Is ethical research possible on the efficacy of interrogation techniques?

Yahoo! News reports today:

CIA officials extracted valuable information from a terrorism suspect after he was subjected to waterboarding, a simulated drowning technique that has been condemned as torture, a former CIA interrogator told U.S. new media.

The Washington Post reports that John Kiriakou, the source of this story, said he was not present and did not personally observe the waterboarding, but is relying on what was told to him by other CIA staffers who did participate in the waterboarding.

So we are presented with a media report, based on interviews with someone we know nothing about, who heard from someone else, that an interrogation technique worked.

There must be a better way to evaluate the efficacy of interrogation techniques than to rely on the subjective impressions of interrogators. Even if an interrogator is honest to himself, and honest to those seeking information on interrogation processes, he would not be a good informant on the efficacy of his interrogation.

A scientific process begins with classification. Are all subjects of interrogation alike, or is there a way to describe subjects that provides information on their responses to interrogation? Are all kinds of information alike, or is there some classification of topics that provides information on their accessibility to different interrogation approaches? What is the most useful classification of interrogators and interrogation techniques for learning about the efficacy of different approaches.

Medicine has been plagued for centuries by the problem of how to separate efficacious treatments from those which are not efficacious. The body has a wonderful array of defenses against disease, and a wonderful capacity to repair itself. People get better with some frequency, no matter what is done to them. How then is the medical profession to figure out which treatments work, and which do not? The double-blind, case-control study has come to be the double standard for pharmaceutical research. Of course, surgery depends on the surgeons knowing what he is doing, so the double blind research approach does not work for the evaluation of surgical techniques. Still, careful statistical analysis is used. It occurs to me that the techniques that have been developed over the last century for the scientific evaluation of psychiatric approaches might provide some clues on how to study interrogation techniques scientifically.

There are, I suppose, interrogation techniques that are the equivalent of medical techniques, in that their intent in wholely beneficial, and harm would only be an unplanned side effect. Some people, I suppose, are helped by an appropriate interrogation that allows them to reveal information, and sometimes that information can be beneficial for society in general. Indeed, from what I read, a lot of interrogation done by intelligence services is of this kind, in which the interrogator takes the time to establish a good relationship with the subject, and indeed helps that subject to reveal the required information (and even to learn about errors in thinking in the past.)

However, our attention is directed to coercive interrogation. The more urgent the need for the information held by the subject, and the more important is that information, the more coercive interrogation techniques that might be justified, supposing that they do in fact work. But how do we find out whether they do work?

And how do we do so ethically? I doubt that we can experiment with informed volunteers, because the circumstances would be so different that the real situation.

I would note that there are descriptive sciences as well as experimental sciences. Science can proceed through theoretically informed observation with statistical analysis of results. Such observation is usually better if done by non-participants.The key to enhancing quality, however, may be subjecting the interpretation of reports from such observations to peer review, the accumulation of bodies of replicated observations, and the social construction of knowledge about interrogation techniques among a network of serious scholars of the subject.

It may well be that the best we can do to improve coercive interrogation techniques, while still maintaining some pretention to ethics, is to have records of interrogations (both more and less coercive) that are believed to be ethical under the circumstances that would be observed and studied by networks of independent scholars who seek to establish a serious scientific understanding of the interrogation process.

Thoughts from others, please.