Over the past month, many journalists have identified a new development in the presidential campaign: Mitt Romney’s decision to begin making coded racial appeals to the white working class. The Washington Post’s Dana Milbank was one of the first on the case. In an Aug. 10 column, he bashed Romney’s misleading welfare ad, which Milbank described as a bit of race-baiting designed to bolster Romney’s standing with “white, working class men.”

A few weeks later, the theme suddenly exploded: writing on Aug. 27 for The New York Times’s “Campaign Stops” section, Tom Edsall (disclosure: a former professor of mine) described the welfare ad as an appeal to the racial and economic anxieties of “whites without college degrees.” (Edsall found a similarly racialized appeal in Romney’s Medicare ad, with the target in that case being white Medicare recipients.) The same day, John Judis wrote in The New Republic that, “With his recent ads on welfare, Romney is playing on the racial resentments of the white working class the same way Reagan did.” A day later, Peter Beinart described the welfare ad as part of Romney’s “bubba strategy” in The Daily Beast. And a day after that, National Journal’s Ron Fournier accused Romney of “playing the race card,” again in order to appeal to “working-class whites.” These are prominent names in political journalism, and they were reaching a consensus.

But according to several scholars of race and public opinion who have been writing their own findings over the past month, the journalistic consensus is only half-right. Romney’s welfare ad is an appeal to whites’ racial resentment, they say—but it’s an appeal that is most likely to resonate not with working-class voters but among college-educated whites.

Let’s first take the part where the journalists and academics agree—the contention that racial resentment shapes the welfare ad’s effect. Some evidence comes from Michael Tesler, a political scientist at Brown University and the author of Obama’s Race, who penned an Aug. 20 blog post examining the relationship between racial attitudes and viewers’ response to the ad. Tesler found that when people hadn’t seen the welfare ad, their opinions about how Romney’s policies would affect the middle class, the poor, or black people weren’t connected to their level of racial resentment. (For the questions used to measure “racial resentment,” a term of art in the field, see here.) Upon seeing the ad, though, viewers with higher levels of racial resentment thought Romney’s policies were more likely to help those groups—while viewers with low levels of racial resentment now believed the opposite. As Tesler puts it, the ad made attitudes toward blacks “a stronger predictor of respondents’ views about the consequences of Romney’s policies.” In other words, the ad apparently polarized opinion, with more racially resentful viewers thinking better of Romney after seeing it. (Interestingly, there was no parallel effect on views of Obama’s policies—presumably because racial attitudes are already strongly linked to peoples’ assessments of the president.)

Another finding, described last week on the polling site YouGov by Harvard professor Ryan Enos, is easier to explain. The Romney ad’s claim that Obama has “gutted” the work requirement for welfare has been roundly debunked by factcheckers. But Enos found that respondents with high levels of racial resentment were far more likely to believe the claim than people with low resentment levels. Moreover, high-resentment individuals were inclined to believe the claim about Obama whether or not they believed there was a work requirement in the first place. There may be ways to unpack that logical conundrum, but as Enos writes, “more likely what is holding these attitudes towards work, welfare, and Obama together is the common association with African Americans—and a resentment of African Americans.”

That’s exactly how coded appeals work—they give people enough information to bring underlying beliefs to the surface.

But that gets us to the second part of the argument, the one that scholars take issue with. While racial attitudes affect people’s views of Obama at all education levels, the political scientists Christopher Federico, Howard Levine, and Christopher Johnston argue in a Sept. 10 “Campaign Stops” piece that the voters who are influenced by the welfare ad (or the Medicare ad) itself “are more likely to be educated than not.”

That’s because, while better-educated whites generally express lower levels of racial resentment, they pay much more attention to politics. As a result, their racial attitudes become connected to a coherent political viewpoint, and they learn the racial signals embedded in political rhetoric. And so when politicians send a coded message, it is college-educated voters who are more likely to hear it. Which means that—at least according to the academic argument—journalistic accounts that describe the Romney ad as an appeal to the white working class are off the mark.

Though it’s well established that better-educated people tend to have more consistent political ideologies, anyone who remembers the welfare reform debate of the 1990s—a group that includes every journalist mentioned above—will find it hard to believe that the racial code around “welfare” would be obscure to any viewer, regardless of education. I do, too. But Federico and his co-authors present a strong case that opinions about welfare are more strongly shaped by racial attitudes among college-educated voters. And Tesler’s research offers some evidence that that’s true of the Romney ad specifically: over email, he told me that the “increased effects of racial resentment” after seeing the ad “were entirely concentrated among higher education respondents.” For his part, Enos agreed that these findings “just [don’t] seem intuitive.” But, he added, welfare “does seem to behave like nearly every other policy… high-education persons can connect them to other policies that fit coherently, low-education people cannot.”

Does that mean the journalistic accounts are wrong? Well, not necessarily. In his dispatch from the Detroit suburbs, Fournier writes about a firefighter and a contractor who seem pretty well versed in the racial codes of politics. He also reports that a GOP pollster identified a small number of white working-class voters moving from Obama to Romney, and declared it was “almost certainly because of the welfare ad.” The pollster might be right: Even if the academic analysis is mostly correct, after all, there could be regional variation or other exceptions. Or the pollster might be wrong, and just telling a racialized story where one didn’t exist—which would be interesting in its own right.

Even if the academic explanation underestimates the ad’s likely effect on working class whites, though, the research presents plenty of reason to believe that the coded appeal is being heard by college-educated whites—which is a part of the story that the journalistic accounts tend to skip over. As a group, the scholars actually tend to think that Romney won’t benefit much from this dynamic, for a few reasons: because racial attitudes are already built in to views about Obama; because the ad does more to remind people of their attitudes than to change their opinions; because, as Tesler’s research shows, many of the people who read the “code” will think less of Romney; and because the better-educated viewers who can decipher coded signals are generally reliable partisans in the first place. These are all good points. On the other hand, turnout can be just as important as persuasion, and Romney could benefit if the ad energizes his supporters. Based on how often the ad has aired, his campaign clearly thinks it’s boosting their effort.

So what’s the upshot for reporters here? It has been encouraging to see so many journalists, including some at the most mainstream of outlets, call out a campaign for its appeal to voters’ baser instincts. But there is, at the very least, reason to ask whether the coverage has reached too readily for familiar frames about racial resentment among working-class whites—and as a result, distorted the ways in which different groups of voters think about policy, and what the actual effect of the welfare ad is likely to be. It’s a concern journalists should keep in mind.

Greg Marx is a CJR staff writer. Follow him on Twitter @gregamarx.