OH: 52-47 Obama, PPP

[J. J.]

Initiate J.J effect, tied race

Probably.  PPP has problems. 

[J. J.]

Initiate J.J effect, tied race

Probably.  PPP has problems. 

Jesus, JJ. You've lost it all now.

Yes... when PPP is saying what you want, they're reasonable... now they've got problems

We've been talking about the same problem for 2-3 weeks.  It looks like they have a fairly strong D house effect.  Their methodology is good, but they have this house effect.  

[J. J.]

Initiate J.J effect, tied race

Probably.  PPP has problems. 

Jesus, JJ. You've lost it all now.

Yes... when PPP is saying what you want, they're reasonable... now they've got problems

Was J.J. actually calling PPP reasonable at some point in this campaign? I don't specifically remember that but I don't not remember it either.

Yes. When other pollsters had Obama +8 and PPP had him +4 in Ohio, and when after the debate Obama fell to +4 for other posters and +1 in Ohio, J.J. frequently cited the PPP polls from this state as indicative of Romney's strength. That lasted as long as Ohio was atypically pro-Romney in PPP results.

What part of house bias don't you understand.  I've been talking about it with PPP in MO. 

[J. J.]

Don't bother. He's not even trying to feign actual level-headedness anymore. If JJ is actually arguing that the only person that's right is Rassmussen and 13 other polls have it wrong, there's nothing more that can be done but wait for Tuesday and watch him vanish.

Here is what I've been arguing:

They seem to lean Republican in some states in 2008, but got others well.  They nailed Florida and Missouri, but had McCain up in Ohio and North Carolina.    They were perfect in 2004.  I really want to see their Ohio poll this year.  They nailed it in 2004.  If Obama is winning in the final Mason-Dixon Ohio poll, that will give me the clue that Obama has won.  


The one time they leaned Dem was in 2010 when they had Alex Sink ahead by three in their final poll in the Florida governor's race.  

Well, even a three point lean puts Romney up by double digits.  

The point I'm making is that we have four polls from MO, all showing Romney with a comfortable lead, all recent.  Three of them have it double digit lead and only PPP has it about half of that.  PPP also shows no movement.  Are these three off and PPP right?  Does PPP have a problem?

This was last week and I think I've been talking about it longer.

What explanation are you going to provide if Obama really wins Ohio by about this margin?

If it is five points I'll say PPP got it right and these others got it wrong.  Now tell me, if it isn't five points, will you say the same thing.  

[J. J.]

It depends what the margin is, J.J., and what happens in other states. If Obama wins by 3 in Ohio but is underestimated in other places...that wouldn't be indicative of a house bias.

Well, I have tried to find another state, where everyone basically agrees on the winner, and where there are several relatively good polls (Rasmussen, M-D).  It is nice how you equivocated, however.

You must be logged in to read this quote.

That wasn't random.  I wanted a state where the outcome was clear and where there were several good polls (one, Yougov is secondary) that were comntemporary.  I wanted to make sure one of them wasn't an outlier.  MO was the state.

We may be seeing a similar effect in PA, where it looks like it is tightening much more than PPP says (and I still give it to Obama).

[J. J.]

J. J.'s First Rule of Elections:  "If a candidate that say something like 'I don't look at the polls,' or 'The only polls that matter are the ones on Election Day,' that candidate will lose."

BRTD, you forgot:

Corollary to the First Rule:  "Never trust just one poll.  Trust several and remember that the electorate changes its mind quickly."

[J. J.]

J. J.'s First Rule of Elections:  "If a candidate that say something like 'I don't look at the polls,' or 'The only polls that matter are the ones on Election Day,' that candidate will lose."

BRTD, you forgot:

Corollary to the First Rule:  "Never trust just one poll.  Trust several and remember that the electorate changes its mind quickly."

But, J.J., we are trusting several polls. You're barely trusting any of the polls at all.

You are focusing very strongly on PPP.  It is a good poll in its methodology, but it produces a house effect.  I really don't know how much, but it looks like it's above D+3.  

[J. J.]

Here are polls since the 10/29:

11-04   Public Policy Polling   O +5   47%   52%   0%   2%   3%   1,000 L   0
11-03   YouGov   O +3   46%   49%   0%   5%   4%   1,620 L   0
11-03   Zogby   O +8   42%   50%   0%   8%   4%   827 L   0
11-03   Columbus Dispatch   O +2   48%   50%   0%   2%   2%   1,501 L   0
11-01   Ipsos   O +2   45%   47%   3%   5%   4%   936 L   0
11-01   Marist College   O +6   45%   51%   1%   3%   3%   971 L   0
10-30   Public Policy Polling   O +5   45%   50%   0%   4%   4%   600 L   0
10-30   University of Cincinnati   O +2   46%   48%   4%   2%   3%   1,182 L   0
10-29   Survey USA   O +3   45%   48%   0%   7%   4%   603 L   0


Zogby and Ipsos are Internet; CD a mail poll.  Marist is not a good college poll.

So, we have PPP, SUSA, UC, and YouGov.  SUSA has just over a D+2 house bias, YouGov, just under 1.  PPP, at least +3.

Edge Obama, but easily within the MOE, and easily affected by GOTV (which I think will determine it). 

[J. J.]

J. J.'s First Rule of Elections:  "If a candidate that say something like 'I don't look at the polls,' or 'The only polls that matter are the ones on Election Day,' that candidate will lose."

You keep forgetting:

Corollary to the First Rule:  "Never trust just one poll.  Trust several and remember that the electorate changes its mind quickly."

[J. J.]

Question, I know the answer, but I would like JJ to say what he believes it to be, since we're now going after PPP, what was their performance in 2008 vs the results?

Also, you just said PPPs house bias was "at least +3" how did you determine that?

The house effect was noted by Silver over the summer.  It appeared to clear up in September, but it looks like it went back in October.  I noticed it in MO, a state that shows a clear Romney victory.

It isn't that PPP is bad, but that it just has a house effect.

PPP was great in 2008, and I think had a slight R house bias.

[J. J.]

Guys, let JJ be. He'll almost certainly disappear for a long time again after election day.

Ah, who told you that I disappeared?  I was posting through June or July 2009?  I had another project that took a lot of time.