Tuesday, 16 September 2008

Race and research

Goodness! I meant to write something yesterday, but a long day at work (is there any other kind?) and work at home had me busy. I want to do a post on trapping, soon, since that's part of what has had me busy (Don't worry, ADF&G! Just getting ready for 1st of Nov!).

Larry Moran from Sandwalk asked about censorship in research that has racism implications. Apparently I've had something to say about this, because I wrote a page and a half easily, before realizing I never answered his question. He got the pared down version, and you all get the full!

That there's genetic correlates to race is hardly surprising - after all, there's heritable phenotypic differences between them! However, as someone who infrequently dabbles in population inference (one who's running Structure in the background even now), what I see is a disconnect between what is being posited, and what is being proven.

What I see demonstrated in the literature is that there are populations of humans who more closely resemble each-other - and frankly, if that wasn't the case, and humans existed in panmixia, I would have just been floored. The differences in allelic frequencies are robust enough that when you saddle enough loci together, you can accomplish the prodigious feats like the recent European structure paper. What is also clear is that the `classical` races (I use that loosely, as I know the definition of each race has changed over time) are composed of many of these populations, and some of which may not closely resemble others they've been lumped in with (e.g., the Finns, grouped in with the Europeans).

You could possibly do some bootstrapping to get some level cluster support - I don't know if anyone has beyond mtDNA. I suspect they have, and it would be valid. But even assuming they fall into line, you're still faced with the same question wildlife researchers are - does this cluster represent an effective management unit? Should we throw our net wider, and group others in with this clade, or should we divide it up into separate units. That something forms a natural clade with others does not mean it is especially useful to treat them the same for some or most things.

And it's a long distance from saying 'behold, a naturally forming unit!' to 'behold, a relevant unit.' Which is what is purported by most people putting forward the notion of race. When people go as far as declaring Race=Subspecies, and we've found valid races, I feel very confident in saying that although we don't have a recognized definition of subspecies, compared to some other declared subspecies in other species, human `races` don't represent the same magnitude of differences.

And there in-lies the problem. We as scientists have let things jump the shark from `we have naturally occurring populations` to `our preconceived notion of race is a) very correct and b) functionally significant.` A lot of that gets lost in the translation between the scientists (of whom I believe, or at least hope, know that there's a long distance between what we've shown, and showing that race is relevant) and the science media/consumer of information (who doesn't know that). Given the nature of how we openly conduct science, censorship won't work - reporters will pick up on articles, and run with them.

So instead of censoring our work, I think there's an ethical obligation for researchers in this field to devote a large portion of their time putting their research in the exact context. They, and other people who engage in research that has the potential to really muck up people's lives, need to make the significance and the non-signifigance of their research exactly clear.

No comments: