Saikat Chakrabarti, candidate for Congress, is promoting a poll that puts him in second place behind State Sen. Scott Wiener in the top-two primary.
The Chakrabarti campaign paid for the poll, and candidates release internal polls to the press for two reasons: They want to convince donors that they have a chance of winning, or they want to convince donors and potential supporters of other candidates that they’re going to wind up on top.
So let’s look at the poll for a second.
First: It’s very hard to get accurate polling data these days, particularly from a relatively small sample (547 likely voters). Polling relies on a truly random sample, which was easy when everyone had a landline and answered the phone when it rang. Now, pollsters try to reach voters on cell phones, and most young people don’t answer their phones at all, particularly when they don’t know the caller, and from texts, which again have tremendous sampling bias.
The poll was only in English, and some 19,000 voters in the district have requested Chinese-language ballots (and another 4,000 have Spanish-language ballots). Sup. Connie Chan, who was ranked well below the two guys, will probably get a significant number of the Chinese-language voters.
The poll claims a margin of error of four percent. Sorry for the math, but: That means Wiener is somewhere between 29 percent and 37 percent. Chakrabarti is between 24 and 32. Chan is between 9 and 17. So it’s entirely possible that the real numbers are Wiener 28, Chakrabarti 24, and Chan 17, with a lot of undecideds. Again, this counts only the English-speaking voters; if more than half of the Chinese language voters break for Chan, she’s at 20 and Chakrabarti is at 21.
With 527 voters, I’d say the margin of error is probably higher than 4 percent.
I know of no independent polling in the race.
Chan’s campaign points to a few problems with the poll, including text messages that identified Chakrabarti as the state Senator.
But it demonstrated a few key things.
If you compare this poll to polls taken a few months ago, before Chakrabarti started pouring more than $1 million of his own money into ads and paying more than 200 canvassers to knock on doors, it appears he has picked up about ten points. Roughly half of that came from Wiener, and the rest from “undecided.” Chan is about where she was.
Wiener’s big-money donors clearly don’t want to run against Chakrabarti in the November election, since he has almost as much money as they do. That’s why we’re seeing so many attack pieces targeting Chakrabarti and aimed at young people.
But those attack pieces apparently haven’t been very successful. It’s surprising the Wiener IE, which has gobs of money, hasn’t spent more—unless the polling indicates that the attack messages aren’t working.
Chakrabarti tested them and his possible responses in the poll:
The real race at this point is not about which candidate is more popular; it’s about who votes. There’s no presidential primary, and the primary for governor of California is such a godawful mess that some progressives might sit it out (who are you voting for? Is there anyone any of us on the SF left are walking precincts or carrying signs for? I’m not seeing it.)
But there are deeply contested supervisor races, with a lot of money and voter outreach, in D2 and D4. D2 is conservative, a place where Wiener has been popular in the past—but a lot of residents don’t like the fact the Wiener has made the gigantic Marina Safeway project possible. If Chan and Chakrabarti can get that message out in D2, it could undercut Wiener.
D4 is mixed, but has a lot of Chinese voters and union members, which could help Chan. Mayor Lurie and his big-money operation are supporting Alan Wong, and several other serious candidates are canvassing for voters—and none of them are openly supporting Wiener, who backed closing the Great Highway.
I’d like to see an independent poll of 1,000 likely voters, in Chinese, Spanish, and English, adjusted for where we’re likely to see the highest turnout. Papers like the Chron used to do that, but it’s a $25,000 nut, and they no longer care.
So let’s take all these polls for what they are—a snapshot, not always a good one, of a moment in time with a select number of voters, released by campaigns for political reasons.
I’m not saying it’s wrong. I’m just saying it’s not necessarily right.
Full disclosure: My daughter works on the Connie Chan for Congress campaign.
