Wednesday, June 25, 2008

Obama 12-Point Lead, in L.A. Times Poll, Draws Controversy

The new Los Angeles Times poll, which finds Barack Obama with a 12-point lead over John McCain, has drawn fire from analysts suggesting an oversample of Democratic respondents.

Don Frederick has the details:

A well-known Republican research firm argues that the voter pool tapped for the new L.A. Times/Bloomberg poll was too skewed toward Democrats - a challenge that causes the GOP strategists to question the double-digit lead the survey gave Barack Obama over John McCain.

The case against the poll, laid out in a memo sent out today by Public Opinion Strategies, in turn sparked a response from survey director Susan Pinkus, who stood by its methodology and findings.

Part of the dispute reflects a long-standing disagreement between independent pollsters and partisan operatives (something The Times
wrote about four years ago) -- whether or not to tinker with a poll to make sure its respondents reflect the nation's political composition at some fixed point, such as the most recent election.

Pinkus, like most nonpartisan pollsters, rejects that notion. Discussing the current survey, she says, "The poll was weighted slightly, where necessary, to conform to the Census Bureau’s proportions of sex, race, ethnicity, age and national region. The poll was NOT weighted for party identification since party ID is a moving variable that changes from one election to another, or when one party may be favored more than the other."

As a result, the survey simply asked respondents their party affiliation or inclination, and came up with this breakdown: 39% Democratic, 22% Republican, 8% something else, 4% refused to say.

There's the rub, insists the memo from Bill McInturff, Liz Harrington and David Kanevsky. They write that these figures, and the 17 percentage-point gap between the two parties, are "greatly out of line with what most other surveys are reporting."

The memo cites several other recent polls in which the party ID gap ranged as low as plus 6 percentage points for the Democrats to as high as plus 14.

It then asserts: "McCain’s double-digit deficit is not a reflection of reality, simply a result of an unusual party identification result in this survey.... If party identification on the L.A. Times survey is recalculated to ... 29% GOP / 39% Dem / 27% Ind / 5% Don’t Know/Refused, the ballot would be 40% McCain – 47% Obama."
Pinkus responds, at the link. She suggests that the statistic of 39 percent Democrats was generated from a random sample of "1,115 registered voters (which includes listed, unlisted and cell phone users)."

These numbers look like outliers to me. For example,
today's Gallup survey finds Obama and McCain Tied at 45 percent:
The latest Gallup Poll Daily tracking update on the presidential election finds John McCain and Barack Obama exactly tied at 45% among registered voters nationwide.

Voter preferences had been fairly evenly divided for the past week, with Obama generally holding a slight advantage of two or three percentage points. This is the first time since Gallup's May 31-June 4 rolling average that Obama does not have at least a slim advantage over McCain. Obama's largest lead to date has been seven points. (To view the complete trend since March 7, 2008, click here.)

Since the changes from Tuesday's results are well within the margin of sampling error, it is unclear at this point if today's results represent a further tightening of the race. The last two individual nights of polling have, however, been more favorable to McCain that what Gallup has shown for most of June.
That sounds much more accurate, but see also Andrew Romano, "Does Obama Really Have a Double-Digit Lead?", and Marc Ambinder, "McCain Campaign Pushes Back On LA Times/BB Poll."