Pages

Monday, September 10, 2012

More on the polling...


Peter "DaTechGuy" Ingemi has a detailed analysis at his blog specifically looking at Public Policy Polling's recent Ohio poll. Basically, although PPP shows Obama up by 5, the numbers seem, well, less than reliable when you look at the underlying figures.

First of all, the poll was slanted 8 points towards women (Obama does better with female voters; Romney does better with men):


...And the poll also seemed to oversample Democrats:


Anyone think this might sway the results in Obama's favor? As Ingemi explains:

Hmm a 4 pt split advantage to Democrats, how does that square with the actual party split in the state? I’m not a big fan of Wikipedia but it is the quickest source I could find it shows the demographic split as in Ohio as GOP+1
So in a state that has a 1 pt GOP advantage PPP and the MSM is touting a poll with a 4 pt Dem Skew, a five point difference and lo and behold, what happens? Obama is up 5.


Further reading:

2 comments:

  1. do they use these kind of slanted numbers with all polls? how can we believe any of them?

    ReplyDelete
    Replies
    1. That's the exact problem with most polls - you can get practically any result you want just by manipulating the sample from which you poll.

      Delete