Verified by 
Key points
- Debates remain whether color-blind or race-conscious policies are better for reducing racism.
- One TED Talk by a black scholar supporting color blindness faced unusual scrutiny and attempted censorship.
- A meta-analysis used to defend TED's actions provides little evidence color blindness is ineffective.
- Color blindness may yet be superior to race-conscious approaches such as DEI or implicit bias training.
A brouhaha broke out recently when it was revealed that TED treated a talk on color blindness by Coleman Hughes (who is black, if that matters), with surprising levels of hostility. Mr. Hughes and TED seem to agree on the broad outlines of what happened. Mr. Hughes argued in favor of color-blind thinking; this offended some staff at TED, resulting in delays, unusual scrutiny, and alleged failure to promote Mr. Hughes’ video. This raised familiar concerns that institutions are throttling free speech and rigorous debate in the name of satiating a few easily offended individuals.
One defense raised by TED for their actions was that Mr. Hughes’ defense of color blindness wasn’t scientifically grounded. They pointed to one particular meta-analysis, which analyzed the impact of color blindness and race consciousness in the form of multiculturalism on several outcomes. TED argued this meta-analysis questions the value of color blindness. Mr. Hughes countered that this meta-analysis actually supports color blindness. So, which is it?
I decided to have a closer look at the study in question. Upfront, it’s worth noting one thing: as I’ve argued before, meta-analyses are poor debate enders. Generally, they tend to artificially smooth over inconsistencies in the data (which the authors of this meta-analysis themselves acknowledge), they tend to overestimate support for hypotheses, and researcher choices can produce divergent meta-analytic conclusions. But, for the moment, we’ll ignore that.
Although the authors purport to compare color-blind to race-conscious approaches, I’m not convinced they actually isolated this. It’s well known that many race-conscious approaches, whether in DEI training, implicit bias training, microaggression awareness, etc., either don’t work or backfire. A thorough discussion of this appeared to be missing from the manuscript.
Instead, the authors focus on a vaguer concept of multiculturalism which they define as “acknowledging differences by learning about, maintaining, or valuing them.” This seems to be a very soft version of identity consciousness, not nearly as likely to concern people as, say, the white fragility concept or segregating school kids by race for affinity groups. I’m not sure their definition of multiculturalism is even contradictory to color blindness.
The authors consider four outcomes: prejudice, discrimination, stereotyping, and what they call “policy support,” which they define as “positive attitudes toward policies aimed at increasing diversity by granting resources to non-dominant groups.” This later category, while undoubtedly good in intention, also appears to be explicitly racially discriminatory (as with affirmative action) and could reasonably be expected to be opposed by color blindness in good faith.
Some of the meta-analytic results appeared to be underpowered, particularly for studies of color blindness. Focusing on effect size, and consistent with Mr. Hughes' read, color blindness tends to be associated with more positive outcomes, some more strongly than others. So was multiculturalism, with the exception that it was associated with more “neutral” stereotyping (a tendency to see groups as culturally different—an outcome I’d argue is actually rather negative for a multiethnic society).
Taken at face value, I think the authors could argue multiculturalism had a stronger impact on some outcomes such as prejudice and discrimination, but we’ll return to that face value in just a moment. Yet the main difference was regarding policy support. Color-blind approaches were associated with less policy support, whereas multiculturalism was associated with more policy support.
Whether that’s good or bad is a subjective evaluation—likely depending on whether the reader likes those policies themselves. I’m concerned that this has been something of a common rhetorical trick by progressive scholars. For instance, some economists claimed that 2016 Trump voters were motivated by “racism,” though the surveys they used for this claim tended to involve disagreement with specifically progressive views of race, such as support of affirmative action or belief that racism is widespread, rather than actually endorsing hateful views of others.
Of course, there are also reasons not to take the meta-analytic results at face value. As mentioned, the authors used a rather vanilla definition of multiculturalism, and deftly avoided the most controversial areas of race-consciousness. This certainly affected effect sizes. Moreover, from the start (including the chosen quotes by Justices Day-O’Connor and Sotomayor at the beginning of the article), I had the sense that the authors really wanted multiculturalism to win, and this may have influenced their interpretations.
In short, there are reasonable debates about the meaningfulness of the comparisons between multiculturalism and color blindness. But there was nothing in this meta-analysis that should have been used to conclude Mr. Coleman’s TED talk was unscientific or contrary to the (very weak) scientific evidence.
