Numbers Don’t Back Norton Camp’s Attempt to Disparage SurveyUSA

Colorado’s big political buzz since Sunday has revolved around the Denver Post / 9News / SurveyUSA polling on the gubernatorial and U.S. Senate races. The governor matchups pretty much reflect Rasmussen’s recently released results, but the latest poll also puts Scott McInnis at a 2-to-1 advantage over fellow Republican Dan Maes in head-to-head support for the primary.

On this point PPC blogger PerlStalker is correct about Maes’ chances:

That’s a lot of ground to make up before the August primary. I doubt that he can do it with so little time left.

Don Johnson and others have been highlighting the most newsworthy result: Ken Buck’s 53-37 advantage over Jane Norton in the Republican showdown for the U.S. Senate contest. The response from the Norton camp at the changing fortunes? Josh Penry attacked the messenger in an email memo to supporters:

It’s safe to say that Howard Dean was more than a little surprised after being trounced in the 2004 Iowa Caucuses by eventual Democratic nominee John Kerry. You’ll recall that was the night that Dean had a now-legendary meltdown on primetime national TV.

But Howard Dean wasn’t the only one surprised by his trouncing in Iowa: so was the national polling firm SurveyUSA. That’s because SurveyUSA predicted that Dean would win the Iowa Caucuses. SurveyUSA missed the mark by a wide margin, overestimating Dean’s support by six percentage points and underestimating Kerry’s by nearly seventeen points.

This is but one of many examples where SurveyUSA has had it wrong – and badly.

What exactly are those examples? And why did Penry choose an isolated instance from 2004 to disparage the pollster showing his candidate 16 points behind? Because it’s an outlier.

Political data uber-geek Nate Silver released his latest comprehensive pollster ratings a couple weeks ago: over 4,700 public polls of Presidential, U.S. Senate, U.S. House and governor’s races have been released since 1998. SurveyUSA conducted 634 of those polls, more than any other individual firm. Of the 63 firms that have done 10 or more polls, SurveyUSA is rated as the third most accurate.

Kind of takes the edge off the spin. The news for Norton is not-so-good.

With this latest information, it looks like the statement I made last week after the release of the Magellan Strategies poll was, if anything, an understatement:

…[I]f Ken Buck can continue his momentum by reporting a significant increase in fundraising through the end of this month, all signs will show him as the come-from-behind prohibitive favorite for the last month of the primary campaign.

Makes me wonder… What if not so much online noise had been made nine and a half months ago when Jane Norton prepared to enter the race, and Ken Buck was urged to leave to effectively clear the field? It’s easy to see how the race might be in a far different place today.

Ben DeGrow is a Contributing Author at People’s Press Collective, Your Source for Colorado Politics.