Is the WaPo/ABC News poll “worthless”?

posted at 1:54 pm on February 6, 2012 by Karl

That’s the verdict from the normally easygoing Ed Morrissey. While I agree with his biggest criticism of the poll, it is still possible to get something out of it.

I agree with Ed that the recent tactic of not disclosing the party breakdown of the sample is simply absurd. In an era where trust in institutions — including journalism — is low and demands for increased transparency are on the rise everywhere, hiding this basic information from public view invites skepticism and ridicule. The WaPo, ABC News, and Gary Langer ought to be embarrassed.

However, Ed also complains that it’s “a poll of general population adults rather than registered or likely voters, so it’s not even a proper polling type for the predictive outcome they claim.” The poll does in fact provide head-to-head results for both adults and registered voters; the WaPo noted both results for each in its accompanying coverage:

In a general-election test, Obama leads Romney 52 to 43 percent among all Americans; more narrowly, 51 to 45 percent, among registered voters. Among all adults, it’s Obama’s first time topping 50 percent in a head-to-head matchup with Romney since July; it’s his first time ever above that point among registered voters.

(Ed has updated his post to reflect this, while noting that other hyped aspects of the story do not give the RV numbers, which is certainly a fair critique.)

The history of this poll, and comparison to other polls, can tell us a bit about what is going on in this particular poll, even without the party breakdown of the sample. Ed correctly notes that the sample in this poll tends to produce Dem-friendly results, which is probably why the recent decision to omit data about the sample really set him off. However, I would add that the dynamic producing those results has been that this poll historically tends to undersample both parties (and disproportionately undersample Republicans). The corollary, which (afaik) Ed has not stressed, is that the result inflates the sample of Independents.

Accordingly, this nugget from the WaPo coverage is doubly notable:

Obama’s momentum since mid-January has evened the score with Romney among political independents. Among independent voters in the last Post-ABC poll, Romney held a 12-point edge; now these voters split 48 percent for Obama, 47 percent for Romney.

First, this reportage tells you that the poll is still collecting the party data but not reporting it in the released results. Second, when you compare this poll’s results to other recent polls (1/12 – 2/5), the Obama +6 result is not particularly out of line. Indeed, the topline results here merely add 2 or 3 points to each side of the Rasmussen poll of likely voters conducted at roughly the same time, which is margin of error type stuff. And it’s not all that different from the mid-January PPP poll which showed a more pronounced Obama surge with independents. This poll’s similar gap with higher numbers suggests this poll’s sample probably includes more Republicans and possibly more Democrats (as the PPP poll did) at the expense of the now supposedly more Obama-friendly Indies.

What accounts for the supposed Obama surge with Indies? One possibility the WaPo coverage raises is the State of the Union speech, which fell within this poll’s window. However, that would not account for the surge in the PPP poll. A more plausible explanation is the modest uptick in the economy (and it’s overhype in the establishment media). This poll has Obama improving a few points not only in overall job approval, but approval on how he’s handling the economy. However, even this poll has his job approval with Indies underwater, so presumably his approval on the economy does not look great with Indies.

Accordingly, the underlying dynamic in this poll is probably similar to that seen in the PPP poll: it’s not about Obama as much as it is about Romney. Q25 in this poll shows 52% say that the more they hear about Romney, the less they like him, which is not as bad as Newt Gingrich’s 60%, but still bad. This is a function of the campaign and its media coverage. Technically, Romney gets marginally better coverage than Obama… but Romney is getting more coverage than Obama. Thus, people are hearing more negative coverage of Romney than Obama. Obviously, the balance will shift once the GOP nominee is effectively known. And this is one reason why head-to-head polling is basically meaningless at this point in the cycle. So it’s a bit ironic that the head-to-head is where the WaPo/ABC poll chose to report the results for registered voters.

—

Update: I wasn’t even going to mention this, but Dem pollster Margie Omero does at the HuffPo:

Today the Washington Post/ABC News released a survey showing Obama over majority support among registered voters (51% Obama, 45% Romney). But as Romney’s pollster Neil Newhouse (a partner in the firm Public Opinion Strategies) pointed out in a blast email, the poll asked about a few of Romney potential liabilities just prior to the vote question. This goes against polling best practices, and it’s possible the survey shows elevated Obama numbers as a result.

Omero also notes that Obama’s liabilities were not questioned before concluding that the underlying issue is Romney’s likability. Again, if Romney is the nominee, that is likely to shift. But Omero highlights that the problem with the poll mirrors the dynamic in the media coverage.

Blowback