Oct. 25 2012 12:00 AM

Respondents could've swayed the poll by almost 7 percent

Carl DeMaio
Photo by David Rolland
When U-T San Diego, in conjunction with USD's Center on Education Policy and Law, came out with a poll last weekend showing Carl DeMaio with a 10-point lead over Bob Filner in the San Diego mayor's race, folks took note of the disclaimer that ran with the poll results: 

Employees of a newspaper, television / radio station, marketing / public opinion research company or the city of San Diego-or who live with someone employed in one of those fields-were excluded. 

While it's not unusual to exclude media from polls, excluding city workers and people who live with them seems odd. I sent an email to Scott Keeter, past president of the American Association for Public Opinion Research (AAPOR) and the director of survey research for Pew Research Center to get his thoughts. Here's what he said: ---
The exclusions for people in the research industry are pretty common. I hear them on occasion when somebody calls my house to do a poll. The exclusion of the city employees seems unusual since they are a part of the voting public and obviously very interested in the election. 
Too interested, perhaps. David Cantor, managing director for The Glover Park Group, which conducted the U-T poll, told me, "People who work in city government are typically excluded from political polls as a way to avoid a conflict of interest." 

The city currently employs 10,132 part-time and full-time workers. Last year, city government was San Diego's seventh largest employer. 

Cantor said that 44 people were excluded from the poll because they lived with a city employee. That's more than 7 percent of the total number surveyed. If included in the survey, those respondents could've swayed the poll by as much as 6.7 percent. Common sense dictates that those votes would swing in Filner's favor due to DeMaio's long-running feud with city-employee unions. 

Michael Smolens, the U-T's politics editor, said city workers weren't polled for a May survey as well. 

"The thinking was, I believe, that city employees were excluded to avoid political entanglements in part because Prop. B (pension) and Prop. A (city contracting/project-labor agreements) were on the ballot, in addition to the mayor's race," he wrote in an email.

I also contacted Marjorie Connelly, an editor with the News Surveys and Election Analysis Desk with The New York Times (she also sits on the board of AAPOR) to see how the Times handles surveys and whether excluding city workers is common. Her reply:

I don't know that its common practice. We often do polls on the mayor's race in NYC.  We may ask in the demos at the end if someone works for the city-but not always. And we don't ask if they work for the media or PR. We ask if they live in NYC. And then if they are registered to vote, we ask the election related questions .  

Actually I have never heard of excluding city employees from a poll-they can vote, can't they?  

Voice of San Diego's Scott Lewis also had some questions about the poll and why every other survey we've seen so far has showed Filner ahead or the race tied. Read his piece, and an excellent analysis by political scientist Vladimir Kogan about how the poll was weighted here. Scott's conclusion: "DeMaio's campaign is likely not assuming he has a 10-point lead."

Dave Maass contributed to this report. 


See all events on Monday, Oct 24