Friday, February 29, 2008

Media Polls

PPP and Civitas don't agree much, so when we do agree on something you can pretty much bet it's true. And one thing we agree on is that the results of today's WRAL/Rasmussen poll sound pretty fishy.

We can't know for sure though, because Rasmussen didn't tell us the party balance of the folks they polled. Or the gender balance. Or the racial balance. Or the age balance. Or pretty much anything that would let us make a basic analysis of the validity of their poll.

You get all those things in any PPP poll, which raises an interesting question. Isn't the media supposed to be a beacon of transparency, certainly more transparent than our 'partisan' polling company?

WRAL isn't the only tv station in the Triangle commissioning a poll. WTVD has been getting a monthly poll too. They get it from Survey USA and it comes complete with all the useful information you would want to analyze a poll.

In my opinion WTVD has it right and WRAL has it wrong. WRAL should either make Rasmussen release the demographic info on its poll so we can see what factors might be leading to this seemingly too friendly to Democrats poll, or else they should contract with someone else to do their poll who will release that information.

It's perfectly logical for media outlets to commission polls- but if they're going to, they should make an extra effort to make sure they're getting a good poll and abiding by the same transparency that they expect out of the decision makers they cover.


Chris said...

SurveyUSA isn't perfect either (but then again, who is?). They seem to have too much of their sample from younger voters.

IIRC, half their sample in their last survey came from the 18-45 age range, which according to voting statistics, is too high. Despite all the rhetoric about younger voters, the facts are, they just don't vote in the same intensity to older voters (especially those over 65).

Anonymous said...

What are you talking about - I just went to the story on -

Right under the graphic it says related documents - one of those documents is called "Breakdown of Poll Results". When you click on that you go to this url...

On this page you will find the poll results broken down by age, gender, party, political ideology, race, income, marriage status, kids/no kids, and religion.

What sort of blogger are you? Did you honestly not take the time to actually look at the story and the related information before posting this blog?

Come on this is weak!

Tom Jensen said...

I'm not talking about the crosstabs. I'm talking about the topline demographics. What was the partisan distribution of the sample polled? What was the gender distribution? What was the racial distribution? What was the age distribution? I'm not looking to see how the candidates fared within each of those categories, I'm looking to see what percentage of those polled was in each of those categories.

My guess is that the numbers of Democrats polled was disproportionately high compared to the number of Democrats in North Carolina, and that led to overly negative approval ratings of the Republican candidates for Governor.

If you're going to write something unpleasant on my blog please have the courage to put your name on it, thanks.

Web Statistics