Taking a Poll…and Some Salt

It’s hard to care much about polls; they’re easily gamed, the questions are usually too narrow to give any real insight, and just because a majority thinks something doesn’t make it right. That said, a new poll from Education Next deserves a bit of comment.


First, I have to repeat a beef I had with last year’s Education Next survey: Why load the No Child Left Behind questions? While the pollsters attempted “survey experiments”—tinkering with question wording to see how it affected results—they just replaced “No Child Left Behind Act” with “federal legislation” in the experimental version of this question:

As you may know, the No Child Left Behind Act requires states to set standards in math and reading and to test students each year to determine whether schools are making adequate progress, and to intervene when they are not. This year, Congress is deciding whether to renew the No Child Left Behind Act. What do you think Congress should do?

The results are pretty damning for NCLB. When it’s identified by name, 50 percent of respondents think the law should either undergo “major changes” or not be renewed at all, versus 42 percent thinking the same way about semi‐​anonymous “federal legislation.” Worse, last year’s results were significantly more positive about the law; the percent of respondents with favorable views of NCLB has dropped by seven percentage points.


Of course, none of this gets to the public’s true opinion about the law because neither version of the question gets rid of the description of NCLB as, essentially, Clarence the angel in It’s A Wonderful Life, intervening to make all schools do well!


So how would the law have fared were people asked what they thought just of NCLB, not “NCLB: The Standardsmaker”? Since I registered this same complaint last year I haven’t seen any polls that have asked about NCLB straight‐​up. But suppose the same changes in NCLB support found by Education Next were applied to the Educational Testing Service poll I mentioned last year, a poll that asked about NCLB unadorned (slide 11 in the link). In 2007, ETS found that only 41 percent of respondents had a “very” or “somewhat favorable” attitude about the law. Drop 7 percentage points from that, and you’re down to a measly 34 percent.


And to think, some people think it’s “foolish” to even consider that NCLB should be scrapped!


Unfortunately, assuming the order of questions in their write‐​up is the same as was presented to respondents, the Education Next folks chose to ask about national standards right after greasing the skids with their encouraging description of NCLB. Not surprisingly, they found that large majorities favored having the feds establish standards and tests for the whole country.


Here we encounter almost all of polling’s shortcomings. For one thing, it’s hard to pin down the effect of the question order, but it certainly seems reasonable to conclude that describing the federal roll as demanding high standards would lead people to conclude that the feds ought to set the standards. But what if the pollsters had described NCLB as a law “that requires states to set standards while safeguarding local control”—which President Bush would tell you it does—or something like that? And what if the national‐​standards issue were explored in some depth, with questions about how the standards would actually be set and what they would be? Suddenly, different thinking would probably kick in. Of course, even with all that it’s possible that a majority of Americans would still support federal standards. Which brings us to the polling problem that majority support doesn’t necessarily mean good policy….


On school choice, the poll offers mixed news: Vouchers keep on struggling, but tax credits seem to have a very bright future. Nationally, only about 40 percent of people support vouchers, versus 54 percent who support tax credits. This is not to say that vouchers are dead—there’s only 40 percent opposition to them as well, meaning you’ve got two evenly‐​matched armies and 20 percent unclaimed territory—but compared to tax credits, vouchers have a long slog ahead. And tax credits fare even better when opposition as well as support is considered; only 28 percent of respondents opposed tax credits.


Of course, wording could have a lot to do with these results as well (for instance, the term “voucher” never actually appears in a question, but the almost as emotionally freighted “government funds” does), and all the other caveats about polls still apply. Even with that, though, the tax credit news, if nothing else in this poll, has to be a little encouraging.