Wednesday, October 31, 2012

The Electoral Dance

Things are not always what they seem.

I've been watching elections since about 1968, and here is what I generally see:  relative to the final results in November, the polls throughout the summer and into early fall show Democrats ahead of where they end up, and Republicans behind where they end up.  Then, sometime around about two weeks before election day, there is some sort of an "adjustment", and behold!  By election day the polls are much closer to the final results.

Classic example of this was the 1980 election.  Carter had been running five points ahead of Reagan all summer long, clear up until their debate.  After the debate, the polls had Reagan pulling up to more or less even, and the morning of the election, the headline on the San Francisco Chronicle's front page (I was stationed in California in 1980) proclaimed, "Too Close To Call!"  Well, nobody told the Reagan voters.  Reagan won by a significant margin, about 51% of the vote in a three-way race (Carter got about 41%, and renegade Republican John Anderson garnered most of what was left over).

Even the elections that Republicans have lost usually turn out to have been much closer than the pollsters had prognosticated.  In '96, there were many polls showing Clinton with huge double-digit leads.  But the "adjustments" appeared at the last minute, followed by a relatively miserable showing for Mr. Clinton -- in neither of his two elections did he even break the 50% mark, but he did beat Dole by nine or ten points.

Even in off-year Congressional elections, the strength of the GOP on election day often appears to catch the media flatfooted.  The Republican takeover in '94 hit them like a runaway train; I don't know whether they were surprised, but they certainly acted surprised.  First time since the Eisenhower administration that both houses of Congress went Republican, and nobody had predicted it.

More recently, in 2010, the media acted like it was "shocking" when some people predicted a Republican congressional landslide.  But election day came, and there it was.

There's a pattern here.  So the next question is, are the polls reliable and the electorate just routinely shifts toward the Republicans as election day nears?  Or are the polls stacked to favor Democrats for as long as they can, until they must either adjust or lose credibility?

The latter seems more likely.

Now, is it some vast conspiracy to demoralize the Republican base to try to lower voter turnout?  Why would professional pollsters risk their reputations to do this?

Pollsters are trying to make money, like the rest of us, and their behavior just might be a simple, rational market response aimed at pleasing their customers.  Most news-media outlets are run by liberals, and liberals want to hear good news about Democrats.  So if you're running a polling company and a major network hires you to produce a poll, and you know the people you're dealing with prefer Democrats, you have a very clear incentive to give them what they want.  That is, until it comes into conflict with another stronger, countervailing incentive -- namely, keeping your reputation as a polling outfit intact.

Either way, it's easy to get a poll to say what you want it to say.  Here's a link to an illuminating article by PJMedia's Charles Martin, explaining how pollsters can achieve whatever results they're looking for and make it look credible:

Most pollsters this year have been (until recently) polling registered voters as opposed to likely voters (which tends to skew towards Democrats); and for the most part they have been oversampling Democrats.  In one Quinnipiac poll, for example, it was claimed Obama had an eight-point lead in Ohio -- but in their sampling, Quinnipiac gave Democrats a nine-point edge.  Ohio is a swing state; does it seem likely that nine percent more Democrats will turn out than Republicans?  You do the math -- in fact, you'd better do the math, as Quinnipiac doesn't seem very interested in doing it for you.

I'm beginning to suspect that the news media understands all this.  Why?  For the simple reason that, all along, Obama has been campaigning -- he's been behaving -- like he's behind.  The Obama campaign has been nasty, surly, and issue-free, focusing its attentions on assassinating Romney's character and trotting out red herrings about how rich he is.  That's not how someone who believes he's ahead in the race runs a campaign.  It shows a lack of confidence.

This tells me Obama's internal polling numbers have been closer to the real mark all along than the polls that get trotted for general consumption.  Obama behaves like he's behind because he is behind, and knows it.

And I don't believe all this has gone unnoticed by veteran reporters.

The polls now have "adjusted" -- of course, the election is next week -- and are giving Romney anywhere from a dead even to five-point edge, but many are still saying Obama still has the advantage in the Electoral College tally.

I think that, where they're saying Romney is now, is where he was all along -- but he is now strengthening his lead.

Right now, I'm expecting a solid Romney win, perhaps even a landslide.  Let's see if I'm right.

But here's another prediction:  if the election is within one state's electoral votes, and that state's election is within two or three points, the Democrats will airdrop a thousand lawyers into that state and try to steal the election.

Wait and see.