Friday, May 20, 2016

Political Science to Trump Denialists: Drop Dead (Those pesky polls)

I started the Trump to Political Science: Drop Dead series to show all of the ways that Trump's success in the Republican contest upended conventional political science.  Now, as we turn to the general election, it is political science's turn.  You see, we have always had better models to explain general elections.

And this time, the polls are cooperating.

If you head on over to RealClearPolitics, you will see that the two most recent polls, one by Fox and one by Rasmussen, show Trump leading Clinton by small margins.  Hillary still has a narrow lead in the polling average, and you should always pay more attention to the polling averages than to any individual poll.  But, what's going on here?

What's going on here is that for the general election, political science is actually pretty good.  How's the economy doing?  According to Federal Reserve Economic Data, first quarter GDP for 2016 is a 0.5% increase.  Small, but positive.  Add to that the fact that the Democrats have won two elections in a row, and that is the set-up for a very close election, or perhaps one that even leans Republican, depending on how the current quarter plays out.

The basic economic conditions and the sequence of past elections predict a close election, with a possible GOP advantage.  The polls are showing that.

And Trump could easily outperform expectations!  I have written a few posts on that in the past, such as this one on how survey respondents may understate their support for Trump, this about the potential benefits of Trump's malleability, and this about how Trump could turn Republican distrust into a general election advantage.

Regardless, I've been saying it for a while now.  Trump can win the presidency.  All he has to do is not dramatically underperform given the political fundamentals, and he has the potential to overperform.  Then, we turn back to the current version of Abramowitz's "Time for a Change" model, with a variable added to reflect polarization.  Partisan polarization gives any given candidate at a floor below which their numbers cannot reasonably fall.  If the economic and political fundamentals create either a near tie or a Republican advantage, then that floor makes Trump's idiosyncrasies less important.

Trump can win.  I've said it before, and I'll say it again.  Don't write him off just because he is Donald Trump.  That's why political scientists kept denying his chances long after his victory became clear.  If you want to get it right, pay attention to the data.

BIG CAVEAT, SO PAY ATTENTION TO THIS TOO!  The trick right now is that while the Republicans have begun the process of unifying around Trump, Bernie Sanders (the Ted Cruz of the left) is still either deluding himself or lying to his followers about the state of the Democratic contest, delaying the process of the Democrats unifying around the person who has already won-- Hillary.  Hillary did the same thing to Obama in 2008, and the party managed to unify anyway.  Does this presage a failure to unify for the Democrats?  Probably not.  Sanders will enthusiastically endorse Hillary over The Donald.  Once that happens, Democratic voters, who have memories roughly the lifespan of a fruit fly, will unify around Hillary.  If you look at the polls in the intervening period between the unification of one party and the unification of the other, you will get a bias towards the party that has already unified.  How much should we read into two polls that put Trump ahead?  Not a lot, but more than nothing.

No comments:

Post a Comment