5 Real Reasons to Ignore Pew’s Arizona SB1070 Poll

Big new polls show majority support for Arizona's new immigration law. Don't believe them.

By Channing Kennedy May 13, 2010

As opponents of immigration will happily tell you, polling companies Rasmussen and Pew have both released surveys around Arizona’s SB 1070, the results of which show support for a law which many have said unconstitionally legalizes racial profiling. But is it that simple? Rasmussen’s in-Arizona poll has already gotten pushback for its methods, and for the company’s right-leaning track record. Pew’s got a good reputation, but its new national poll, conducted last week and showing "broad appeal" (59% approval) for the law, has its problems too — both in execution and concept. Here’s the grains of salt you should be taking with it. 1. Not enough cell phones were called. From Pew’s statement on methodology, "a national sample of 994 adults living in the continental United States, 18 years of age or older, from May 6-9, 2010 (662 respondents were interviewed on a landline telephone, and 332 were interviewed on a cell phone, including 123 who had no landline telephone)." Pew itself has done research showing the age-based slant that landline polls exhibit — and on a race-heavy issue like this, we’d do well to remember that people of color, especially Latinos, are some of the biggest cell phone users. How would this survey have turned out if they’d called 994 cell phone numbers? Who couldn’t pay the phone bill this month, and didn’t get called as a result? How many people do you know who only have a landline? 2. Some people have jobs, school, kids, second jobs, and other reasons not to be at home. Some people don’t. Who’s at home in the middle of the day, has a land line, and answers it to talk to a stranger? Retirees. Who watches Glenn Beck? Ding ding ding. I’m not saying your elderly uncle back in Oklahoma is a racist; I’m just saying you delete his email forwards without reading them, so you can be civil at Thanksgiving. 3. The survey was only conducted in English! Self-explanatory, and pretty unforgivable on a topic like this. This scrambles the numbers for Spanish-speaking households across America, of any ethnicity and status, as well as knocking out big chunks of multi-language urban centers like New York, San Francisco, and Minneapolis. Arizona itself spoke more Spanish than English until the 1940s. For that matter, why poll on a subject like this without breaking down the results by race? 4. Phrasing matters. Remember the poll from a couple months back in which respondents supported gays in the military, but not homosexuals? (Yes.) In this survey, Pew asked people if they supported "… allowing police to question anyone they think might be in the country illegally." What if they had changed "illegally" to "without documents"? 5. and the big one: So what? I’m not a pollster or a statistician, and my comments shouldn’t be taken as anything more or less scientific than common sense. (If you’re a professional in this area, I’d love to hear from you.) And I know that good pollsters understand the weaknesses they’re up against, and use tested methods of making their work as relevant as possible. And of course, no one poll tells the whole story. Here’s one saying that support in Arizona for SB1070 has dropped from 70% to 52% in just a few weeks. Here’s a poll saying that 70% of Latinos oppose the law. I haven’t seen a poll that compares against political engagement, or familiarity with the issue. But ultimately, referendums and polls can’t inform issues of human life. The rights of the few are not to be determined by the will of the many. The ideal that guides our society, cynicism aside, is true equality for each member, regardless of frequency of occurrence within a given sample. Japanese internment camps probably would have polled really well ("Agree or disagree, a nation must protect residents at risk of remote brainwashing by Charlie"), along with the Trail of Tears, Jim Crow, Operation Wetback, the Patriot Act. These initiatives all looked sensible on paper, because they were designed to. They all happened with rational people watching. And the populations they targeted were small, i.e. "minorities." Which of these small, targeted stains, do you think, will wash clean first? This Pew poll is something of a gift to those with dogeared Heritage Foundation pamphlets, because it lets them move the conversation away from other, less convenient numbers, and to scare politicians to the right in the 2010 election with phrases like "out of touch with real Americans." And so, title of this post aside, this poll can’t be dismissed out of hand. Understanding how it was made, and what it says to whom, is a step toward making ‘vulnerable’ a subset of ‘human,’ and in popularizing justice.