Prediction time

So, like a good social scientist, I think I’m going to go with what makes the best sense as a social scientist.  At last check, Drew Linzer, Sam Wang, and Pollyvote all basically had this map, that’s good for me.  Nate Silver has NC, NV, and FL all tilting just Trump, so it will really be interesting to see who performs better.  For what it’s worth, the betting markets also are pretty much on this map.

So, where do I think I (and this somewhat consensus) might be wrong?  I actually NC is the likeliest of those 3 to actually be red (I really think Hispanic voters are being improperly polled in the other two states– just not that many in NC).  Given that HRC is making a last minute push in OH, I also would not be at all surprised to see her grab that and push up to 341.  I could also potentially see IA and NH flipping colors.  As for all the remaining states, I would be quite surprised if they did not end up the color on this map.

As for the final national margins?  RCP is at a 2.2 margin, Pollster is at 4.4 (2.2 with less smoothing), Sam Wang’s meta-margin is at 2.2.  So, why not go with that?  Because, based on seeing too many polls with really questionable internal crosstabs (on Hispanics, especially), I suspect that there is a systematic bias in favor of Trump.  I’m thinking Pollyvote’s 5 point margin is probably more like it, so I’m going to say something around there.

Long ago, I predicted Gary Johnson would be over 5%, now I’m not so sure.  Would not be surprised to see him at only 4%.  Though, I would be surprised if he goes lower than that.

Now, all that numbers and social science aside, I find it truly, truly horrifying that a transparently racist, misogynistic, xenophobic, shockingly ignorant, breathtakingly immature, etc., individual can get anywhere near this close to the presidency.  Truly, someone of Trump’s character and beliefs should be losing by 20, and as much as I get the power of partisanship, it depresses me that, no matter what, Trump will not have the full-on repudiation he deserves.


About Steve Greene
Professor of Political Science at NC State

One Response to Prediction time

  1. ohwilleke says:

    The big potential inaccuracy I see is that Dems seems to be performing very well in early voting turnout, really more or less unprecedented compared to historical patterns, and I have real doubts about whether the polling models are properly comparing likely voters (who may be perhaps 90% likely to vote in reality) against already voted individuals (who are 100% likely to vote and may be less inclined to speak to pollsters), resulting in underweighing of early voters who seem to be more liberal than later voters this cycle based upon the surveys I’ve seen so far (such as a University of Denver poll with a very stark early (45D v. 38R) v. late voting (37D v. 45R) partisan split in Colorado with a 4.2 percentage point MOE).

    The early voting trend also seems to be present in FL, NC and NV. And, with a weak ground game for Republicans, election day GOP turnout may be lower than usual.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: