Tuesday, July 21, 2009

O my...

Here's something you don't really want as a Democratic polling company:

Rush Limbaugh gleefully reading excerpts off your blog about Barack Obama's approval rating.

But our numbers are straight, and sometimes that's going to be the result of it.


Ranjit said...


That is a nice blessing. Your business is going to boom. Best of luck

Anonymous said...


Have you considered polling both Likely Voters and Registered or Adults? There is a big, big problem it seems with the various polling methods and most of the time the pollster, whether it's you or Rasmussen, always points out that the reason they have lower approvals for President Obama is because you're polling Likely Voters.

I think it has to do with your polling method--IVR. Using the information on Pollster.com the average of the Live polls during the election had the race at Obama +7.9, while the average of the IVR polls had the race at Obama +6.6.

Contrast that to the approval ratings:
Live: +21.4
IVR: +4.1

How do you explain that huge discrepancy? And also I understand the importance of Likely Voter models but only when it comes to voting. At this point in time doesn't it make more sense to at least poll Registered Voters?

I hate criticizing individual polls, but there are only two pollsters that have Obama with such high disapproval numbers: you and Rasmussen--both IVR. (I choose to ignore Zogby (Internet) because, well, because it's Zogby).

And in regards to Limbaugh...what do you expect? I hope you're not happy about that.

When nearly every other pollster shows completely different numbers--and were just as, if not more, accurate as PPP during the GE--why are you so confident in your results?

I guess that's why your MOE is so high?

Anonymous said...

^^^ Forgot you do RV, not LV (confused you with Rasmussen). Regardless, the point I was making stands. It's still hard to explain why there's such a large difference with other polls.

Tom Jensen said...

A lot of the polls that show Obama with better approval ratings assume a double digit Democratic party ID advantage and that the percentage of Republican voters in the country is in the low 20s. Neither of those things were true on November 4th and they're both less true today because the political climate, while certainly still solid for Democrats, is not as good as it was then.

But I agree that IVR may have something to do with it- but I don't think that means it's wrong. Across the board we find worse approval ratings for all politicians, of both parties, than most live interview pollsters. I think respondents are more willing to be negative about a politician on one of our polls than they are if they're talking to a real person.

One particularly big difference I find between our polls and others on Obama is that virtually no Republicans say they approve of him on ours but that number is more in the 20-30% range on some other polls. Not nearly that many Republicans voted for Obama, and if they didn't like what they saw then I don't know why they would like what they see now.

Still there's no way of knowing who's right and who's wrong on approval polls. If we're wrong because Obama really is more popular than we're showing, that's fine by me. Thanks for your thoughtful comment.

Anonymous said...

Thanks for the comment.

I would only say the following. If you miss on your projection of the Party ID, that'll mess up your Liberal/Moderate/Conservative breakdown (which I believe to be more important than the Party ID breakdown).

While I agree with you in that Republicans aren't likely to be very supportive of the President, I don't agree with that assessment when applied to Conservatives in general.

Wouldn't you say your breakdown for the various ideologies is more Republican-leaning than what most pollsters believe is actually the case?

You have it 18/43/39 in your latest release. The 2008 exit polls showed the breakdown at a 22/44/34. A +10 Moderate advantage in the country has shrunk to +4 according to your polling at the expense, mainly, of Liberal identification.

Those numbers just seem off (in addition to the favorable Republican identification numbers). That will obviously skew his numbers downward (and his disapprove numbers upward).

Web Statistics