Joho the Blog » [2b2k] What do we learn from our failure to believe the polls?
EverydayChaos
Everyday Chaos
Too Big to Know
Too Big to Know
Cluetrain 10th Anniversary edition
Cluetrain 10th Anniversary
Everything Is Miscellaneous
Everything Is Miscellaneous
Small Pieces cover
Small Pieces Loosely Joined
Cluetrain cover
Cluetrain Manifesto
My face
Speaker info
Who am I? (Blog Disclosure Form) Copy this link as RSS address Atom Feed

[2b2k] What do we learn from our failure to believe the polls?

There’s lots being written about why the Republicans were so wrong in their expectations about this week’s election. They had the same data as the rest of us, yet they apparently deeply believed they were going to win. I think it’s a fascinating question. But I want to put it to different use.

The left-wing subtext about the Republican leadership’s failure to interpret the data is that it’s comeuppance for their failure to believe in science or facts. But that almost surely is a misreading. The Republicans thought they had factual grounds for disbelieving the polls. The polls, they thought, were bad data that over-counted Democrats. The Republicans thus applied an unskewing algorithm in order to correct them. Thus, the Republicans weren’t pooh-poohing the importance of facts. They were being good scientists, cleaning up the data. Now, of course their assumptions about the skewing of the data were wrong, and there simply has to be an element of wish-fulfillment (and thus reality denial) in their belief that the polls were skewed. But, their arguments were based on what they thought was a fact about a problem with the data. They were being data-based. They just did a crappy job of it.

So what do we conclude? First, I think it’s important to recognize that it wasn’t just the Republicans who looked the data in the face and drew entirely wrong conclusions. Over and over the mainstream media told us that this race was close, that it was a toss-up. But it wasn’t. Yes, the popular vote was close, although not as close as we’d been led to believe. But the outcome of the race wasn’t a toss-up, wasn’t 50-50, wasn’t close. Obama won the race decisively and not very long after the last mainland polls closed…just as the data said he would. Not only was Nate Silver right, his record, his methodology, and the transparency of his methodology were good reasons for thinking he would be right. Yet, the mainstream media looked at the data and came to the wrong conclusion. It seems likely that they did so because they didn’t want to look like they were shilling for Obama and because they wanted to keep us attached to the TV for the sake of their ratings and ad revenues.

I think the media’s failure to draw the right and true conclusions from the data is a better example of a non-factual dodge around inconvenient truths than is the Republicans’ swerve.

Put the two failures together, and I think this is an example of the the inability of facts and data to drive us to agreement. Our temptation might be to look at both of these as fixable aberrations. I think a more sober assessment, however, should lead us to conclude that some significant portion of us is always going to find a way to be misled by facts and data. As a matter of empirical fact, data does not drive agreement, or at least doesn’t drive it sufficiently strongly that by itself it settles issues. For one reason or another, some responsible adults are going to get it wrong.

This doesn’t mean we should give up. It certainly doesn’t lead to a relativist conclusion. It instead leads to an acceptance of the fact that we are never going to agree, even when the data is good, plentiful, and right in front of our eyes. And, yeah, that’s more than a little scary.

Previous: « || Next: »

Leave a Reply

Comments (RSS).  RSS icon