The Weekend's Numbers

+ More

On the way to Fox News's New York headquarters for the election night dress rehearsal on Saturday, I ran into a young analyst at a Democratic polling firm.

"I've never seen numbers like this," he said enthusiastically. He was citing results showing Democrats competitive in heretofore solidly Republican districts. I'm sure he hasn't seen such figures before; he hasn't been in the polling business that long. As I've often noted, we have been in an extended period of political stasis, with the electorate divided just about evenly between the two parties and the contours of political support by geographic area and demographic group remaining very much the same. It has lasted from the 1996 election, which seems to represent where people fell out after the extended budget battles between Bill Clinton and Newt Gingrich, up through the 2004 election. Now we seem to be into, or at least headed for, a different period. In the 1996-2004 period, Democrats faced a ceiling of 49 percent–that's Clinton's percentage in 1996 and the best the Democrats have done in the popular vote for the House. The Republicans also have had a low ceiling, though just a tad higher, of 51 percent–that's George W. Bush's percentage in 2004 and the best Republicans have done in the popular vote for the House.

Now the Democrats seem certain to break through their ceiling and to rise above 49 percent in the popular vote for the House. The question is just how far above and how that will break down seat by seat. Larry Sabato predicts the Democrats will gain 27 seats, giving them 230–exactly the number Republicans ended up with after 1994. Charlie Cook and Stuart Rothenberg think they may win even more. Karl Rove and Ken Mehlman are gamely predicting that Republicans will lose only 12 to 14–which would leave a Republican majority. An experienced Republican seat watcher puts the losses at 22.

Sunday morning's polls provide a glimmer of hope for Republicans. The Washington Post/ABC News poll, conducted November 1-4, has the Democrats' advantage in the generic vote for the House among likely voters at 51 to 45 percent–down from 55 to 41 percent in the poll it released October 22. And in this latest poll, Republicans are a little more apt than Democrats to pass the likely voter screen; on the preceding poll, and on many other polls this season, it was the other way around. Did John Kerry's disparagement of the soldiers in Iraq give Republicans more motivation to vote? Maybe. These numbers are a bit of evidence in that direction.

The Mason-Dixon Senate race polls released Sunday are also interesting. They've got Rick Santorum still way behind in Pennsylvania but Mike DeWine trailing only 50 to 44 percent in Ohio. They show Bob Corker way up (unrealistically, it seems to me) 50 to 38 percent in Tennessee. Claire McCaskill is ahead 47 to 46 percent in Missouri–throughout the cycle the closest race in the country. They've got Montana a 47-to-47 percent tie and Lincoln Chafee ahead 46 to 45 percent in Rhode Island–both an improvement for Republicans over almost all recent polls. In Maryland, Ben Cardin is up only 47 to 44 percent; in Virginia, Jim Webb is up 46 to 45 percent, less than his lead in other polls released last week. Incumbent Democrat Bob Menendez in New Jersey and Republican Jon Kyl in Arizona still have leads despite spending by their challengers' national parties in the past week. Most analysts have assumed that Pennsylvania, Ohio, Rhode Island, and (at least until the last week) Montana were already lost to the Republicans. The Mason-Dixon results suggest that more races may still be open than previously thought.

I'll be on Fox News on election night, analyzing the tabulated vote and providing such other thoughts as I may be called on to share. Democrats hope that it will be pretty much all over by 10 p.m., and that's still certainly possible. But it may go on till the small hours.

Poll Update

Two more polls have come in confirming the trend in the ABC/WaPo poll. First, there was Pew, which shows the Democratic advantage on the generic vote among likely voters diminishing from 50 to 39 percent two weeks ago to 47 to 43 percent November 1-4. Then came Gallup, which showed the Democratic edge declining from 54 to 41 percent to 51 to 44 percent. Gallup, because of its way of culling likely voters from registered voters, tends to show greater volatility than other polls. The movement here is, technically, not necessarily statistically significant. But it points in the same direction as ABC/WaPo and Pew. These polls portray Republicans as more motivated to vote than Democrats. And they portray the preference for Democrats generically as diminishing.

Let's look at the relationship between the generic vote and the actual voting on Election Day. The ABC/WaPo poll is helpful here, because the release shows the generic vote in the run-up to previous elections. As I have noted before, the generic-vote question is not a good indicator of how people will actually vote–and Democratic pollsters I have raised this question with agree. The numbers speak pretty loudly: Here are the WaPo's generic vote numbers (R-D) compared with the actual popular vote for the House (to the best of my recollection) going back to 1994. For some years, WaPo data are lacking.

2006 45-51 ?
2004 50-47
2002 48-48 51-47
2000 9/6 42-49 49-48
1998 43-51 49-48
1996 10/8 47-51 49-49
1994 42-47 52-45

The generic-vote question persistently underprojected Republican vote share from 1994 to 2002. It did a better job of projecting Democratic vote share but missed the important thing: Republicans trailed on the generic vote every time except 2002, when they were tied, but Republicans won the popular vote for the House every time. The current WaPo numbers look like the average of 1996 and 1998, when Republicans won the popular vote 49 to 48 percent. As you may recall, they were on the defensive in the campaign dialogue in both cycles, in 1996 because of the backlash against the highly unpopular Newt Gingrich, in 1998 because of the backlash against the impeachment of Bill Clinton. They won the popular vote, and they won most of the seats anyway.

That's not to say they necessarily will this time. Scandals and other unforced errors seem sure to cost them 10 or more seats, and a loss of 15 seats produces a Democratic majority. But there are some other interesting numbers in the WaPo poll. Was it worth fighting in Iraq? Registered voters say no by a 44-to-53 percent margin. But that's not statistically different from the margin among registered voters in the WaPo's September 2004 poll: 46 to 51 percent. And the current WaPo's likely voters split, just barely, the other way: 49 to 48 percent. Similarly, on which party is better at handling Iraq, the WaPo has likely voters at 42 to 42 percent. Not a great endorsement of the president's party. But not the stinging rebuke that so much of the MSM coverage suggests.

Last summer, I wrote that the voters had decided that the Republicans deserved to lose but had not decided that the Democrats deserved to win. Sometime in October, as we spent our two or three weeks mulling over the Mark Foley scandal, one of Charlie Cook's ace staffers said that the voters had decided that the Democrats were an acceptable risk. Now I wonder whether that was right: whether in fact voters in the past week or so have been considering whether the Democrats deserved to win. The movement of independents in these polls to somewhat smaller anti-Bush margins and the apparent greater motivation of Republicans than Democrats to get out and vote suggest voters have been mulling over that question and that the Democrats, with their pounding anti-Bush rhetoric but their absence of much in the way of positive policies, might not be passing the test. And then along comes John Kerry. Voters may want to see George W. Bush checked by his opposition. But maybe not all that much.

All this said, I have been looking at three polls, and others may come along and point the other way. Republicans are plainly on the defensive in Senate and House races, and if they lose all or almost all the close Senate races and if they forfeit as many House seats as they seem likely to, Democrats could end up with majorities in both the House and Senate. On the House side, Republicans, even while holding most of their seats that have long been recognized as seriously contested, could lose overwhelmingly Bush '04 seats where Democrats are running attractive candidates and Republicans nominated by plurality candidates with serious liabilities (Idaho District One, Nebraska District Three) or where Republicans who have never had to campaign much have been caught unawares (New York 25). Many outcomes are possible. But those possible outcomes include some that seemed unrealistically optimistic for Republicans only a few days ago.