Thursday, April 3, 2008

More On The "Latest" Schaffer-Udall Poll

The "latest" poll was tackled by Ben earlier this morning, but although it was released on March 28, the survey itself was conducted between March 6-9, nearly a month ago, and came before the most recent Rasmussen tracking poll that had Mark Udall up 46-43, within the margin of error, just like every other poll to this point:

As Ben points out, there is too little information here, hidden behind a fee wall--just like many of the other polls.

Here is the methodology blurb
:
Methodology: McLaughlin & Associates conducted statewide surveys among general election voters in Colorado (n=400), Maine (n=400) and Minnesota (n=500) between March 6th and 9th, 2008. All interviews were conducted by professional interviewers via telephone. Interview selection was at random within predetermined election units. These units were structured to statistically correlate with actual voter distributions in statewide general elections. The accuracy of the samples of 400 likely general election voters is within +/- 4.9% at a 95% confidence interval. The accuracy of the sample of 500 likely general election voters is within +/- 4.5% at a 95% confidence interval.
While their stated methodology may be sound, it is hard to put much credence into a poll with 25% undecideds, and no breakdown on which way those undecideds lean.

Rocky Mountain Right has a good tracking poll (blogroll-worthy, if you have not yet checked them out), but places this poll after the last Rasmussen poll, instead of before it, in correct chronology (not their fault, as the McLaughlin poll was released after the Rasmussen poll).

Our SvU tracking polls put the McLaughlin poll in its proper context (click to enlarge):


With trendlines:


Only the McLaughlin poll shows anything but a tight race, consistently within the margin of error.

But the national media and Beltway politicos have always expected this to be a Udall pickup for the Dems, and have dismissed the other polls showing a toss-up in favor of one outlier:
Given his lackluster performance in the 2004 Republican Senate primary, many expected that former Rep. Bob Schaffer's bid to replace retiring Senator Wayne Allard would fall similarly flat. Schaffer is assumed to be too far right for the increasingly-Democratic Colorado, and his opponent, Democratic Rep. Mark Udall, has a great name and a fat bank account.

But in poll after poll, Schaffer has trailed Udall by exceedingly small margins, virtually always within the margin of error. In December, Schaffer trailed by two points. In October, the gap was just one point. Was labeling Udall as a Boulder liberal finding success? Is Colorado still a red state? Or is Schaffer manager Dick Wadhams, a former top aide to Virginia Senator George Allen, the next Karl Rove?
. . .
Before Democrats get too thrilled and claim they are guaranteed to pick up the seat, they might want to wait for a few more surveys to come out. Schaffer has a talented political team, led by Wadhams, and Udall remains well under 50%. But a twelve-point lead is what most Beltway politicos expected to see, and the McLaughlin survey proves that Schaffer, who has trailed in every poll Politics Nation has seen on the race, has a ways to go to climb out of a hole.
It is surprising to see the RealClearPolitics blogger parrot the typical MSM-Beltway consensus, when every other poll has shown a maddeningly (for Dems) tight race. The Cook Political Report's most recent analysis (March 20) still has this race as a competitive toss-up, with either party having a good chance of winning.

And as for polls in general--the polls said Hillary Clinton would be the Democrat's nominee and that John McCain was on death's door (politically), just a few months ago.

So much for polls. There will be plenty more over the next seven months, and as the saying goes, anything can happen.

1 comment:

Ben DeGrow said...

Great analysis, El Presidente. Thanks for providing the necessary context.