Picking on Pollsters

It’s only April and I am getting poll addicted already. Sorry to bother you, but I’d like to update my views on some Pollsters whom I suspect of Republican bias (in the statistical sense).


First Rasmussen. Have you noticed that Rasmussen reports much higher Obama job approval numbers than other pollsters ? I have. I think that they (that is Scott Rasmussen) recognized that the had a problem with overestimating Republican share of votes and they corrected their sampling — I know they use some web based polling, because they are not allowed to call cell phones (computers may call fixed lines but not cell phones). I fear they have over corrected.

But our old buddy Gravis Marketing is still reliably red. They are largely responsible for the fact that the Pollster average shows McConnell and Lundergan Grimes tied in Kentucky.
http://elections.huffingtonpost.com/pollster/2014-kentucky-senate-mcconnell-vs-grimes

So far in 2014 Grimes is ahead in 2 polls, one is tied and McConnell leads in four — two by one point in each and two by five points. Both the polls reporting a solid McConnell lead were conducted by Gravis Marketing.

Now I recall someone from Gravis marketing showed up in a comment thread here and was very patient and polite, so I the “friends” wasn’t (all) snark. But I do think they need to work on their sampling strategy.

Also Wenzel Strategies — they had two polls with huge McConnell leads in 2013.