Once Again Marquette Poll's (WI) Sample Grossly Oversamples Conservatives, Undersamples Moderates

[WI_DEM]

Polls are only relevent if their sample is reflective of the electorate. As the graph below demonstrates, compared to exit poll data (averaged from 2010, 2008 and 2010) the recently released Marquette poll grossly oversamples conservatives and undersamples moderates.

This is, of course, quite significant considering that that same poll shows Barrett beating Walker 50 to 42% among moderates.

Click on link to see graph showing over and under sampling:

http://www.uppitywis.org/blogarticle/once-again-marquette-polls-sample-grossly-oversamples-conservati

[Scuba]

1. Only election fraud can save Walker now.

[liberal N proud]

2. That is how you set the stage to steal an election

Then you make sure the rhetoric includes some good falsehoods and get both national parties to proclim the Wisconsin election a trial run for November.

[cthulu2016]

3. The sample used

favors Obama big over Romney, and has a distinctly unfavorable view of Romney. Walker's approval = 51% Romney's approval = 40%.

In 2008 Obama won WI by 14% (13.9%). In this sample his approval disapproval is plus 9% and his favorable unfavorable is up 14%. Is there a reason to think Obama ought to be polling better than that in WI today? That's certainly where I'd expect it to be. An easy state for Obama but by a few percent less than in 2008.

The sample is more hopeful about the local economy than it probably was when the recall movement started, with predictable benefit to both the incumbent in the White House and in the Governor's mansion. And a state that is distinctly pro-union except for public service unions which it dislikes.

Maybe the poll has a terrible sample that just happens to coincide with most other polling.

Maybe the poll has a bad sample that will be spectacularly wrong when Barrett wins.

Maybe more RW self-described moderates are fired up about their antipathy to public workers and are self-identifying as conservatives.

Maybe people answer questions about ideology differently in exit polls versus pre-election polls.

Or perhaps averaging exit poll data from 2006, 2008 and 2010 is an arbitrary choice a blogger made to maimize the appearance of something and that does does not offer an accurate picture of current ideological self-identification. 2010 was a RW year but 2006 and 2008 were probably the two biggest Democratic years since 1974-1976. Why three elections? Why not two elections or four elections? Using the last three elections would surely produce a more left-leaning result than using the last two, or the last four.

It's a campaign season. People on all sides massage the facts until they get a happy ending.

The truth could be any of those things, or something completely different. But if, hypothetically, a RW blogger started showing how if you used this particulary measure over this particular time frame and look at it just right it looks better for his candidate, you would laugh.

I very much want Walker to lose, but that doesn't mean that I should accept any happy-talk someone comes up that seems to support my hope.

We will see what election day brings.

https://law.marquette.edu/poll/wp-content/uploads/2012/05/MLSP6_Toplines.pdf

[madrchsod]

4. i was looking at that poll and thought....

what the fuck? how in the hell could that be anywhere near the what is happening ...

thanks for the real story behind the numbers

[hrmjustin]

5. Walker is likely ahead by 2-4%. This poll shows...

...Obama ahead by 8% so that means that people who say they will vote Obama will vote for Walker. That sucks! I hope this poll is wrong, but I don't think it is.

[Fresh_Start]

6. Thats good news....let them think they are winning

so that they sit back and chuckle instead of doing something effective