Friday, October 23, 2009

Barack Houdini

OK: the most important point, before I start this, is that it really, really, really, doesn't matter whether Barack Obama's approval rating in the Gallup daily tracking poll is at 50% or 49% on any particular day. The entire consequence of falling to 49%, as compared to falling to 50%, is that there will be one mini-cycle of stories that practically no one with an open mind will even notice (since people who follow politics most closely are also the most committed partisans).


Obama first hit 50% on Gallup in their August 24-26 polling, sparking a story about when presidents typically fall below 50. Actually, the story was written when he was at 51%...and then he hit 50. And stayed right there, for four consecutive days, before climbing back to the relative safety of 51%. He has stayed stuck in the low 50s ever since, returning to exactly 50 in the polling periods beginning on September 22, October 3, and then October 17 and 18. In fact, his Gallup approval average for the entire quarter was 53%, but not a single one of the ninety reported measurements fell lower than 50%. One cannot compute the odds of that (since we don't know whether the true average stayed at 53% over the three months or jumped around some), but I think it's safe to say that the odds are long. It's good to be lucky!

At any rate, Obama has "spiked" back up to 54% in today's report, if you can call a four-point swing a spike. His high since he first dropped to 50% approval has been 56%, and in fact since mid-July his high has been 58%, so at least in the Gallup numbers he's basically been flat for a bit over three months. See also, which has Obama falling until mid-August and then flat, depending on which settings are used.

Two complaints to Gallup, by the way. First, their headline, "Obama Sees Approval Slide in His Third Quarter," is technically accurate but misleading; his approval fell in his second quarter, and was stable (at the new lower level) in the third quarter. Second, their front-page graph is scaled in order to make very small changes look significant. No surprise -- they've switched from polling every two weeks or so to a constant daily track, and so they need to sell it. I love the daily track; political scientists always think that more data is always a good thing. But the downside is that we're apt to get a lot of noise misinterpreted as something more than that, and I think Gallup has a responsibility to do what they can to minimize that.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.

Who links to my website?