The Society for Ethical Data

The other day Joshua Tucker asked a hard question:

Nevertheless, it does seem that a significant number of Americans always want to come back to "24" style scenarios, where a ticking bomb is about to blow up an American stadium somewhere and some terrorist holds the code to defusing it but won't talk. Even absent this kind of extreme scenario, there is a belief - not exactly discouraged by Dick Cheney - that torture could help save American lives by revealing information about terrorist plots. There are a lot of arguments floating around in the blogosphere and media now about whether torture does in fact yield valuable intelligence information, and as I read them I thought, wow, here's a place where good social research with modern methodological tools could really make an important contribution to a policy debate (which, not coincidentally, is one of the goals of this blog).
But here's where things get complicated. My original thought was that good social science research that shows that torture does not extract useful intelligence information would be the final nail in the coffin in any public argument in support of torture. But what happens if one of us gets access to the relevant data, does the empirical analysis, and then discovers the opposite: that torture does lead to useful intelligence information. What do you do then? Sit on the results? Would any political science journal publish such a paper? How would that look in a tenure review? ("Right, she's the one who said torture was valuable . . . ")

Dan Drezner gives an unequivocal no:

I, too, would welcome good empirical findings showing that torture does not work, but my answer to Josh's questions are "no."  You have to publish your findings regardless of what you discover.  That's the only way this business can work. 

From a practical perspective, it makes little sense.  Uncomfortable findings, if they hold up, will get discovered by someone.  Sitting on them merely magnifies their impact.  One of the few currencies social scientists can use is their research integrity.  A short-term compromise of this integrity simply magnifies the impact of the discovery.

From an ethical perspective, social science results do not upend ethical arguments for or against a particular issue.  In other words, even if torture works in extracting information, there are strong normative reasons to oppose its use.  Covering up results, however, does compromise the ethical position of the person making the anti-torture argument. 

In theory, I am entirely with Dan:  the spirit of free inquiry should not be compromised--which is sort of why I argued that people should not employ effectiveness arguments against torture.  When an important moral debate rests on the outcome of an empirical question, the study of that empirical question gets compromised.  You haven't resolved the debate; you've just produced another round of dueling crap science.

In practice, however, the end of Tucker's second paragraph has a lot of weight.  The guy who finds out that torture works is not going to have a happy life either in or out of academia--particularly if his work is used to justify something he finds morally repellant.

So what we get in practice on a lot of tough questions is that the people who are willing or able to do objective research on a question bow out, leaving the people who are only interested in finding (or publishing) one of two possible answers.  The uncomfortable results don't get discovered . . . by anyone credible.

The few conversations I've had about the Bell Curve with professionals who work in cognitive sciences indicate that this is why most of the work about race and IQ seems to be written by crypto-racists with an axe to grind.  Given what we know about evolution and cognitive science, it is possible that there are real and heritable differences between genetically isolated groups (and just as possible that there aren't.)  What reputable scientist wants to risk being the guy who found credible evidence of a persistent, heritable, IQ gap?  Moreover, given all the interesting questions there are to study, why on earth would you pick this one unless started out fairly determined to prove either that blacks are genetically handicapped, or that they aren't?

Is this a big problem?  Well, for one thing, both sides of these debates end up over-reliant on poor quality evidence that allows the other side to reasonably proclaim that they are not arguing in good faith--it makes it harder to solve either the empirical or the political question.  Indeed, the mere fact that Tucker is willing to ask this question will no doubt confirm the suspicion of many conservatives that academia is engaged in the business of putting out heavily biased "science" to undermine them.