The New York Times this morning reported two national polls completed Sunday that show a double-digit lead for Barack Obama over Hillary Clinton in preference for the Democratic nomination – the Times' own, with CBS News; and one from Gallup with USA Today. There were, however, two others – from Gallup alone and from AP/Ipsos – that show a much closer race, virtually dead heats (Obama +2 and +3).
With these differences beyond sampling error, the reason why is a puzzlement. Frank Newport of Gallup has a piece on Gallup.com discussing how his organization has two polls on the same day, one with Obama +12, the other with Obama +2, but he arrives at no firm conclusions. Question order always is a possible culprit, but it doesn't usually make for differences like these. All four polls had about the same numer of undecideds. One was among likely voters, but the other three were among the general population. And the two conflicting Gallup polls had the largest sample sizes.
I’ve noted before that primary polling, lacking the anchoring influence of partisanship, can be a squirrelly affair. Sampling approaches can cause differences (e.g., ABC News and The Washington Post alone have oversampled African-Americans in every poll we've done this cycle). And there can be flexion points around which polls simply differ because attitudes are unsettled.
For the moment, with the cause of these differing estimates up in the air, when considering the two national polls that show Obama ahead it would be prudent also to keep in mind the two that show the race essentially tied.
But all of them, of course, are a far cry from the substantial Clinton lead we saw in ancient times – like a month ago.
Source End date Interviews Pop. Clinton Obama Diff.CBS/NYT 2/24/08 427 LV 38 54 +16 AP/Ipsos 2/24/08 473 GP 43 46 +3Gallup 2/24/08 1,294 GP 45 47 +2USA Today/Gallup 2/24/08 1,009 GP 39 51 +12