Virginia 2009 Megathread

[Verily]

The polls have McDonnell getting at least 21% of the black vote against each of the candidates. Needless to say, that's not going to happen. But it brings up an interesting point, why do blacks poll for Republicans but don't end up voting for them?

subsample MoE

It's 17% of the sample. It's not that small.

Yes, but the MoE is still quite large. 17% of 1,396 is only 237. The MoE on a sample size of 237 is 6.5%. And MoEs are just an estimate for medium-ish sample sizes. At small sample sizes (smaller than ~400), MoEs really should grow faster than the standard calculation, which is 100/sqrt[sample size].

[Verily]

Huge for Deeds indeed.

WaPo gets it.

Nah, it's huge for McAuliffe. Probably does more for him than endorsing McAuliffe would have. The politically highly informed people the WaPo caters most to are pretty opposed to McAuliffe from the beginning, and right now they're mostly supporting Moran. But if a bunch of them switch from Moran to Deeds as a result of this endorsement, Moran, who is the only candidate polling competitively with McAuliffe, will fall back, with essentially no damage to McAuliffe's own numbers (since few of his voters read/care about the WaPo). So what might have been a 40-35-25 race (McAuliffe-Moran-Deeds) might become more like a 40-30-30 race (and those are pretty pessimistic numbers for McAuliffe).

Maybe it would have been better for McAuliffe to be endorsed by the WaPo, but I suspect most of its readers would have just ignored the endorsement and voted for Moran anyway.

[Verily]

I'd argue that a higher turnout favors Deeds because more Independents and Republicans will have voted.

You'd be stupid.

[Verily]

I'd argue that a higher turnout favors Deeds because more Independents and Republicans will have voted.

Something tells me Democratic turnout matters more...

Well obviously. But the more Independents show up the better it is for Deeds.

No, not really. The scant polling we have suggests Moran does best with Independents. (These may be liberal Independents; they probably are, since they're voting in the Democratic primary.) Deeds is not the sort of candidate with a lot of Independent appeal for a primary. His appeal is to old-time Democrats.

Independents will not be more than, say, 15% of the electorate, anyway, and Republicans less than 5%.

[Verily]

The poll is way off.  McAuliffe winning gun owners, and people who regularly attend church?  No way that happens.  These numbers are way off.

"Regularly attend church", voting in a Democratic primary in Virginia... sounds like blacks to me. Yes, McAuliffe should win them quite handily.

[Verily]

Wow, is this race over?  Sorry I haven't been paying attention to it, there's a governor's race that actually matters going on to the north

No, the race is far from over.  Deeds is down, but he'll be fine in the end, and manage a state that isn't full of corruption.

Is there anything short of pure optimism that makes you think he can overcome what has become a double-digit deficit?

That PPP and SUSA aren't even attempting to weight their polls for the 2008 vote is a bad sign for the polling. Yes, the differential turnout for 2009 compared to 2008 will be bad for the Democrats---but not nearly that bad. They should be weighting to around a 2008 tie, which makes the race quite competitive in both polls.

[Verily]

Wow, is this race over?  Sorry I haven't been paying attention to it, there's a governor's race that actually matters going on to the north

No, the race is far from over.  Deeds is down, but he'll be fine in the end, and manage a state that isn't full of corruption.

Is there anything short of pure optimism that makes you think he can overcome what has become a double-digit deficit?

That PPP and SUSA aren't even attempting to weight their polls for the 2008 vote is a bad sign for the polling. Yes, the differential turnout for 2009 compared to 2008 will be bad for the Democrats---but not nearly that bad. They should be weighting to around a 2008 tie, which makes the race quite competitive in both polls.

They didn't weight at all in 2008, and nailed it perfectly. I didn't hear you complaining then.

Oh, I would. Pollsters should always weight, and they didn't nail it perfectly. That's what I like about British polling; it's miles ahead of American in that way.

[Verily]

Wow, is this race over?  Sorry I haven't been paying attention to it, there's a governor's race that actually matters going on to the north

No, the race is far from over.  Deeds is down, but he'll be fine in the end, and manage a state that isn't full of corruption.

Is there anything short of pure optimism that makes you think he can overcome what has become a double-digit deficit?

That PPP and SUSA aren't even attempting to weight their polls for the 2008 vote is a bad sign for the polling. Yes, the differential turnout for 2009 compared to 2008 will be bad for the Democrats---but not nearly that bad. They should be weighting to around a 2008 tie, which makes the race quite competitive in both polls.

They didn't weight at all in 2008, and nailed it perfectly. I didn't hear you complaining then.

Oh, I would. Pollsters should always weight, and they didn't nail it perfectly. That's what I like about British polling; it's miles ahead of American in that way.

Too bad it isn't in terms of accuracy.  Go ask Al.

Couple of points:
1) Summer polling sucks.  Repeat.
2) That being said, folks seem to forget that pre-2008, pretty much every Virginia election had a party ID of 2-3 points GOP advantage.  I haven't forgotten.

The bad years of British polling general stem from the pre-weighting period. Sure, they're not excellent anyway--no one is. But they don't deliver huge surprises now that weighting is used, pretty much ever, except at the local level where no one except news agencies really pretends you can poll effectively anyway.

And, sure, you're right about the voter ID. I'm not talking about voter ID, which is a stupid measurement and highly fluid. I'm talking about vote recall, which is much more important (maybe the most important) for weighting.