Today we have two new polls of the L.A. mayor's race, with two different results. Both of them have some problems which make them less than totally reliable.
That's a bigger lead than we've seen in other surveys
, and the crosstabs explain why.
First, here's the result:
Eric Garcetti 26
Wendy Greuel 18
Kevin James 12
Jan Perry 12
Emanuel Pleitez 4
The firm surveyed 544 likely voters from Jan. 12 to Jan. 15. The margin of error is 4.3%.
The problem is with the likely voter sample. Looking at the crosstabs, we can see that Survey USA believes that 44% of the March 5 electorate will be Latino, and only 35% of the electorate will be white. That's a red flag.
Garcetti does well with Latino voters. (In this poll he takes 35% of the Latino vote, to Perry's 13% and Greuel's 12%.) So the more Latinos in the sample, the better Garcetti's numbers. So how big will the Latino vote share be? Nobody really knows, but the Garcetti campaign has predicted a Latino turnout of 25% to 30%. The Greuel campaign puts it at 21%. Nobody is saying 44%. For reference, when Antonio Villaraigosa was on the ballot in 2005, Latino turnout was a record-setting 25%.
Survey USA arrived at its likely voter sample by asking a larger pool of respondents whether they intend to vote. The firm did not screen the sample based on past voting history, which some other pollsters do in low-turnout races.
"We start with no preconceptions," Jay Leve, the CEO of Survey USA, told the Weekly. "We let that water seek its level. We found active interest among Hispanics in this election."
John Shallman, Greuel's strategist, took issue with the firm's methodology, arguing that robopolls are unreliable. Shallman also questioned how Survey USA could project that 63% of registered voters are likely to vote, given that turnout in municipal elections is usually in the mid-20s.
"This poll is bogus and contrary to every
public and private survey we've seen," Shallman said.
With that, let's move to the second poll, which comes from the Greuel campaign. The poll memo is embedded below. Here's the result:
Wendy Greuel 20
Eric Garcetti 18
Jan Perry 8
Kevin James 7
Emanuel Pleitez 2
The poll was conducted by Celinda Lake, a Democratic pollster who also does work for EMILY's List. Her firm surveyed 803 likely voters from Dec. 5-9, using live interviews. The margin of error is 3.5%.
The first obvious problem is that the poll comes from Greuel's campaign. They wouldn't release it if she were behind. The second problem is that it's more than a month old. It appears to have been released to slow any momentum that Garcetti might pick up from the Survey USA poll.
"We've had three polls that have all had Eric ahead," says Bill Carrick, Garcetti's strategist. "She's trying to come up with something to push back. She put out an old poll that predates much of the negative coverage she's gotten."
Another problem with the poll is that we don't have the crosstabs, so we can't tell whether demographic groups are sampled in a reasonable way.
Carrick's bottom line: "We all know the reality is this is a very close race. It looks like Greuel and Garcetti are going to be in a runoff."
Greuel Poll Memo by