Recall Turnout Matters. Where Voters Turn Out Matters More

Here we go. These are likely the final independent polling numbers we'll see for the gubernatorial recall election in Wisconsin tomorrow.


[Reuters:]

Two public opinion polls released on Sunday show Wisconsin Republican Governor Scott Walker with a lead of three and six percentage points two days before the election to recall him because of a new law reducing the power of public sector unions.

Public Policy Polling, a Democratic polling firm, said Walker was leading 50 percent to 47 percent over Democratic challenger Milwaukee Mayor Tom Barrett in its final survey. Angus Reid polling had Walker ahead 53 percent to 47 percent. Both findings were within the margin of error so the results could be even tighter.

The PPP margin of three percentage points compared with a five-point Walker lead in their survey three weeks ago.


"If the folks who turn out on Tuesday actually matched the 2008 electorate, Barrett would be ahead of Walker by a 50-49 margin," says PPP's Tom Jensen. "It's cliche but this is a race that really is going to completely come down to turnout... This is a close race, closer than it was a couple weeks ago. Scott Walker's still the favorite but Barrett's prospects for an upset look better than they have in a long time."

Meanwhile, Democratic internal numbers look much closer. Prior to the release of the much-anticipated PPP findings, dems announced their numbers showed about as close to an actual tie as you're ever going to see: 47.8% Walker to 47.7% Barrett.

Throughout this race, Democratic numbers have been much better for Barrett than independent ones. It's easy to write off dem internals as spin, but the truth is that these are the numbers they're relying on to track the progress of their work. If the numbers are bad, then they're no help at all. Democrats have every incentive to create the most accurate and predictive polling results out there. So why the difference between their numbers and independent numbers?

To explain that, allow me to tell you a brief story. In 1998, Russ Feingold was in a tight reelection race against Republican Mark Neumann, who was favored to win. Come election day, Russ eked out a 2% margin of victory over Neumann, thanks to massive turnout in Milwaukee and Dane Counties. "It looks like Dane County's chosen their senator," Republican then-Gov. Tommy Thompson said of the result. It's not a lesson that state Democrats are likely to have forgotten.

Democrats and unions have been targeting Milwaukee and Dane especially. That's where the big organization advantage they have comes from -- huge crowds of blue state voters, all bundled together for easy access. Phone banking in rural areas is one thing, but door-to-door canvassing in urban centers is another. Recall supporters have had much more face-to-face talks with voters and aggressive organizing in Milwaukee and Dane are the explanation for that.

So, if you knew turnout was going to be higher in the areas you've been targeting, wouldn't you want your polling sample to reflect that? If dem polling disagrees with indie polling because the Democrats are oversampling Milwaukee and Dane, then I'd put my money on dem polling as the most predictive. Behind the scenes talk right now is turnout in Dane of 70-80% (based on early voting numbers) and I don't know what they're expecting in Milwaukee -- but I do know that the organizational effort there is of heroic proportions. They'll have vans running all day tomorrow, taking voters to the polls.

In the end, the cliche is true -- what matters here is turnout. But just as true is that it matters a great deal where that turnout is heaviest.

-Wisco

[image credit: adapted from a photo by ra_hurd, via Flickr]