A big part of Sarah Palin’s appeal has always been that she had the right enemies. Palin’s resentment of sneering East Coast elites, one of the subtexts of her bizarre resignation speech last Friday, has been one of her defining characteristics since she burst onto the national stage last fall—and, by all appearances, one of the big draws for her supporters.

Ross Douthat, in his column in today’s New York Times, seizes on that dynamic and argues, plausibly, that in contrast to our overachieving president, Palin’s popularity comes from the fact that she “represents the democratic ideal – that anyone can grow up to be a success story without graduating from Columbia and Harvard.” But when Douthat tries to extend his argument—claiming that Palin’s appeal “extends well outside the Republican Party’s shrinking base,” and that “her popularity has as much to do with class as it does with ideology”—the column becomes an object lesson in how to manipulate polling data.

Douthat’s argument hinges on one data point from a recent poll by the Pew Research Center: 48 percent of respondents without a college education view Palin favorably. But his analysis obscures as much as it illuminates. First, as with all stories based on polls, it’s best to broaden your sample to minimize the effect of outliers. (This is one of the lessons of FiveThirtyEight and similar Web sites, which use an aggregate view of polling data to predict election results with impressive accuracy.) But more to the point, Douthat’s position isn’t substantiated by the Pew poll—as Pew’s own interpretation of the data makes clear!

Where to start? For one thing, Palin’s favorable/unfavorable split among the general public is 45 percent to 44 percent. That makes her pretty much the definition of “polarizing”—but it also means that her support among the general population (45 percent) is pretty much identical to that among those with less than a college degree (48 percent). When you account for the likely imprecision of the latter number, which is based on a subset of the 1,502 respondents in the entire poll, there’s not much evidence for above-average Palin support among working-class voters. In fact, the last time Pew asked this question, in mid-October, it found exactly the same level of support for Palin among those with a college degree or more and those with a high school diploma or less: 41 percent.

So who provides Palin’s base of support? As Pew notes right in the headline of its write-up of the poll, the “Republican base.” Eighty percent of conservative Republicans, and 84 percent of white evangelicals, view her favorably. Among Republicans, she trumps Newt Gingrich, Michael Steele and even the rebounding Mitt Romney with a 73-17 favorability split. If that’s not an ideological base of support, what is?

As Pew notes, there’s nothing new in these findings. A Gallup poll conducted just after the election found a sharp partisan divide on views about Alaska’s governor, with less than half of the general public wanting to see her become a major national political figure. There is also evidence suggesting that unlike most vice-presidential nominees, who are electoral non-entities, Palin may have had a real—and detrimental—effect on the Republican ticket.

Douthat may be motivated to overlook this information because of his ongoing program to rebrand the Republican Party and make it more appealing, and responsive, to working-class families. There are good things to be said about that effort—but in Sarah Palin, he’s picked the wrong symbol.

If you'd like to get email from CJR writers and editors, add your email address to our newsletter roll and we'll be in touch.

Greg Marx is an associate editor at CJR. Follow him on Twitter @gregamarx.