On The Mendacity of the Polling; -or- Egg Counting for Conservatives [UPDATED]

One, two, three, four, five... Wow! Look at all those chickens!

When it rains, it pours… All over Tom Barrett.  The most recent Marquette University Poll showing that Governor Walker has passed Tom Barrett in the race for Governor is all the rage among the Walkerbots on Twitter and the squawkers on the radio.  But, like the Walker job numbers, if you scratch at the surface, the rot leaks through to the surface.

The recent shift in polling results couldn’t possibly have anything to do with this…

Data assembled by the author from the data published by Marquette

See that little shift in May? More conservatives and fewer moderates.  And a drop of 5% or so of liberals.  I imagine that this shifting demographic has something to do with the skew in the May polling results.

UPDATE: I went back to the polling data and pulled the age information from the samples.  Here’s what the data shows.

Visually, it appears that age tracks to political views quite nicely.  This replicates what I’ve seen in other polls.  Younger people tend to be more liberal, so failing to control for the age ranges in your sample produces a seriously skewed result.  My quick back of the napkin calculation shows a very strong correlation (r=.6174 at 95% CI) for these two variables (meaning that they have a very high likelihood of moving together).

(h/t  Jud Lounsbury over at Uppity Wisconsin)

 

Share:

Related Articles

7 thoughts on “On The Mendacity of the Polling; -or- Egg Counting for Conservatives [UPDATED]

  1. Careful, though. While yes, it appears the sample had more conservatives in this result, it’s also possible that’s because more people in the sample are starting to identify themselves as conservative. The entire sample population could be shifting to the right.

    I’m not saying it is, only that it could be.

    In contrast I offer an Obama poll from a couple of weeks ago that appears to have over-sampled to the left giving Obama a great lead. Or, of course, the sample could be moving further to the left of that spectrum.

    Touting the latest poll is fun. We all do it left and right, but in the end, it’s that one vote that counts.

    1. While technically correct, I struggle to see how a population would shift in the space of 30 days like that. It’s much more likely to be a sampling error. I’ve added the age data to the post and a correlation number. I think you’ll agree that this is very explanatory, more so than a mass cultural shift in Wisconsin.

      Their sample was insufficiently random.

    2. More wingnut wishful thinking from the sweetly obnoxious, and so much less than fairly conservative, Cindy Kilkenny. Wisconsin becoming more conservative, Miss K? I don’t think so. You’ve jumped into the deep end, and you don’t know how to swim.

      No one knows whats going on for sure, but the fact that Fat-Ass Fitz “deigned” to debate “fringe candidate”, Lori Compas, gives us some idea. I will tell you this, though. ALEC and the Koch brothers are running scared, not to mention bathed in glorious sunshine

      More conservative, Cindy. Yeahhhh, I don’t think so. Cindy, be careful of that which your partisan blinders won’t allow you to see.

      It’s a new day in Wisconsin, Cindy, and all the cinnamon cookies, bedazzled coffee cans and conservative kool-aid in the world isn’t going to change that.

      Here’s a little light reading for you. Spoiler alert. If you want to have a even the remotest chance of being able to “fairly” process it, you will definitely have to come down off of that high horse of yours. Oh yeah, and take THOSE partisan blinders off, too, a’ite?

      http://www.dailykos.com/story/2012/05/17/1092522/-All-Bets-and-Polls-are-off-in-Wisconsin

      From MadisonProgressive

      “The right is trying to discourage us by pushing that Marquette poll saying that Barrett is 5 point down from Walker. I teach statistics and I can tell you that all polling is meaningless at this point. There are too many hangups from the Barrett supporters to get a meaningful sample. What’s really going on is that what we’re doing is too new to know what’s going to happen. All bets are off.

      We are in uncharted territory. We’re on the cutting edge of unprecedented political change. This has the DNC running scared. The Koch Brothers running scared. ALEC running scared. This level of political involvement by normal middle Americans is everything they don’t want to see. They are trying to drag us back to their narrative of dismal expectations. Forget it.

      They foresaw none of this. Not Scott Walker being recalled. Not Scott Fitzgerald being recalled. Not Dale Schultz standing strong and blocking the mine. None of it. They saw none of it. They don’t know what’s gonna happen June 5th.

      Steve Jobs changed 3 industries in one fell swoop. We’re that good. We’re changing the political landscape in the same way. We’re still the heart and soul of the resistance to corporate America taking everything away.

      Everything is up in the air. Let’s make history together.”

      Right on, MadisonProgressive! Right on!

  2. If the pollsters are generally reputable it is not very useful to attack one poll or another as it comes out. You have to look at trends over a number of polls. I rather doubt there is any intentional skewing.

    1. Reputable pollsters don’t publish obviously skewed data without a warning, either. I make no assessment of intent, rather lassitude. Sloth. Although the timing is quite suspect.

Comments are closed.