!) races that are less than a percentage point apart, so it could literally go either way. Each of these races will probably end up going with the Presidential victor, since it would only take a very small number of "extra" people to vote to put these Senate candidates over the edge. Republicans can rest a little easier, though; for a while it looked like they might lose Missouri and North Carolina, which would more or less seal their fate, but the polls have improved there.

Why You Should Be Crying Throughout The Night
Similar states tend to correlate with one another, depending on their demographics. For example, if Clinton wins Ohio, don't pin your hopes on Trump winning Michigan--if Trump has lost Ohio, he's already lost Michigan. So if you're going to spend the night thinking "Oh, he or she is behind, but returns from this nearly identical state haven't come in yet!" you are going to be sorely disappointed. The caveat is if the won state was unusually close, there's always a shot for a similar state to flip--but then again, if the results are unusually close, no one is going to be calling it anytime soon.

What To Look For Today:

Maine: Maine's 2nd congressional district is slightly leaning Trump. But not by a lot. If Trump wins this district, it probably means the polling errors nationwide are going his way, and expect it to be a long night for the Democrats. 

New Hampshire: New Hampshire is one of the states Trump needs to win--if he loses NH, he'll have to win some other, harder state later in the night, which of course gets more and more unlikely as the night goes on. If Trump does win, all is not lost for the Democrats, but, again, it's going to be a long night.

Rust Belt Returns:  Most pundits will be looking at both turnout and the margins in places like Pennsylvania, Ohio, Wisconsin, and Michigan. If it is unusually close in Pennsylvania, for example, that spells trouble for Clinton even if she ends up winning that state--that would most certainly put Ohio out of her reach, and might even signal trouble in what should be a sure bet in Michigan. The good news for Clinton is that these results will mostly be coming in at once, and there are few other similar states, so as long as she passes this firewall she should be fine. A win is a win, regardless of the margin.

Arizona: Arizona is Clinton's safety valve. Even if she does poorly elsewhere in marginal states, she might be able to pull off a win in AZ. It's not likely, but it's possible. The good news for her is that AZ is largely an outlier--good rust belt returns for Trump won't mean much in the southwest, for example, and most other similar states aren't really in play. If Clinton is having a bad night, this may be her saving grace. The drawback to this is Florida--if Hispanic turnout is enough to flip Arizona, chances are it would have affected the only other similar state, Florida, hours earlier. Florida is its own beast, of course, but it's something to take note of.

So What Will Be The Final Outcome?
Who knows, but the signs still point to a Clinton victory. However, there's enough statistical noise that it probably means there will be at least one or two razor-thin victories one way or the other; the question is where these will happen. For some states it won't matter for Clinton, but in others it absolutely will. Fivethirtyeight's blog pegs it at around 65%-35% for Clinton--a pretty solid outcome for Clinton, but far from a trivial chance for Trump.

*Polls from Ohio are pulling much more for Trump lately, and has been for a while, but I'm still rating it as too close to call.  That's probably being overly cautious.

No comments:

Post a Comment