Polling and the Midterms

Stolen in its entirety from Kos's DemFromCT. Lots of polling-related issues we should think about. The Hotline article it cites is worth a gander, too.

I seem to be among the very few who will not be at all surprised if the Republicans hold on to both the House and the Senate, albeit with slimmer majorities. This touches upon some of the reasons.

What Makes Republicans Confident?

by DemFromCT

Sat Oct 28, 2006 at 05:15:38 AM PDT

Winning. Oh, sure, there are details behind that, and other reasons we should and will review. But had it not been for 2004, no one would be paying attention to what Rove thinks. The flip side is that what makes Democrats incapable of overconfidence is losing (despite a record GOTV in Ohio in 2004). But we talk about Democrats often enough... let's talk about Republicans.

In this Hotline post by Marc Ambinder, Seven Reasons Why Karl Rove Is Optimistic, we look at the major factors. Of course, the most important reason is that Rove can't afford not to show optimism because the troops, already shaky, will be demoralized and stay home. But even 'optimism' means losing.

Let's define our term, first. "Optimism" doesn't mean that these Republicans are convinced that they'll pick up seats. The White House knows that its majorities in both chambers will be reduced. Optimism also doesn't imply that these Republicans are blind to the probability of a Dem House takeover and the possibility of a Dem Senate takeover.

What optimism means is that these Republicans believe that there are enough reasons to believe that Republicans can hang on to enough seats in the House and enough in the Senate to barely miss the guillotine.

Some of those reasons include

• GOP internal polls
With more money available, there's more polling. this helps Mehlman, et al direct Cheney to Idaho and other weak third tier areas that Democrats (no D polling) may not realize are weak. Follow the monay, follow the surrogates, hope it's not a fake like Bush to CA in 2000, or GOP spending in Ohio for DeWine.
• The GOP GOTV


There are about 20 additional races where the D candidate either leads the R candidate WITHIN the margin, trades leads with the Republican, or occasionally leads outside the margin of error. The Rove Optimist believes that the national Republican turnout effort - the 72 Hour Program - can add one to two percent to the margins of Republicans. So if these Republicans can stay within the margin of error - within two points - of Democrats until Election Day, there's a chance that Republicans can eek out victories in 70 percent of those contested races. Many of these races are located in congressional districts won by President Bush. Many involve incumbents who have had time to develop party-independent personas. Many represent districts drawn especially to preserve their seats.

• The mantra
What happens during the last week of the election matters as much as what happened during the last month; what happens during the last three days matters as much as the last week. Republicans might catch a break from exogenous events; they might win news cycles in critical areas.

Included in the discussion is the idea that the core can always be mobilized. Dispirited or not, they'll show. In this interesting post at RCP, John McIntyre lays that idea out more clearly.

In some ways there are two very different ways to look at what is going to happen on election day.

1) Republicans are in big trouble. The generic ballot shows a huge lead for Democrats (over 15%) with fewer than 10 days until the election. Republicans in contested races are either trailing or polling in the mid-40's, and given the national mood toward the GOP as seen in the generic ballot, it is reasonable to assume that these races will break for the Democrats. With the close races tipping the Democrats way they are poised for substantial pickups in the House of 25 seats or more and perhaps the six seats needed for a majority in the Senate.

2)The generic ballot is problematic and is over sampling Democrats, pushing the raw numbers higher for the Dems than they should be. Trying to use the generic ballot to predict who will then win x, y and z house races is a jump that can't be made soundly. In 2004 the voter turnout was 60% of eligible voters. In 2002 and 1998 in the two previous midterms it was 40%. What if a significant number of that 15%-20% who aren't going to show up at the polls this year come from soft voters in the middle? These are the exact group of voters that are helping drive the big polling numbers for Democrats. What if they don't show up in these contested races at the same proportion they are representing in many of these polls? Following this line of thinking, it is possible the bulk of the races that the polls now say are close will actually go to the GOP because the pollsters aren't sampling a representative field of who will actually vote in the contested races.

Simplifying things dramatically, the first view is essentially the one taken by Charlie Cook, and it's why he is out forecasting a 20-35 seat pickup for the Dems in the House and a very good shot for them to take the Senate. The second view is the one taken by Karl Rove, which is why he believes the GOP will hold both chambers, losing less than 15 seats in the House and 3-4 in the Senate.

Then again, there are those that believe that while only a rough guide, that generic ballot means a lot. Frank Newport has studied the Gallup generic ballot and is a believer in Gallup's likely voter model, at least for turnout. (When the interview was done, there was a tied generic LV, which has since skewed D by 13%). Here's another group that believes (click the link to see where the Figure comes from):

Although the Democrats hold a large advantage in generic ballot polls, there has been considerable uncertainty regarding whether the Democrats would win the most House Seats. Doubts are often expressed about the accuracy of the generic ballot polls. How district lines are drawn raises further doubts about whether the Democrats could win a sufficient majority of the vote to win the most seats. We estimate how the generic ballot "vote" translates into the actual national vote for Congress and ultimately into the partisan division of seats in the House of Representatives. Based on current generic ballot polls, we forecast an expected Democratic gain of 32 seats with Democratic control (a gain of 15 seats or more) a near certainty.

In the end, someone's right and someone's wrong. If the Republicans retain the House, there are going to be a good number of embarrassed professional and academic prognosticators. OTOH, whereas most of the reasons to believe there's a strong Democratic wave are based on published data, the reasons for Republican optimism are based on conjecture, anectodes, head games and fear.

My guess is that there's less "there" there for the GOP for the simple reason that this election is about the moderates and not the base, and the moderates have had enough.If Rove's polls were so great, he'd be leaking them like a seive. I'll put my money on the wave, and watch Rove and Mehlman try to make history.