« Colorado Pics of the Day: Aspen Cascade | Main | Guest Blog: Divine Strake and the Rebirth of Democracy »
Thursday, September 21, 2006
Even Survey USA Shows Madrid-Wilson Race Neck and Neck
Local KOB-TV News just announced the results of a recent poll by Survey USA (SUSA) that shows Republican incumbent Rep. Heather Wilson at 51%, Dem challenger Patricia Madrid at 46% and Undecided at 3% in the congressional race in NM-01. Since the poll has a reported margin of error of 4.5%, this means the results have Wilson only a hair ahead of Madrid. Still, numbers like these might normally make me a little nervous. Not this time.
SUSA has a generally poor reputation because its methodology isn't viewed as scientifically sound by many serious analysts. It uses what's known as "computer response automated calls." Auto-dialers make calls until an arbitrary number have been answered. This method isn't perceived as statistically reliable by a number of observers. Its samples are also known to often contain too many Republicans. It's a pattern. In fact, SUSA's reputation is shaky enough that the Associated Press, the National Journal and other top-of-the-line news outlets won’t even report SurveyUSA poll results.
In addition, all the other polls in this race have been consistent in showing the race as a tie. The results of polling by Constituent Dynamics, Research & Polling and Lake Research Partners all show a virtual dead heat between Madrid and Wilson. This one is very probably what's known as an outlier -- a poll that falls outside the margin of credibility.
The good news is that, even with all of these problems, Madrid is shown at only 5 points back and just about within the margin of error -- a strong position for a challenger at this stage of the race. We can (and will) win this thing!
Of course the only poll that really counts is the one that happens on November 7th. Are you registered to vote? Turnout will make the difference in this neck and neck race. If we get a strong turnout by Dems and Independents, Madrid will win. You know what to do.
September 21, 2006 at 06:45 PM in Candidates & Races | Permalink
Comments
An by the National Council on Public Polls, just released, found SurveyUSA outperformed all but one traditional pollster in the 2004 presidential election.
An independent analysis completed in November 2004 found that SurveyUSA outperformed all other pollsters in 2004.
SurveyUSA’s own analysis found that SurveyUSA outperformed all other pollsters in 2004, by any measure. (caution: large Excel file, takes time to open)
SurveyUSA, the nation’s most active election pollster, is the only opinion pollster to publish an exhaustive compendium of all election contests it has ever polled, compared to all other pollsters. Our record speaks for itself.
Posted by: | Sep 21, 2006 8:53:42 PM
"Still, numbers like these might normally make me a little nervous. Not this time." I sure hope your right, however the Madrid campaign really needs to step it up a couple of notches. It seems to me that their TV ad's are only beating the connection to Bu$h and tainted money drum, although important, there is more available, Wilson's voting record is a gold mine of usable material. Veterans issues would be a good one: Project Vote Smart shows Rep Wilson supported the interests of the Disabled American Veterans 40 percent in 2005. Rep Wilson supported the interests of the Disabled American Veterans 0 percent in 2004. Yet, I keep seeing Veterans for Heather yard sign all over the place, apparently those Vets are not getting the truth on her.
Posted by: VP | Sep 21, 2006 8:54:39 PM
SurveyUSA cites studies from SurveyUSA to conclude that SurveyUSA is correct. Questionable. What do independent analysts say?
The Washington Post's Chris Cilliza:
Since I spend so much time parsing the polls, I have formed opinions about what makes a poll worthy of attention and what doesn't. Ever since my time at Roll Call newspaper, I have generally viewed telephone surveys that employ live interviewers as the most reliable. I'm much more skeptical about polling that uses automated interviewers -- Survey USA and Rasmussen Reports being the two firms best known for using this approach -- so I largely avoid citing these surveys on The Fix.
More on auto-dial polls:
https://blog.washingtonpost.com/thefix/2006/07/parsing_the_polls_of_auto_dial.html
Posted by: Not SurveyUSA | Sep 21, 2006 9:51:55 PM
SurveyUSA did not site its own studies. SurveyUSA cited independent research and then linked to its own side-by-side comparisons of our polls and competing polls.
You can find still more independent analysis of SurveyUSA's work in this article by pollster Mark Blumenthal and in this paper by University of Oregon professor Joel Bloom.
If you have evidence showing that any of our data is incorrect -- that we've mis-reported the final outcome of an election, or reported a competitor's penultimate poll and not actually their final poll in a given contest -- please contact us to let us know, and we'll happily update our data. Otherwise, I suggest you (and Mr. Cilliza) spend some time examining the data so that you can make accurate statements in the future.
Posted by: < | Sep 22, 2006 6:55:09 AM
Guess what? The tobacco industry used to be able to find and tout "independent studies" that said smoking wasn't harmful. Industry touts itself so that it can make money.
Now, let's look at some facts.
1) The Washington Post (above) said that they are "skeptical" about SUSA's methodology and as a result they "largely avoid citing these surveys."
2) The Associated Press and the National Journal also refuse to report on SUSA polls because they use questionable methodology.
3) University of Michigan Professor Michael Traugott on SUSA: "There is no sound theoretical basis for the way in which these surveys are conducted."
4) Spokesman for Texas Republican Gov. Rick Perry: Perry spokesman Robert Black dismissed the Survey USA report as "bogus," claiming its methodology could not be trusted. It's not just a partisan thing, Democrat Chris Bell has agreed with the Perry spokesman regarding SUSA's accuracy.
5) SUSA is clearly an outlier in the NM-01 race. Research & Polling (very good track record polling NM), Constituent Dynamics and Lake Research (Madrid internals) all show a much closer race. Thusfar, we haven't seen a Wilson internal - the good bet is because the results aren't very favorable to her.
I've got more, but, I don't have more time right now to post. Sorry!
Posted by: Not Survey USA | Sep 22, 2006 7:32:55 AM
Geez, these SUSA people come out of the woodwork immediately. Might be because they aren't taken seriously by most knowledgeable sources so they have to be out defending themselves quite often.
From what I know, local TV news outfits often use SUSA because they are cheap. You get what you pay for. I'll take Brian Sanderoff any day over SUSA. He has an excellent reputation and has local political knowledge as well.
Even if someone would take SUSA as a good outfit, the results are within the margin of error on the poll again making the race just about even.
This is Madrid's race to win or lose. It is up to Democrats to get out in droves to vote or else be quiet as Wilson continues another term of playing yes woman to Bush and his terrible agenda.
Posted by: Not a SUSA fan | Sep 22, 2006 9:05:22 AM
