Before a day had passed, the torture debate had flared. The chairman of the House Homeland Security Committee, Representative Peter T. King, Republican of New York, told Fox News that the success of the hunt for Bin Laden was due to waterboarding. The next morning, Senator Dianne Feinstein, Democrat of California and chairwoman of the Senate Intelligence Committee, said just as flatly that “none of it came as a result of harsh interrogation practices.”

This from a New York Times article this morning on how the killing of bin Laden has revived the debate over the efficacy of torture. The evidence? Some prisoners who were tortured gave useful information; others didn’t.

But a closer look at prisoner interrogations suggests that the harsh techniques played a small role at most in identifying Bin Laden’s trusted courier and exposing his hide-out. One detainee who apparently was subjected to some tough treatment provided a crucial description of the courier, according to current and former officials briefed on the interrogations. But two prisoners who underwent some of the harshest treatment — including Khalid Shaikh Mohammed, who was waterboarded 183 times — repeatedly misled their interrogators about the courier’s identity.

I’m morally opposed to torture, so what follows is mostly just for argument’s sake. But if policy makers want to know whether torture “works”, presumably it shouldn’t be hard to conduct a random experiment to find out.

National security policy seems to raise a lot of interesting causal questions. One question is why the US is using torture. As the Times article mentions, there is a debate about whether torture is effective at extracting information from suspects. (Note the amorality of such a debate; yet such concerns seem to be outside of the domain of causal inference.) If the government were interested in useful information, they possess the means to put such concerns to rest, as I mentioned above: just do an experiment. The policy doesn’t seem as carefully considered as that, though. Instead, and these are just my impressions, the interrogation policies were assembled haphazardly, and were largely a function of machismo and the desire for revenge.