The Washington Post and ABC News released poll numbers on the 2016 presidential primary and general election on Thursday morning. They were met with the political Internet's ever-endearing mix of snark and outrage.
This tweet, from longtime Democratic political hand Guy Cecil, was our favorite of the morning. (We mean this without snark. Guy's tweet made us laugh out loud and was not indicative of the ad pollinem -- is that a thing? -- attacks on Twitter more generally.)
I love that old political adage: the only poll that counts is the one 3 years before the election.
— Guy Cecil (@guycecil) January 30, 2014
Guy is, of course, right. If polling done at this point in the 2008 race was right, we would have had a general election fight between Hillary Clinton and Rudy Giuliani. And, we didn't.
But, there's also two false assumptions made in much of the criticism leveled at the idea of polling this far out from an election. (There are 701 days between today and Jan. 1, 2016.) The first is that the Post and ABC (and all organizations that conduct early polling) are somehow complicit in the all-politics-all-the-time climate in Washington these days. The second is that these numbers are meant to be predictive, rather than a snapshot in time.
Let's tackle the first assumption first. The idea that polling on 2016 somehow makes candidates interested in the race begin to make moves is entirely backward. Look at the number of trips people like Kentucky Sen. Rand Paul, Texas Sen. Ted Cruz, Maryland Gov. Martin O'Malley and many, many others have made into early primary and caucus states like Iowa and New Hampshire already. Look at the remarkable campaign-in-waiting -- Ready for Hillary, Priorities USA -- that is already built for Hillary Clinton. Look at the wink-wink-nudge-nudge statements the candidates themselves issue about their interest in the race.
The conclusion of all of it? People who want to run for president begin putting the pieces -- staff, donors, early state organizations -- in place years before a single vote is cast or before a single poll is conducted about the race. Polling doesn't create the race. Candidates create the race.
Now for the assumption that simply by the act of conducting a poll about a far-off race you are predicting that the poll results MUST match the ultimate outcome. Talk to any pollster -- non-partisan, Democratic or Republican -- and they will tell you that polls are by their very nature snapshots in time, not predictors of outcome. (Modeling, the sort of stuff Nate Silver does, is, on the other hand, meant to be predictive.) And, if you kept talking to that pollster they would tell you that the head to head number in any poll -- especially one conducted 700 days (or so) before an election -- is probably the least important number in the survey. The perceptions that people have about the candidates' character attributes/strengths/weaknesses is far more telling as an indicator of the shape of the race to come than who is ahead in the horserace question.
In short: You need to take the WaPo-ABC poll -- and the dozens (or maybe even hundreds) that have come before it and will definitely come after it -- for what they are: Early measures of name identification and voter perceptions about the candidates.
It's easy to blame polling -- and the media organizations that conduct them -- for all the bad we have in our politics. And it's easy to mock early polling as trying to put certainty where none is possible. Easy and wrong.