i've been stewing about this one for a while, trying to come up with an appropriate response to timothy burke's recent (ok, less recent now that i've been stewing for a while) post "a tapeworm on the body academic." the key bits:
"This is one of those issues where there isn’t a reasonable accomodation, where people of good faith can sit down and say, 'Ok, sure, maybe in a few cases, an oral history project ought to go through an IRB'. Flatly: there are no conceivable works of oral history, journalism or cultural anthropology that require such a review. There is virtually no survey research that requires such a review ...There is a narrow class of research that IRBs justifiably should supervise: those that involve direct medical and psychological experimentation on individual human subjects. Everything else is pernicious and wasteful at best, and at worst, constitutes a kind of creeping stranglehold on free inquiry."
whoa there! i can think of at least two broad classes of "works of oral history, journalism, or cultural anthropology" for which, unless the researcher is completely unconcerned about the consequences of the research for her subjects, IRB review is both good and necessary.
class the first is about direct harm to the subject through the process of investigation. this is what happens when you are bumbling around, unconnected to solid oversight, trying to uncover memories, record narratives, or otherwise memorialize a traumatic event. the psych term is "retraumatization," and as the term suggests, it's something you might reasonably worry about whenever the subject population has high incidence of trauma (violence, famine, migration, persecution, whatever).
[i'd be less concerned about retraumatization if it weren't so attached, in my thinking, to the question of wartime sexual violence. here is a phenomenon that is incredibly culturally taboo and for whose survivors, in quite a number of conflict and post-conflict settings, absolutely no mental health (or other) services exist. so you talk to someone, who has freely decided to speak with you for whatever reason (and on this, see below), and s/he tells you about what the soldiers, or the neighbors, or whomever, did -- and then you leave your subject alone (s/he would never tell anyone but an outsider what happened) with the memories you've just helpfully freshened up. we can expect that a lot of people -- maybe even most -- will be fine. but there exists a subset that ceases to function after you fly back to new haven. or wherever.]
a subgroup of class the first consists of cases involving indirect harm. here, the subject gives informed consent to speak with you and, in doing so, risks serious harm to self and others. [you were seen leaving the informant's house; tomorrow a death squad kills her elderly parents.] this is a relatively common situation in conflict and post-conflict settings.
class the second is trickier than the "first do no harm" considerations above, and i suppose i'm less wedded to this as a reason for IRB meddling. but i think it's absolutely worth considering (esp. if it leads to indirect harm as discussed above). it has to do with whether, when a community is especially poor or traumatized or fearful or marginalized or what have you, any sort of inquiry can be *actually* non-coercive. at some point, it doesn't matter how carefully you disavow any concrete exchange for your informants' services, or how clearly you explain that you don't necessarily have power over their situation. you are the powerful outsider, and you will be told...something...regardless, because the situation is desperate. as with all sorts of coerced testimony, your concern might be with the veracity of the information. but the ethical questions certainly (should) outstrip your worries about bias.
to return to timothy burke's very provocative statement: there are plenty of post-conflict oral histories, and plenty of cultural anthropology about people who are, in one way or another, "disadvantaged" or traumatized; all these fall within at least one of the classes i outlined above. i suppose one might argue that these are fields with relatively highly developed (relentlessly, boringly self-reflexive?) ethical standards *outside* the IRB process, that political science is a different animal, and that my critiques somehow don't apply when a field has spent a long time independently considering the ethics of its methods. unfortunately, though, i'm not sure "highly developed ethical standards" applies everywhere or to everyone. in particular: i am a graduate student. i've never done this before. everything feels like a race. and i'm going to magically rely on my professional ethics when it feels as if my career's at stake? burke's argument places far too much faith in the virtue and diligence of academics in a publish-or-perish world.
in any case, i can hardly believe the assertion that "direct experimentation" is the only context in which researchers can harm or coerce their subjects.
the remaining question is whether IRB's are the right vehicle for those considerations -- and about that i'm not certain. at yale the human subjects board has been careful and thorough with social scientists but, in most cases, aimed at helping you not-do something you might regret rather than at setting up mindless roadblocks. if IRB's aren't the right way to review qualitative research designs, though...what is? individuals, individual departments, and individual disciplines all develop weird methodological blind spots and outsize faith in their own moral rectitude. without delving into analogies about historians revieiwing quantum physics or veterinarians standing in the way of oral history projects (as burke does), it's worth considering who, exactly, might do better than the IRB. is it that IRB processes are necessarily bad, or that some IRB boards suck? i think the latter.
it's also worth remembering that, however bad your IRB "horror story" might be, it's likely tamer than some of the horror stories to which you might expose a vulnerable population if you wander into its midst with a notebook and tape recorder.