September 29, 2010

The Loudest Minute

Hatched by Dafydd

A Lizardian maxim is that in elections, the last minute is the loudest minute. That is, last-minute roaring surges are the norm, rather than the exception.

In particular, in a race in which incumbent (or much better-known) nominees are running against little-known challengers, the same pattern typically emerges: If the incumbent cannot break through the 50% ceiling by a significant margin -- say consistently averaging 54% or better -- then in the last month, undecideds will generally break for the challenger.

The reason is simple. Voters have waited and waited for the incumbent to give them a reason to reelect him; but if he cannot make the sale by October, he likely won't do it at all: He's so well known already that he has no surprises left. By contrast, the lesser known candidate still has a great surprise-potential; and as voters become impatient for a reason to reelect the incumbent, they take a longer, friendlier look at the challenger.

I dub this the "stale incumbent" factor.

Of course, the challenger's "surprise" could also be something terribly negative, leading to a surge for the incumbent. It doesn't usually happen; if such a deal-killer existed, it would almost certainly have already come out earlier. In Delaware, Democrat Chris Coons is wasting no time bringing out all the nutter utterances of Tea Partier and Republican nominee Christine O'Donnell; he's not waiting for the last week, he's been pounding on her since the moment she won the nomination!

In most (two-person) elections where (a) an incumbent is not noticibly above 50% by October, and (b) the challenger or his party has a tailwind, the challenger will win -- even if he is running somewhat behind the incumbent right up through the last poll.

I think we're seeing that dynamic right now in Washington state, based upon Paul Mirengoff's reporting in Power Line. He begins:

I've been a bit disappointed by the polls I've seen of the Senate race in Washington State. Dino Rossi, an attractive Republican challenger who very nearly was elected Governor in 2004, has been consistently behind incumbent Patty Murray. Murray's average lead, according to Real Clear Politics is 5.3 percentage points.

The pattern occurs in many, many races this fall:

  • Sen. Patty Murray (D-WA, 95%) has led challenger Dino Rossi in the U.S. Senate race there in sixteen out of 24 polls since January (one was a tie), and this month she has led by 5-9 points. But she has never managed to get to 54% the entire year -- not even once. In fact, in all the polls in which she has been ahead, she has only been above 50% four times. Most of the time, Murray has been mired in the 40s, even when she led Rossi.
  • Similarly in California, Barbara Boxer (D-CA, 100%) is limping along in the mid-to-high 40s in most polls (the CNN/Time and the LA Times polls have her above 50%, but no others). Republican nominee Carly Fiorina is riding 6.8% behind... which would be a likely loss, if the split were Boxer 53 to Fiorina 46, instead of Boxer 49.5 to Fiorina 42.5.
  • And in Nevada, Senate Majority Leader Harry "Pinky" Reid (D-Caesar's Palace, 95%) has never been above 50% in any poll this year, and he only touched 50% one time. (Republican nominee Sharron Angle hit 50% three times, and once she even nosed above to 51%.)

These are classic examples of incuments who simply cannot close the deal. Given that, I expect that starting in October, Rossi, Fiorina, and Angle will "unexpectedly" surge forward by at least the amount of the undecided respondents, which ranges from three to eight percent.

As Paul reports, such a jump seems to have begun a bit early:

But two very recent polls suggest a closer race. A Survey USA poll released on September 23 shows Murray leading by a margin of 50-48. And a Fox News/Pulse Opinion Research survey of 1,000 likely voters, taken on September 25, has Murray leading by only 48-47. Both "leads" are within the margin of error.

That puts the race dead even within the margin of error. But there is another factor at play here: When voter sentiment for Republicans is rising, pollsters typically underestimate it; but when voter sentiment for Democrats is rising, pollsters typically overestimate it. I call this the "I can't believe it's not butter!" factor: Pollsters see surging Republicans -- and they just can't believe it.

They conclude they must have accidentally "oversampled" GOP respondents, so they "correct" their mistake by reweighting the poll, reducing the number of likely GOP voters by enough of a percent that the final results show... well, whatever number seems more "reasonable" to the pollster.

They're not deliberately cheating; they're just dead certain that Republicans cannot possibly be doing that well, and they don't want to report such an obvious "outlier" and be embarassed on election day. And hey, none of their friends are voting for the Republican; how well could the GOP possibly be doing?

Contrariwise, when a typical pollster sees Democrats rising, it's just what he's been expecting all along. He gets excited and again monkeys with the weighting of likely voters, giving the Dems the boost that he believes, to the bottom of their soles, is what's really happening.

This pro-Democrat, anti-GOP fudge factor typically amounts to at least 2% and sometimes as high as 5% - 6%. The "I can't believe it's not butter!" factor and the "stale incumbent" factor are additive: If Republicans are ascendent in an election cycle, most races in which the final poll is 50-50 or even 52-48 for the Democrat -- will "unexpectedly" break for the Republican when the actual vote is counted, prompting Democrats to file a lawsuit and try to sue their way into office. (And assuming Democratic voter fraud is not so rampant that it overcomes all obstacles.)

That is why we predict that the GOP will in fact win all the so-called "toss-up" races in November and may even pick up one or two Democratic leaners or likelies. And that is why we're nor surprised to see Dino Rossi suddenly neck and neck with "Patsy" Murray.

Nor will Bill Clinton campaigning for Murray turn the race around; nobody in Washington state doubts that Murray is a good liberal or that Clinton supports Democrats over Republicans... so of what value is a campaign turn by the popular former president?

In general, campaigning by more senior politicians only has a significant impact when the lucky recipient of such help is himself little known; in that case, support from a better-known and popular figure can reassure the base. For example, Sarah Palin campaigning for her virtually unknown "Mama Grizzlys" is extremely helpful. But Bill Clinton stumping for embattled incumbents -- not so much.

I'm pretty sure Rossi will win on November 2nd... just as I'm pretty sure Carly Fiorina will beat Barbara Boxer in the Golden State and Sharron Angle will defeat "Pinky" Reid in the Silver State. None of the incumbents seems capable of closing the sale, despite -- or because of -- many years in office.

Hatched by Dafydd on this day, September 29, 2010, at the time of 5:21 PM

Trackback Pings

TrackBack URL for this hissing: http://biglizards.net/mt3.36/earendiltrack.cgi/4611

Listed below are links to weblogs that reference The Loudest Minute:

» The "Loudest Minute" Begins in Florida as Crist Sinks from Big Lizards
In yesterday's post, the Loudest Minute, I offered a few examples of longtime incumbents in Washington state, California, and Nevada who couldn't break through the 49%-52% ceiling of support; I opined thus: These are classic examples of incuments who s... [Read More]

Tracked on September 30, 2010 1:22 PM

Comments

The following hissed in response by: MikeR

I'd be interested in hearing your comments on the recent series of posts in fivethirtyeight.com (Nate Silver) on the accuracy of polls.

The above hissed in response by: MikeR [TypeKey Profile Page] at October 3, 2010 7:26 PM

Post a comment

Thanks for hissing in, . Now you can slither in with a comment, o wise. (sign out)

(If you haven't hissed a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Hang loose; don't shed your skin!)


Remember me unto the end of days?


© 2005-2009 by Dafydd ab Hugh - All Rights Reserved