More on IRBs

Fair warning:
If you’re not interested in Institutional Review Boards (IRBs)–and there’s no reason you should be–you should probably just skip this whole post.

Brief background: Federal regulations require IRB approval is required for all professors’ research on people. Since 1991 (I just learned this from Shrag’s blog), IRB approval was expanded to cover, among others things, participant-observation research (that’s what I do, with an emphasis on participant). You want to interview or observe somebody? You need IRB approval first. The purpose of the IRB is to protect research subjects. There’s a bit too much history about scientists doing bad things.

Professor Zachary Schrag, a man who keeps a blog about Institutional Review Boards, left me an interesting comment in regards to a previous post:

It’s quite possible that an inept IRB would have blocked Venkatesh’s research. But Venkatesh, by his own admission, “f[oul]ed up” by passing on information that people meant to tell him in confidence. So it is also possible that an adept IRB would have permitted the research while mitigating the harm. Not likely, perhaps, but possible.

Rather than deal in such speculations, I hope you will elaborate on your own experiences with IRBs that led you to distrust them.

I don’t distrust IRBs. My practical experiences have been more or less favorable. I just fundamentally question the very notion of needingIRB approval for non-experimental social-science research on capable adults. And I firmly believe that the simple nuisance and fear of conflict with an IRB limit social-science research.

But first let me deal in some speculations:

Not only do I think an IRB wouldn’t approve Venkatesh’s research. I don’t think an IRB doing its role of protecting research subjects shouldapprove Venkatesh’s research. The risk of some harm from his research was so great as to be virtually inevitable. But I think Venkatesh should do his research, and hence my problem with the IRB in general.

I read one review of Venkatesh’s book that went so far as to to use the word “evil.” I don’t think it is. With regards to potentially violating protection of research subjects, I would draw the line between malum in seand mala prohibita. In other words: drug dealing OK, murder not. I don’t think Venkatesh crosses this line. But for Sudhir’s sake, I’m glad nobody shot somebody while exclaiming, “Am I going to make your book now, Mr. Professor?! Is this what you want?!”

While in theory the IRB could have mitigated harm in Venkatesh’s case, it’s hard to imagine a scenario where this would actually happen. I’ve never heard of an IRB with follow-up. Thus the IRB, while failing in its basic mission, still managing to hinder good qualitative research.

Too often, the quantitative researcher’s goal becomes how to outwit the IRB. I know of many good researchers and otherwise ethical people who admit to lying and deceit when it comes to the IRB. They feel, rightly or wrongly, that such deceit is necessary for research in their field to survive.

For my book’s research, I honestly don’t remember or have a copy of what I gave to Harvard’s IRB ten years ago. I know it took three of four drafts and I agreed to make an announcement on day one of the academy stating who I was. It was good to be forced to make this announcement as it wasn’t easy to make (so thanks, IRB). That being said, I also didn’t want to remind my classmates every hour that they were being watched by a researcher. Is this ethical? I think so. My point was to be honest and overt. And I was. But I think I was supposed to post something that never got posted.

I never got IRB approval for my switch in research plans when I actually had to get hired as a police officer. Between the heat from the P.D. brass disallowing my original research plan (for not being a cop) and the heat from my dissertation adviser for my new research plan (for being a cop), well, I had other priorities.

Alas, my fear of preserving confidentiality and protecting my research subjects went so far that I didn’t leave any way to link specific people at different times to their actions and quotes. What if I get subpoenad(?!), I thought. This was a mistake. I could have written a better book if I had characters. I would have, if it weren’t for the IRB.

My own experiences with John Jay College’s IRB have been pretty good. But there are still problems:

1) The line between doing research and not doing research is not as clear cut as the IRB want to believe. My book is done. I’m no longer doing research for it. But when a friend calls or writes, I’m still going to note information for my records. Am I supposed to get IRB approval before going to a Bull Roast? Am I to stop going to Baltimore now that my “research” is complete? Am I to ignore everything I see and do since I don’t have IRB approval?

2) Signed consent forms are simply impossible in police research. First of all, cops won’t sign them. Second, the social situation is flux. You arrive on a scene and officers come and go. Are you supposed to ask every arriving officer for signature? If signed consent forms were required and if this requirement were actually enforced, qualitative police research would simply end. Researchers who say they’re going to get signed consent forms from police officers are lying. Yet IRBs love signed consent forms. It’s like court overtime pay is for police. Just give ’em some and they’re happy.

3) IRBs want a guarantee of confidentiality. I won’t do that unconditionally. Luckily, Baltimore police behave pretty well, at least within the bounds of reason.

For those interested (and you might be if you’ve read this far), I’ll include the more original parts of my successful IRB submissions for approval withoutsigned consent forms and withoutunconditional guarantees of confidentiality. And they said it couldn’t be done.

Requiring a signed consent form from every research subject would so limit my participant-observation research as to effectively kill it. Given the fluid nature of social interactions in a police station, it is not possible to have every police officer who enters a room or call for service to read, understand, and sign a consent form. My research methods are overt, but informal. While the bulk of police officers will be familiar with me and my research, it is inevitable that I will see and hear police officers who initially will not know who I am. Such confusion is usually clarified immediately by asking me or another police officer.

My research is confidential and offers minimal, if any, risk to police officers.

Police officers are not normally considered an “at-risk” group (quite the contrary, police officers are often seen as the group that places marginalized people at risk). Given the constant risk of observation by their supervisors, the public, the media, and internal affairs, an outside researcher who promises confidentiality present little risk. Police officers are used to being on-guard around people they do not know.
…
A Note on Criminal Behavior and Confidentiality

While IRB boards consider harms that may come to research subjects, the potential of research subjects harming others should not be ignored. Should a researcher remain quiet if confronted first-hand with research subjects who commit genocide, mass torture, or major war crimes? Of course not.

In reality, I do not expect to witness any war crimes. But I believe that in extreme cases, a researcher’s obligation as a human being come first. Were I to witness a police officer in severe violation of criminal law, I would have to weigh any promises of confidentiality with my moral and legal responsibility to do the right thing.

I would not violate confidentiality for violations of departmental regulation or even most criminal acts. In my years both as a police officer and as a researcher (often the two overlapped–I was a Harvard University graduate student researcher during my entire two-year tenure as a Baltimore City police officer), I have never participated in nor witnessed an act that would make me consider violating confidentiality.

But if, hypothetically, I witnessed a police officer rob and kill, or sexually abuse a 10-year-old child, or anally violate an innocent man with a plunger, I would feel little compunction legally and ethically to violate a vow of confidentiality. Of course the odds that I will witness such a scene are almost zero. I do not expect such a situation to occur. But I and the IRB must be aware that the possibility, however slight, exists.

I realize that it might facilitate approval of this project if I stated that all research subjects will sign consent forms and confidentiality would never be violated. But to pretend the former denies the reality of participant-observation research on police officers; to pretend the latter is morally irresponsible. I believe it is the combined duty of researchers and the IRB to promote ethical research and protect research subjects. This process begins with the presentation, discussion, and approval of an ethical and honest research proposal.

All researchers are free to use this for their own IRB submissions. I’m happy to try and help the field.