Earlier this month, the University of Washington’s Institute for the Study of Race, Ethnicity & Sexuality released some data from a survey it had conducted. In seven states (Georgia, Michigan, Missouri, Nevada, North Carolina, Ohio, and California), the Institute had asked white people whether they supported the tea parties, and also whether they agreed with a variety of race-related statements — such as that blacks and Hispanics are hardworking, trustworthy, and intelligent. Forty-five percent of those who said they strongly supported the tea parties said they agreed that blacks were intelligent, for instance, compared with 59 percent of strong tea-party opponents.
Left-wing media outlets, most notably two different writers for Salon.com, jumped on the numbers. The Institute itself, apparently unfamiliar with the difference between correlation and causation, claimed the results suggested that tea-party supporters are “motivated by more than partisanship and ideology.” (That statement has since been removed from the Institute’s website, but a blogger’s quote of it is available here.)
But almost immediately, skeptics raised some questions about the results the Institute had provided — they pointed out, for example, that only the views of strong supporters and strong opponents of tea parties were included, and that the Institute’s table didn’t provide the full range of responses to the questions about blacks and Hispanics. Relative to the average white person, did tea-party supporters express more racism, did tea-party opponents express less, or both? Did the people who didn’t “agree” with these statements actively disagree, or did they refuse to express opinions?
It took requests from three different journalists (that I know of), but at last, the truth is out: The Institute’s full data set provides no evidence whatsoever that supporters of the tea parties are disproportionately racist.
First up was the politics website FiveThirtyEight.com.
In response to a request from the site’s Tom Schaller, the survey’s lead investigator, Prof. Christopher Parker, provided this table — which doesn’t even include the questions about industriousness, trustworthiness, and intelligence. Instead, the questions measure what Parker calls “racial resentment” — though they can reasonably be called transparent attempts to bait conservatives into giving “racist” answers. Here’s an 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.
Mainstream conservatives believe that people of all races can improve their lot in life through hard work, and further that most poverty is the result of behavior. (These are not crazy beliefs: Those who finish high school, work full-time, and marry before having children rarely end up in poverty (pdf) and usually end up in the middle class.) When presented with a statement that racializes that sentiment, should conservatives decline to agree? Maybe, but liberals — who believe that structural barriers, more than behavior, affect people’s outcomes in life — face no such tradeoff, so any gap between liberals and conservatives could represent mere differences in political beliefs.
So, I asked Parker for more data about the questions that were in the original table. He provided some (pdf). These results show that the more a white person supports the tea parties, the more likely he is, when prodded by a pollster, to disagree with statements that express positive views about racial minorities.
But a key fact about the survey had yet to enter the discussion: Respondents were also asked whether whites are hardworking, trustworthy, and intelligent. Parker provided the details to Cathy Young of RealClearPolitics, who wrote:
While only 35% of strong Tea Party supporters rated blacks as hardworking, only 49% described whites as such. While the gap is evident, these responses are close to those for all whites (blacks are rated as “hardworking” by 40%, whites by 52%). While whites who are strongly anti-Tea Party seem free of bias on this item — blacks and whites are rated as “hardworking” by 55% and 56%, respectively — this is not true for intelligence and trustworthiness. Whites in every group are less likely to rate blacks than whites as “intelligent” by similar margins: 14 points for Tea Party supporters (45% vs. 59%), 13 points for all whites (49% vs. 62%), 10 points for Tea Party opponents (59% vs. 69%). On “trustworthy,” the gap is smaller in the pro-Tea Party group (41% vs. 49%) than in the anti-Tea Party group (57% vs. 72%). One could write headlines about the “racial paranoia” of white liberals who consider blacks less trustworthy than whites!
So, the question is: When the Institute released its initial summary of the results, why did it include results for opinions about blacks and Hispanics (which suggest disproportionate racism among tea-partiers across all questions) but leave off opinions about whites (which show a completely different trend)?
In a phone interview, Parker explained that he simply hadn’t been able to analyze that portion of the data yet — when Young “pressed” him for the additional numbers, he had a grad student put them together, and Young had to wait for “days.” “You’ve got to understand, this is a huge data set, and I’m the only one working on it,” he said. “The question at the time was whether tea-party supporters are racially intolerant, so I started right in on that.”
Of course, the new data — especially when seen in the light of the Institute’s speculation about what “motivated” tea partiers — raises questions about bias. “First and foremost, I’m a scientist,” Parker said. “The data is what the data is, whether I agree with it or don’t agree with it.”