A better post from Ruffini
ArchPundit
A better post from Ruffini on Zogby
I’m still not clear why the headline reads Is Zogby a Fraud?. However, he makes an argument closer to the evidence.
The problem is he isn’t grasping how this poll is reported. He mentions the black turnout numbers and says they are a problem because they assume 100% turnout by African-Americans. It does no such thing actually. It assumes a turnout of 12% of the voting population based on a likely voter turnout model that is proprietary. The voters included in those numbers have already been determined by Zogby to be likely voters. The only people in the final sample are likely voters as determined by the model so non-voters are already weeded out. If one is saying he got the proportion wrong, that is a legitimate argument, but understanding the samples for what they are is important.
The 95% shouldn’t be surprising and fit with past results and recent history. As I mentioned here while Missouri traditionally has had more African-American support for Republicans, the Ronnie White case radicalized the population.
Strangely, Jim Talent is probably the most open to African-American concerns of Missouri GOPers. However, Bond and Ashcroft have the advantage of being former Governors and used patronage to gather up some support as individuals. Talent never had that position and thus is not likely to get the same numbers of support. I think Talent could have made some inroads, but instead of running a bold campaign in the black wards and towns, he was timid and has allowed the RNC to push a suppression tactic instead. Patrick’s description of the small differences needed to even change the calculus a little bit is very good. Given a sample of only 95 people, the MoE is quite high too. But there is little evidence that Talent will break 10% of the black vote. More importanly, the higher the black turnout, the lower Talent’s percentage will be. This is part of the story Judis didn’t mention. There aren’t many black Republicans, but those few vote. Where the turnout varies is in leaners who are overwhelmingly Democratic and that is why knock and drag is so vital to Democratic prospects.
Ultimately, it all boils down to the GOTV tomorrow.
UPDATE: Let me add that I agree at least two of the results are most likely skewed. Looking for where the problems are is a perfectly legitimate exercise. I’ll largely be leaving that alone until Wednesday when we have more data.