A Note On Summer Polls [UPDATED]

There are all kinds of reasons to ignore the polls in the summer, but understanding the issues raised here will serve you well once we get into the fall.

If you’ve read the news over the last 48 hours, you’ve undoubtedly heard that the race has “tightened” and that Hillary’s “negatives” are rising. Most, but not all, of this is being driven by the results of Quinnipiac’s polls. It’s way too early for polls to mean anything of lasting importance. The only contest that matters is the one that happens on Election Day, so this contest isn’t like a regular season GSW game where the team can run up the score in the first three quarters and then coast at the end. Just as the Warrior’s ability to blow out Cleveland in the regular season meant nothing once they met in the NBA Finals, so too here.

That said, the movement in the polls does teach us something about polling itself, so it’s worth spending a moment or two looking at the polls driving this latest wave of otherwise pointless stories.

This early out, the pollsters can do nothing more than guess at what the eventual turnout will look like. The pollsters at Quinnipiac, to their credit, explicitly say as much in their own explanations of their latest results. As they explain, they used “guess work” to predict that white voters will account a much larger share of the vote in 2016 than they did in either 2012 or 2008, and more specifically that whites will turn out at levels we haven’t seen in decades WHEN MEASURED AS A PERCENTAGE OF THOSE WHO VOTE. And yes, that’s in all caps because it is an essential point to understand. They aren’t simply predicting a surge in white turnout; they are instead predicting a surge in white turnout that isn’t accompanied by a similar surge among non-whites. They are, in essence, predicting the Republicans will be much better at mobilizing their base than Democrats will be at doing the same AND that they will do so in ways that are unprecedented in our modern hyper-partisan era.

That might end up happening, but this far out there’s no reason whatsoever to think it will. If we were a few weeks from the election AND there was real evidence to suggest this would happen, then these findings would be meaningful. Not only is there no evidence yet that this will happen, this far out it’s not even yet possible for there to be reliable evidence suggesting this will happen. So when they say it is “guess work,” they REALLY mean it, which is why they and all other pollsters should be ignored until September.

But hey…since we’re guessing…I’ve made the prediction before and I stand by it: non-white participation as a percentage of the vote will hit roughly 1/3 of the voting population, up from the approximately 30% we hit in 2012. My best guess has been and will remain precisely the opposite of Q’s. I’m not going to change, but they might! Either way, we’ll have ourselves an answer on Election Day.

Much more here from Vox.com

UPDATE: Just in from Marist, polls that show a totally different race. Why? Haven’t seen the internals yet, but when they are available I bet they show a much less white electorate than Q is predicting. Elections in a partisan age are about turnout, not persuasion. One of the campaign’s is building a turnout machine, the other seems to want to rely totally on free media. Will be interesting to see how it all goes down…

Advertisements

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s