Willie Colón’s salsa beats crackle over the speaker, his cadence dipping and dancing over the heavy brass. A new hit on La Mega 97.9? Nope, it’s the loop that callers hear when on hold at Mike Bloomberg’s campaign headquarters, encouraging them to vote for the mayor.
In his quest to squelch opponent Freddy Ferrer’s largest support base, Bloomberg has been pursuing the Latino vote with a vengeance, and the tune from Bronx-born Colón (who is, like Ferrer, of Puerto Rican descent) is one stratagem among many. According to recent polls, his efforts have not been in vain.
Ferrer’s numbers in both the Quinnipiac and Marist polls showed, at best, no bounce following the September 13 Democratic primary. In Quinnipiac’s poll of likely Democratic primary voters, Ferrer got a predicted 63 per cent of the Latino vote; in the most recent poll of all likely voters, he had 57 percent of the Latino vote; Bloomberg took 31 percent, indeed attracting a significant chunk.
Marists’ director, Lee Miringoff, observes that, in the current Mayoral race, “some people think the Latino vote is out there and not being captured.” There are several reasons to believe that to be the case.
Both Quinnipiac and Marist rely on random digit dialing, in which telephoned respondents answer a series of multiple choice questions, and pollsters then determine the likeliness that an individual will vote and for whom. An alternative way to conduct a survey is to use a list of registered voters. Each method has its weaknesses.
Don Green, professor of political science at Yale who has conducted several studies on polling techniques, describes the choice between the methods as a “balancing of alternative biases.” For example, random digit dialing would weed out “low information voters,” while a registered voter list would eliminate all unlisted phone-numbers, inadvertently skipping over extremely low and high income voters. Jerry Skurnick, who makes a living selling registered voter lists, acknowledges that, “Public polls always have problems with minorities and in minority neighborhoods.”
Pollsters also face the challenge of obtaining views from non-English speakers. Mostly student interviewers for both Quinnipiac and Marist polls make calls from their respective suburban campuses, which boast a 4 and 6 per cent Hispanic population, respectively. Quinnipiac’s interviewing team of 155 includes six Spanish speakers, who, though screened for fluency, were not actively recruited by the Institute. Marist’s numbers are comparable.
Another stage in a random digit dialing polls, referred to as “weighting,” allows the pollster to adjust the percentages in a group of informants to match the local demographics according to the census. So, if blacks make up 20 per cent of the population but only 15 per cent of a poll sample, the pollster would give more weight to existing black responses. But Green notes that a problem arises here “when the blacks you’re talking to are not representative of blacks.”
Not surprisingly, opinions vary regarding how the Hispanic community will vote. “My feeling is that Ferrer’s going to do better in the Hispanic and black communities than these numbers show,” says Maurice Carroll, director of Quinnipiac’s Polling Institute. “[The polls] are underestimating the zeal for Latino presence.” But he adds, “The Latino vote here will not be as high as the black vote was for Dinkins, it’s not as monolithic as the black community.”
Miringoff, however, feels that “the Latino community is itself quite diverse and isn’t going to fall behind one candidate. People don’t vote based just on ethnicity.” Bloomberg’s band of Hispanic supporters-from former Bronx President Herman Badillo to the Latina Political Action Committee-reveal the diversity of the Hispanic opinion.
Making things even more murky is the fact that no exit polls were conducted on primary day, leaving experts a bit in the dark trying to evaluate the actual ethnic composition of Ferrer’s vote. Ferrer’s own pollster, Jeffrey Pollock, went through the back door to evaluate returns in particular districts, an oft-used method in the absence of an exit poll. Experts admit that polls are not omniscient. Miringoff says, “clearly, six weeks out, theirs is more art to this question than science.”
Pollock remains optimistic after his own analysis of the primaries, in which he says he saw “no reason to think the Hispanic community would be any less excited than they were in the 2001.”
“The bigger question is: ‘What’s going to happen with these numbers?'” he says. “When Freedy goes with his media campaign is where the numbers begin to move. It’s all about the closing two weeks.”