New Los Angeles Times Poll is Outlier: Democrats Oversampled in Survey From Left-Leaning Greenberg Quinlan Rosner

The headlines for the new Los Angeles Times poll are falsely indicating a November blowout for the Democrats in California. My hardcopy newspaper headline reads: "Brown's Lead Doubles in One Month." And at LAT's homepage, "Brown's lead over Whitman widens to 52%-39%; Fiorina not gaining ground." And the killer is the New York Times' blog post, "Brown and Boxer Have Significant Leads, New Poll Shows."

Note first that other recent surveys find the governor's race no better than "lean Democrat" (see Rasmussen's survey out yesterday, with Brown holding a four-point edge with a +/-4 percentage point margin of error). And Survey USA, which routinely tilts left, has a poll out this week with Brown at 47% and Whitman 40%. That same survey throws some particularly interesting light on the CA Senate race. The survey is leaning Democrat in all the statewide races, but is a toss-up for Boxer-Fiorina, Boxer 46%, Fiorina 44%. (This needs more investigation, but there's some interesting speculation that support for marijuana decriminalization among highly-motivated voters is lifting Boxer's numbers. I'm predicting a defeat of Prop. 19 at the polls, so perhaps Survey USA's numbers hold even worse implications for Boxer's chances.) And Rasmussen's poll out yesterday had "Boxer picking up 48% of the vote, while Fiorina draws support from 46%."

I don't want to overstate the point, since it's always uphill for Republicans in California. But there's something suspicious with the methodology at the Times, as indicated at Greenberg Quinlan Rosner, who conducted the poll:

These findings are based on a random sample survey of 1,501 registered voters in the state of California conducted from October 13-20, 2010. These findings are also based on 922 likely 2010 voters. Likely voters are defined as registered voters that meet certain conditions based on previous vote history as determined from a voter file, likelihood of voting in 2010, and enthusiasm in the election. This includes respondents who have already voted, voted in both the previous two general elections who indicate they are “almost certain” or “probably” will vote in 2010 and those who have registered since the 2008 election due to ineligibility who are “almost certain,” all of whom must respond as a 5 or higher on a 0-10 enthusiasm scale.

And:

An oversample of 400 Latino registered voters were interviewed by telephone. All interviews from the Latino sample were carried out by bilingual Latino interviewers, and conducted in the preferred language of the survey respondent, English or Spanish. Overall, 41 percent of interviews with the Latino sample were conducted in Spanish and 59 percent in English. The technique of using fully bilingual interviewers is greatly preferred because it does not terminate calls with Spanish-language households and require a callback, which can be difficult to schedule with language barriers ... Upon completion of all interviewing, the results were weighted to bring the Latino oversample population into line with the racial and ethnic composition of registered voters in California. The data were weighted to reflect the total population of registered voters throughout the state, balancing on regional and demographic characteristics for gender, age, race and education according to known census estimates and voter file projections.

Folks can check other analysts (Nate Silver, for example), but given the huge Democrat advantage in California registration (see 2008 totals here at Page 4), and the tremendous Democrat enthusiasm in previous elections, it's pretty clear that the Times survey has oversampled Democrats. Likely voters are tilting toward the Dems at the Times, and Latinos are supporting Democrats this year by a roughly 2-to-1 factor: "Boxer has opened up a whopping 64% to 28% lead among Latinos."

The Los Angeles Times is pressing its fingers on the scales to favor the Democrat Party. There's widedly varied results across polls (the Field Poll had the governor's race tied at 41% last month), and Whitman may indeed be washed up, but I'd hardly count out Fiorina. Additional influences include the ground game over the next week, GOTV efforts on election day, and any last minute bombshell surprises.

ADDED: The Other McCain links: "Disinformation is the new objectivity."