Well, we might know what pollster the GOP is using to get their numbers (OK, that’s a joke…sort of) with Mason-Dixon releasing a poll this morning showing McCain only down 4% in Pennsylvania, 47-43. Of course, the issue with this poll is that it’s completely unlike any other poll in the state.
This isn’t the first time Mason-Dixon has released such a poll:
- Michigan - In their Sept. 18 - 23 poll, Mason-Dixon showed Michigan tied, even though polls released at the same time showed the race +5, +7, +7, +8, +10, and +13 in the state
- New Mexico - In their August 13 - 15 poll, they showed McCain up 4% when other polls at the time showed Obama up 4%, 6%, and 13%
- Virginia - In their Oct. 20 - 21 poll, they showed only a 2% Obama lead when other polls at the time showed Obama leads of 10%, 6%, 10%, 9%, 9%, and 8%
To be sure, Mason-Dixon has also released quite a few polls which are right in line with the other polls in the state. However, Mason-Dixon has a bad habit of having a slight GOP bias and, as I’ve shown above, sometimes a pretty significant GOP bias that amounts to McCain performing anywhere from 5% to 8% better than he is in virutally any other poll released at the time.
Now, 538.com shows Mason-Dixon as slightly above-average compared to other pollsters in the primary, but of course, that’s the primary, where you don’t have to worry about things like…party weighing, for example.
October 30, 2008
Well said - I have had question about the validity of the polling numbers provided by Mason-Dixon and this post confirms my thoughts:
Non bias with an underlying party alliance.