Friday :: Feb 11, 2005

Don't Get Too Excited Yet About Today's AP/Ipsos Poll On Bush's Falling Approval Rating


by Steve

I would like to make a big deal out of the poll released today by AP/Ipsos that shows Bush’s job approval rating plummeting to 45% and his disapproval rating increasing to 54%, even after the SOTU and the Iraq elections. But I cannot do so, for the same reasons that I can’t accept Gallup’s recent numbers. You’ll recall that Gallup tells us in their latest poll that Bush has a 57% approval rating, without telling us that the poll contained 37% Republicans and only 28% Democrats. However, the latest AP/Ipsos Public Affairs poll out today showing the plummeting numbers for Bush and consumer confidence was built on a sample that had a disproportion in favor of the Democrats. Apparently Ipsos takes independents and rolls them into being leaners either for Republicans or Democrats after additional questioning. And yes, Ipsos like other pollsters weights their samples to reflect census demographics. But even with that weighting, the latest AP/Ipsos poll was based on a sample that ended up 52% Democrats, and only 40% Republicans.

Now, my problem with Gallup and its nondisclosure of the party ID of their samples is well known. I also simply do not feel that there is a nine-percentage point advantage for the GOP over the Democrats right now (37%-28%), or at any time in recent history. I think that Gallup should disclose these numbers up front when they release their polls so that consumers of their polls can judge for themselves the poll’s credibility. But I also don’t think that the Democrats enjoy a 12 percentage point advantage over the GOP now either, even with independents included, unless the bottom has dropped out of Bush’s support of late. Yet that seems to be what the AP/Ipsos poll is stating. And as much as I may have problems with this poll because of party ID, I also note however that Rasmussen reported this morning that Bush’s approval rating has fallen to 48%, only three points above what Ipsos is reporting.

Again, to Ipsos’ credit, they fully disclose in their release the party ID and how they factor independents towards one party or the other through leaning. And Ipsos is weighting for demographic factors like other pollsters are, including Gallup.

I have always assumed that Independents, when pressed, break evenly between the two parties, which is why I have a problem accepting this 52%-40% split. However, this data from Ipsos shows that when pressed, more Independents leaned to the Democrats than the Republicans. If this is an accurate reflection of current opinion amongst adults, then this may portend a shift that will cause Bush problems for the next four years and bode well for Democrats in the 2006 midterms.

Anyway, Bush had better hope that AP/Ipsos poll is not representative of adults as a whole right now. In this poll, the right track/wrong track number is bad for Bush, with only 38% feeling the country is on the right track and now 58% feeling we are on the wrong track. In another bad sign for Bush, 59% of these respondents were under 50 years of age, and 39% of the poll’s respondents had a household income of over $50,000 a year. In other words, it is Bush’s base that is unhappy with the direction of the country as well.

Steve :: 11:00 AM :: Comments (8) :: Spotlight :: Digg It!