Sunday, January 21, 2007

The Language of Exemption

In an earlier posting, I asked

How many examples must critics accumulate before Sieber, Menikoff, and others will ask if these problems are not the fault of individual IRBs, but of the application of rules designed for experimental research to non-experimental research?

A correspondent comments:

I do not think 'experimental' versus 'non-experimental' gets at the problem you are trying to address. The rules are not really designed for experiments. For example, the 'Tuskegee Experiment', despite the name, was not an experiment.

I am not sure I am ready to concede this point, but the comment does highlight the need for language that precisely delineates the kind of research I and others would like to exempt from IRB review. I therefore present four proposals, made over the years, that would exempt qualitative interview research:

1. In 1979, the Department of Health, Education, and Welfare (HEW) announced planned revisions to 45 CFR 46 (the federal regulation requiring IRBs). Overall, the proposal would have greatly expanded the reach of IRBs, but one of the possibilities it considered would have exempted "survey activities involving solely product or marketing research, journalistic research, historical research, studies of organizations, public opinion polls, or management evaluations, in which the potential for invasion of privacy is absent or minimal." (Federal Register, 14 August 1979, PDF)


2. The announcement provoked outcry from a variety of scholarly organizations. The American Association for the Advancement of Science "resolved that the Association urge that further government regulation designed to protect the human subjects of scientific inquiry not be made applicable to studies that involve no more than the free exchange of information between adult subjects, competent to make their own decisions, and the scientist."
(American Association for the Advancement of Science, "AAAS Resolution: Protection of Human Subjects of Research," Center for the Study of Ethics in the Professions at IIT, 14 August 1979.

3. In 1980, several groups endorsed the American Council on Education's demand that the regulations include the proviso,
“These regulations do not apply to research using legally competent subjects that involves neither deceit nor intrusion upon the subject’s person nor denial or withholding of accustomed or necessary resources.”
[J. W. Peltason, "Comment on the Proposed Regulations from Higher Education and Professional Social Science Associations," IRB: Ethics and Human Research 2 (February 1980), 10, JSTOR ]

4. I haven't traced all the proposals since then, but in 2006, the American Association of University Professors recommended "that research on autonomous adults whose methodology consists entirely in collecting data by surveys, conducting interviews, or observing behavior in public places, be exempt from the requirement of IRB review—straightforwardly exempt, with no provisos, and no requirement of IRB approval of the exemption." (AAUP, "Research on Human Subjects: Academic Freedom and the Institutional Review Board," 2006)

Since any of these proposals would get historians off the hook, they all strike me as improvements over the status quo. But there are significant differences among them.

The HEW proposal is the most limited. By listing by name only a few disciplines to be exempted, it would have kept IRB review of disciplines, such as folklore, even further removed from the biomedical and behavioral research that was the target of the National Research Act of 1974. Far better to list methodologies, such as interviewing. Moreover, from the recent debate over the meanings of "generalizable" and "in general," we can now see that the proposal would have left unresolved the question of who gets to determine when "the potential for invasion of privacy is absent or minimal."

The 1979 AAAS proposal and the 2006 AAUP proposal are quite similar to each other, in that they both stress the right to talk to competent adults. However, the AAAS is less precise in its language. Many researchers who have found themselves subject to IRB review do not consider ourselves "scientists." Moreover, the AAUP recognizes the need for self-exemption. If only IRBs (and their designees) can exempt researchers from IRB review, then researchers will remain subject to the caprice of the IRBs.

A comparison of the AAUP proposal to that of the ACE raises an interesting issue. Since the AAUP's exemption targets "methodology [that] consists entirely" of surveys, interviews, and observation, it would not encompass research that withheld resources, such as the penicillin denied the subjects of the Tuskegee study. But what about deception, which can be part of surveys and interviews? The AAUP's report is ambiguous on the issue. On the one hand, it raises the specter of Milgram's obedience experiments (which relied on deception) to suggest that "an across-the-board exemption for all social science research is arguably overbroad." On the other hand, it cites Keith Humphreys and Michelle L. Brandt to note that we have no empirical evidence that IRBs protect human subjects. Thus, even if we concede that research using deception raises significant ethical problems, that's no reason to assume that IRB review can identify or resolve those problems.

The AAUP's proposal still needs a bit of tweaking on this question, but it is very close to a plausible amendment to the current regulations.

No comments: