(If you didn’t read my previous post on this topic, you might want to before starting in on the following. It’s fairly wonky. I’ll address Joan Walsh’s similar self-defense in a separate, briefer post.)
Christopher Parker, the lead investigator of the University of Washington poll I criticized in the American Spectator, defends himself against similar criticisms by Cathy Young:
Apparently, Young doesn’t appreciate the capacity of social scientists to take the analysis further. If she did, I doubt she would’ve claimed that support for the Tea Party was simply a proxy for conservatism. Suppose that most Tea Party supporters are what some call “principled conservatives.” That is, they’re simply about a small federal government, fiscal discipline and free markets. In fact, our data show that when you account for/control for conservatism, and partisanship to boot, there’s still a strong statistical connection between support for the Tea Party — rather than conservative politics generally — and racial resentment. Indeed, ideology does matter: If one is conservative, he or she is 23 percent more likely than a liberal to hold racially resentful attitudes. Even so, we found that Tea Party supporters are even more likely than conservatives who don’t support the movement to believe that blacks simply need to work harder, and that the legacy of slavery and discrimination has no effect on blacks’ current condition in America society.
First of all, I’m a little bit baffled by the idea that Young should have just trusted that as a “social scientist,” Parker had examined the data in detail, “tak[ing] the analysis further,” before releasing it and letting liberal journalists go to town. As I detailed in the Spectator, he plain and simply did not do this: He released an incriminating-looking chart with a few data points on it, and only later — after requests by at least three different journalists including Young and yours truly — analyzed and released other data (such as the numbers concerning what white tea-party supporters thought about other whites). In fact, until seeing this article from Parker, I was not aware the survey had even included a question about conservatism directly — so far as I can see, none of the charts he’s released to this point include any numbers from it.
Further, I’m not convinced that Parker’s methodology does what’s demanded of it. He is responding to the allegation that his “racial resentment” questions (example: “It’s really a matter of people not trying hard enough; if blacks would only try harder, they could be just as well off as whites.”) measure not only racial sentiments (blacks are lazy and we shouldn’t help them) but also a right-of-center political belief (disadvantaged people of all colors can and should work harder to improve their lot). To fix this problem, he tries to adjust the data to account for this political belief — if this can be done, it will leave only the difference in racial sentiment.
The problem is that he doesn’t control for this belief specifically. Instead, he controls for overall ideology and partisanship — and to whatever degree these measures don’t correspond to the specific belief he’s trying to control for, he fails in his task. (Though he doesn’t provide the questions he used, I assume he simply asked respondents how conservative/liberal they are and also whether they identify as Republicans or Democrats.)
Conservatism and Republican ID are indeed correlated with the belief that poor people should help themselves, but they are nowhere near perfectly correlated — even at the extremes. For example, the General Social Survey often asks respondents to rate, on a scale of 1 to 5, whether government action or self-help should improve life for the poor. Going by data collected from 2000 to 2006, even someone who calls himself “extremely conservative” has a 23 percent chance of favoring the latter solution (answering 1 or 2). For “strong Republicans,” the number is 12 percent. Combining the two to create the most extreme measure possible — looking only at people who call themselves extremely conservative strong Republicans — brings the number below 5 percent, but even in this group, 29 percent of people say they “agree with both” rather than saying poor people should help themselves.
In other words, even a third of extremely conservative strong Republicans don’t agree that poor people should help themselves instead of getting help from the government.
So, it’s very possible that Parker’s conservative Republican tea-party supporters were more likely than his conservative Republican tea-party opponents to support self-help over government help — indeed, since the stated purpose of the tea parties is to protest government economic intrusion, I’d guess that’s the case. It is further possible that this difference, rather than racism, caused the response patterns Parker finds so damning. This is the fundamental problem with using vague proxy variables to “control for” very specific opinions.
A good way to control away differences in opinion would have been to ask the same question without a racial component — “It’s really a matter of people not trying hard enough; if the poor would only try harder, they could be just as well off as everyone else.” — and compare the answers to those of the racialized question. (Maybe they did ask this question, and we’ll find out weeks from now; who knows what data still lurks up the sleeves of the University of Washington Institute for the Study of Race, Ethnicity & Sexuality?)
An even better way would be to ask questions that measure racial sentiments without also measuring political beliefs to begin with, and in fact the survey included some: Respondents were asked whether they thought blacks were hardworking, trustworthy, and intelligent. But as Young pointed out, these questions yielded mixed results once you take into account what each group thought about whites.
Parker continues:
It should also be noted that the support for small government and “states’ rights” expressed by Tea Partiers have rarely been entirely free of racism, or racial motivation. Dating back to before the Civil War, the South leaned hard on the doctrine of states’ rights as a means of preserving slavery. During the 20th century, claims about states’ rights animated the Dixiecrats, who had fallen out with Truman’s Democratic Party as it installed a more racially liberal party platform. Indeed, research conducted almost 30 years ago, by David Sears and Jack Citrin, suggests that support for small government and tax revolts are motivated by some whites’ desire not to have their taxes redistributed to help blacks and other minorities.
I suspect that the leftists who buy this argument would like it less if it were applied to, say, gun control — which historically has been used to keep blacks from arming themselves, as Stephen P. Halbrook explains in Securing Civil Rights. Or abortion, which historically has been advocated as a way of controlling the black population. A policy’s use for racist purposes does not make the policy inherently racist for decades, nor does it make support for the policy code for racism.