Chron/Murray poll story has legs : Lone Star Times

Chron/Murray poll story has legs

by David Benzion · 01/18/2006 10:05 am

Serious methodological doubts have already been raised on other right-leaning local blogs regarding Professor Richard Murray’s recent "Tom DeLay poll" on behalf of the Houston Chronicle.

This story isn’t going away. Your blogger was personally present yesterday afternoon as a reporter from FOX News (national) conducted an extensive interview with Dr. David Hill of The Woodlands, who happens to be one of the most respected political pollsters in the nation. (Full Disclosure: And my former boss.)

I won’t share all of Hill’s goodies– my understanding is that you might want to tune in to "Special Report w/Brit Hume" this afternoon– but insights gleaned by yours truly include the following:

By the time you get to the small sample of just 149 "reasonably likely in the ‘real world’ (and not just according to some academic theory) to vote in the Republican primary voters" the data has a margin of error approaching plus OR MINUS 9-percentage points.
That means that if (as the Murray would claim) 41% of likely Republican primary voters questioned said they would vote for DeLay in the election were held today, the ACTUAL number might be as high as 50% (41 plus 9) or as low as 32% (41 minus 9).
Well, which is it? 50% or 32%? These sorts of numbers are obviously useless– which is why newspapers typically don’t publish political poll results based on the opinions of just 149 people.
In his "defense", the "Murray is a bumbling professor" theory (as opposed to the "Murray is a Democrat partisan" or "Murray is a Democratic hack" theories) is strengthened by the manner in which belatedly-released cross-tabular information is presented.
Not to bore you too much, but in essence we are told (on pages 5 and 6 of this pdf) that 84% of Tom DeLay supporters are White, 2% are Black, 5% are Asian, etc.; and also that 58% of DeLay’s supporters are born-again or Evangelical Christians, while 35% aren’t, etc.
Who the hell cares? In the real-world of professional politics, the relevant question is "How is Tom DeLay doing among Whites OVERALL" or "How is Tom DeLay doing among Evangelicals OVERALL". The very presentation of the cross tabs reveals an academic mindset, rather than a hardball political mindset to the numbers– which, remember, have an absurd margin of error anyway.
Finally, let’s not overlook this simple fact– respondents are informed at the beginning of the survey that the poll is being conducted on behalf of THE HOUSTON CHRONICLE. Given the paper’s reputation among area conservatives generally– and DeLay’s hard-core base in particular– could that information not reasonably be expected to bias (A) the willingness of many conservatives to participate in the survey; and (B) the willingness of those Republicans that did participate to answer all questions truthfully and fully?

As Laurence might say, "This poll…"

Filed Under Uncategorized ·

Comments

9 Responses to “Chron/Murray poll story has legs”

malcolm on January 18th, 2006 at 10:22 am

Oneth!
Laurence Simon on January 18th, 2006 at 10:40 am

I envision a hill called Pollgotha, where pollsters who use lousy sampling methods or bias their polls to force a particular conclusion are crucified and left to die.

The dream usually ends with me jamming a spear in Zogby’s side, then cutting him down, beheading him, and burning the body into ashes which I toss to the four winds.
Wino on January 18th, 2006 at 10:46 am

One word, Laurence, “Prozac.”

j/k

But I agree that polls have become tools to prove a foregone conclusion rather than a method to determine probable outcomes.
malcolm on January 18th, 2006 at 11:21 am

Many pollsters and their polls have pretty much lost their credibility with me. It seem that they all have something to prove except the true results of the polls themselves. We all know (or should) that the way the questions are worded will skew the poll a certain way, will cause a certain answer and predetermine the outcome of the poll results.
Channel 13’s website home page today has such a question that reads as follows:
“Should the govt. have authority to spy on citizens without warrants?”
Yes ___
No ___

Obviously, that question, the ways it’s worded will get mostly a “no” response.
Worded differently this would certainly have caused a different response.
“Should the gov’t. have the authority to spy on suspected terrorists and their accomplices in order to protect the American people?”
What do you think the pol results would show with that question?
jimb on January 18th, 2006 at 12:09 pm

This is why I tend to dismiss those posts which use polls to play NIGYSOB…

On a lighter note, I had a boss who had a tendendency to use malapropisms in his emails. We were going to do a team lunch and I was asked to ‘pole’ the group and see where most of the folks wanted to eat. I responded to him and said that I don’t think very many of the guys (or gals) in the group would appreciate that…
malcolm on January 18th, 2006 at 1:00 pm

#6 Has it been proven that DeLay is dirty as yet
or is that an opinion?
Ken Kelley on January 18th, 2006 at 4:30 pm

#6 Minor Cog —
Actually, the “arrest” thing has already been handled, and you most certainly have my permission (indeed, you have my encouragement) to contact Ronnie Earle and insist upon a trial at the soonest possible date.
But before you decide that “DeLay is dirty”, would you mind waiting until the trial actually takes place?
pmartin on January 18th, 2006 at 6:16 pm

DFT!
publius on January 18th, 2006 at 10:46 pm

Outstanding claim to fame:When I attended UH, I was an active participant in one of Dr Murray’s election night polls. I went to a polling place on the eastside, and phoned in the final results as soon as they became known. He uses those to make his final predictions. Autographs are available; well, for a minimal fee anyway.