Speaking of bad university polls, did you see Indiana State U.'s new poll of the Hostettler race in Indiana? Sample with 60 percent women has Hostettler down 23 points, no surprise. (However, they said when the poll was weighted equally by gender, Hostettler was still down 21).
Yes, I saw it. I've adopted IN-02, IN-08 and IN-09, BTW. Yes, this poll had an overly large number of females (60.5%). The female vote broke for Ellsworth by 56.3% to 29.6%, which was very close to the final 54.5% to 32.1% margin; while men broke for Elsworth by 51.4% to 36%
So they weighted the results to equalise men and women and came up with a weighted result of 54.1% for Ellsworth to 32.1% for Hostettler
Personally, I don't think Ellsworth is leading by that much. I'd need to see a non-university poll in IN-8 to confirm this kind of lead
I've been number crunching the internals for the RT Strategies/CD poll (September 8-10), which had Ellsworth leading Hostettler by 6% among all voters:
1) Among women, Ellsworth leads Hostettler by 60% to 36%
2) Among men, Hostettler leads Ellsworth by 56% to 42%
However, among those most likely to vote, Ellsworth leads Hostettler by 56% to 41%
So, if the Indiana State University poll is anywhere near accurate, which is arguable, there has been a sizeable shift among men from Hostettler to Ellsworth
Dave