Global Warming Caused by Racism

MITHRIDATES
You're joking, right?

Yeah, you caught me. Racism doesn't cause global warming. It just makes people fat. At least that's what a new study by Boston University researchers claims.

There are several unassailable truths in the world, and among those are:

• Racism exists and can have serious consequences
• Social scientists publish studies overstating the impact of racism
• Journalists write articles about these studies overstating the studies' own conclusions

More...
Now, of course, not all of these axioms hold true all of the time. Not everyone is a racist (at least not all the time); sometimes a study employs sound methodologies, the data are fairly presented, and the stated conclusions are appropriate; and sometimes journalists get it right. But let's take a look and see how the scientists did on this one. And the point is not to pick apart the study over irrelevant details. Real science is imperfect, after all. We're just trying to see if the conclusions are way out of line with what the data suggest.

• First, let's examine the proposed correlation between perceived racism and increased weight gain. As far as I can tell (and one always has to trust the authors a bit), the statistical analysis is sound. The ordinal trends appear to be statistically significant for "everyday" racism. But the 95% confidence intervals for the highest and lowest quartiles do overlap and the "lifetime" racism trends were insignificant for each Body Mass Index group. So there does appears to be some correlation for "everyday" racism, but the evidence is far from overwhelming.

• Second, lets tackle the ubiquitous "correlation vs. causation" issue. E.g., did CNN liken the black woman to a turtle because she's obese; or is the black woman obese because CNN likened her to a turtle?

  Seriously, one of the good things the researchers did is to make this a prospective study and get temporal correlation the right way. But that does not give us causation. There still could be a (very plausible) confounding variable.

  The data only concern "perceptions" of racism. It's possible that "perceptions" of racism do contribute to stress that, in turn, adds to weight gain. But this doesn't stop the authors from boldly declaring in their conclusions that "weight gain increased as levels of everyday and lifetime racism increased" and "experiences of racism may contribute to the excess
  burden of obesity in U.S. black women." These pronouncements despite the fact that there were no data about actual experiences of racism, just perception. It may be that the increase in perception of racism was attributable to increased levels in actual racism. But might women who are more insecure or willing to blame others for their problems be more likely to perceive racism and be less responsible about their health? Of course, this isn't the conclusion those of raised with bleeding hearts want to hear. It might not be true, but it's at least as plausible as "perceived racism makes people fat." Let's not bend the science towards our pre-conceived conclusion.
  

• Finally — and this is the big one — suppose you believe there is a correlation, that perceptions of racism translate directly to actual incidence of racism, and that the causality is correct, in that increased racism causes an increase in weight gain. The lowest quartile of study subject for "everyday racism" gained 11.2 pound, and the highest quartile gained 12.5 pounds. That means a baseline gain of 11.2 pounds and an extra 1.3 pounds gained from the lowest to highest quartile. There's a bit of wiggle room here for extrapolating this result to lower levels of racism beyond what anyone in the study experienced — but far better to add some white women to the study for that.
  
  But there's not even any wiggle room with "lifetime racism." Those who answered no to all the "lifetime racism" questions gained 11.6 pounds compared with 12.6 pounds for those who answered yes to all of these questions. In this case we can say that a typical study subject who suffered no "lifetime racism" would gain 11.6 pounds — whereas a subject who experienced the highest level of "lifetime racism" would gain one extra pound. That means only 1 out of 12.6 pound gained is attributable to racism! And that's if you swallow all the assumptions and analysis wholesale (which, as you might have guessed, this author doesn't).
  

Now, surprisingly, the journalists actually did OK on this one. Sure, there's the usual selection bias in promoting this story prominently over other issues. So they fail by putting the race story on the front page, because "race sells." But at least the Globe's headline, "Perceived Racism Linked to Weight Gain," more accurately states what can be fairly concluded from the data than the authors' own claim, which overlooked the gap between perceived and actual. Usually the journalists take a point made by scientists and blow it out of proportion; so let's give some credit to the journalists who actually rein in the scientists . . .

Actual racism exists. But its real, pernicious effects can be drowned out by the absurd headlines about "scientific" studies blaming it for all of society's ills. This author sympathizes, but so far his few experiences with racism — when the waiters at a certain 55th street establishment don't come by to ask him if he needs anything else — actually make him thinner.