Background

As a postdoctoral student at Yale 35 years ago, I once bumped into the distinguished sociologist Al Reiss, who promptly invited me to lunch. He carried under his arm a clutch of papers that he had just received at a conference in Washington and was anxious to talk. The conference had been sponsored by the major national funding agencies—NIH, NIMH, and NSF—and called in response to a new, rising concern with ethical standards in social science research. The task addressed by conference participants, including Al, was that of devising ethical standards for research involving human subjects, and then establishing a process for enforcing them. Their emerging plan was to develop a self-enforcement mechanism that would ward off more far-reaching oversight by Congress or federal agencies—the classic reaction of any would-be regulated industry. In a nutshell, the proposal Al outlined to me that afternoon was to condition receipt of federal funds upon guarantees that the research proposals would be vetted by local ethics committees at their home institutions.

I was in the midst of researching the criminal process in New Haven, hanging around court all day and talking to everyone—reporters, bail bondsmen, bailiffs, defendants, victims, witnesses, judges, defense attorneys, prosecutors, gallery audiences, police, everyone—so of course I was quite interested in what Al had to say. I listened intently and asked numerous questions.

After lunch, as we left Mory's and walked down High Street, Al showed me a set of proposed guidelines that had received tentative approval at the Washington meeting. I cannot recall their language, but I distinctly recall my reaction: “Al,” I said, “this rule would require me to first get permission from the occupants of those cars at the corner of High Street and Whaley Avenue if I wanted to conduct a study that correlates the make of the cars with the number of occupants. Right?”

“Correct,” he replied, “but don't worry.” He assured me that such sweeping principles would soon give way to narrow and nuanced provisions. After all, he continued, researchers themselves would refine and apply the rules; a sensible common law would emerge.

Al, who passed away last year, was almost always right on almost everything. But on this, his optimistic prediction was flat-out wrong. Rather than a common law–like process that developed narrow rules in response to limited, well-defined problems, we have seen the adoption of a federal law requiring the ethical review of all federally funded research that affects human subjects, the emergence of sweeping regulations and guidelines, and the emergence of a vast bureaucracy—in Washington, in the learned societies, in the universities—that oversees this law and derivative regulations. This process has spawned an aggressive enforcement process at universities across the country, and these regulatory powers are enhanced by risk-averse university administrators and atomized faculty researchers who passively accept the consequences. It casts its net over a vast range of work—far more vast than anything originally intended, and indeed much more than is necessary under current federal law, which requires only that federally funded research be vetted by such committees. Most universities have extended this requirement to apply to all research, whether funded by federal sources, funded by other sources, or wholly unfunded. This vast “licensing authority,” as Philip Reference HamburgerHamburger (2004) calls it, threatens the research of countless scholars.

IRBs Institutionalized

Before I turn to criticize this development, let me be clear: there is good reason to be concerned about ethics in social science research. Universities have a duty to sensitize researchers to ethical boundaries. Those who harm their research subjects should be sanctioned, and those who are harmed need redress. But we must distinguish between some problematic forms of biomedical research, on the one hand, and the bulk of social science fieldwork, on the other. And we need to guard against risk-averse instrumentalism that embraces an expansive regulatory regime when none is needed and when more modest and effective responses are available.

Campus committees established under federal law are now generically called institutional review boards, or IRBs (though, on my campus, the Committee for the Protection of Human Subjects is known to graduate students as the Committee for the Prevention of Research on Human Subjects). These committees are charged with reviewing research proposals to determine if they pose risks to human subjects. If these committees perceive such risks, they may request that research protocols be revised to better protect subjects. If researchers do not receive IRB approval, the campus will not permit the research to go forward, and federal funds cannot be used to support it. Researchers who evade these requirements risk dismissal; journals that publish such research risk severe sanctions—even if it can be demonstrated that no harm to subjects occurred. The stakes for noncompliance are high.

Further, though federal regulations do not require it, most campuses—indeed, virtually all American universities—have applied IRB procedures to all non–federally funded and unfunded research as well. The rationale offered for this is a combination of principle and practicality. If ethical standards are important, they should be applied to research without regard to particular sources of funding. And if there is an effective enforcement mechanism already in place to vet larger-scale projects that are funded by prestigious sources, it makes sense to apply that mechanism to other research as well. The result has been the emergence of an elaborate bureaucracy to vet social science research. Contrary to Al Reiss's prediction, it has neither narrowly tailored its concerns nor focused on a handful of real problems. The reverse has occurred: the process has expanded exponentially over the past 30 years, spawning an immense bureaucratic apparatus that is poised to consider the remotest risk. In an increasingly risk-averse environment, IRBs employing an instrumentalist cost-benefit analysis regularly withhold approval or require significant research modifications in order to preempt speculative harms.

In short, in the name of minimizing risks, IRBs subject researchers to petty tyranny. Graduate students and junior scholars are particularly likely to be caught in their web—and for them IRB tyranny is often more than petty. Senior scholars are generally more adept at avoidance, evasion, and adaptation, but they too are hardly exempt from this tyranny. A number of prominent social scientists, including some members of this Association, know all too well the harms of running afoul of campus IRBs.

IRBs are not advisory boards. They tell you what you can and cannot ask, what you can and cannot do. They can suppress data. They can prohibit publication. They are powerful. Here, it is important to consider the biomedical model implicit in this procedural structure. No one, I take it, would object to the requirement of rigorous protocols and informed consent in the testing of new drugs, or in experiments with new medical procedures. However, it seems inapt that, after 30 years, social science research would still be assessed according to a biomedical model, its disciplinary distinctiveness largely ignored. Indeed, the expansive regulatory approach toward the social sciences is at odds with the critical nature—the praxis—of the social science research that Michael Burawoy celebrated in his 2004 address. Under existing IRB protocols, much of the work he cites as models of public and critical sociology could not be conducted if IRBs exhaustively fulfilled their missions, or if the authors of those works had positions in today's universities. IRBs, as they have developed, pose an impediment to Burawoy's call for a socially relevant and critical sociology. They have taken on the form and function of censorship boards, and as such the universities that embrace them uncritically indulge community standards rather than provide researchers with a shield against them, as has been one of their traditional functions.

Let me catalogue some of the troublesome issues about and actions of IRBs.Footnote ¹ What follows is hardly a comprehensive survey or an unbiased sample. But it does represent issues that have been raised repeatedly by my colleagues and graduate students, frequently enough to suggest that the issues are not idiosyncratic oddities.

1. 1 The IRB review process, in effect, imposes a licensing requirement to conduct research. Researchers must be licensed by the federal government through a power delegated to universities and then to research committees, and this power expands rather than contracts as it is transferred down the chain of authority. If you are not licensed, you cannot conduct research. Some have argued that this licensing amounts to prior restraint and hence contravenes the American constitutional tradition of free expression (Reference HamburgerHamburger 2004).

2. 2 Even if a study involves subjects who are public officials, namely those who are experienced in fielding office-related questions by hard-nosed investigative reporters, researchers still must submit proposals—including a list of questions to be asked, a list of expected findings, a statement of the work's social significance—to their campus IRBs. The IRB, not the researcher, determines whether a project qualifies for a “public figure” exemption. And the IRB determines this based on a detailed application for exemption, an application just about as detailed as one submitted for full review. That is, an “exemption” here does not mean what it normally means in ordinary language: it becomes a privilege bestowed by an IRB, not a right to be exercised.Footnote ²

3. 3 The process of applying for an exemption itself can create a moral dilemma. How candid should a researcher be, if IRB review jeopardizes even projects that should be obvious candidates for exemption? Consider researchers taking up the call advanced by Michael Reference BurawoyBurawoy (2005) in his much-commented-upon ASA Presidential Address, or the call for a critical social science by still others such as C. Wright Mills, Pierre Bourdieu, and Nancy Scheper-Hughes, whose concerns he echoes. Researchers following their examples could easily encounter IRB approval problems. Does this sound far-fetched? Consider some not-so-far-fetched possibilities.

   • • In his ongoing research, former LSA president and LSR editor Rick Abel explores conditions under which American lawyers, judges, and other officials embrace torture at the expense of legality. Abel is apparently keen on exposing his research subjects to risk; I take it that he would be delighted if what he wrote tarnished reputations and forced resignations. Since his work is drawn entirely from public print sources, Abel is free of IRB oversight. But suppose he began to talk to some of his research subjects, at, say, his college or law school reunions. I am fairly certain that he would be happy to obtain unflattering information about high administration officials. Should Abel report his aspirations—in response to the question about “significance” in the IRB form—in his exemption request to the UCLA IRB? Indeed, would he be in violation of IRB rules, thus barred from using what he learned in informal chats that occurred prior to his application to the IRB? I think yes.

   • • Consider something closer to my own research interests: cross-case vote trading between appellate court judges or between prosecutors and defense attorneys (“I'll vote your way in this case, if you vote my way in that case”). Judges and prosecutors are good candidates for the public figure exemption, but still, would I jeopardize IRB approval if I were candid about my interests in my application? Or would an IRB require me to disclose the purpose of my research to my subjects? If I were to ask clerks or other members of courthouse workgroups about this and had their informed consent waivers, could an IRB also insist on third-party consent—by judges and prosecutors, who are the objects of and not the subjects of the research? IRBs have been known to impose such requirements.

   • • Consider this dilemma, not uncommon in field research: IRB regulations require that we obtain approval before initiating research, but we often don't know we have started our research until after it is begun. To wit: last spring at a conference I spoke to several current and former state commissioners of correction. These conversations led me not only to formulate a research question, but to partially answer it. Should I be able to draw on my notes from these conversations in my writing? I think so. IRBs, though, would probably say no. Prior IRB approval may feasibly be required for biomedical tests or questionnaire-based research, but it poses severe problems for other types of research whose beginnings are often discerned retrospectively, and whose nature often leads researchers to revisit field notes to investigate questions not formulated at the outset. Perhaps this type of research is not so good and should not be fostered. But is it an IRB's province to make this determination? In fact, IRBs can and do make precisely this determination, not only on an individual basis, but institutionally, as they selectively privilege certain types of inquiry within a prior restraint-like framework. Recently the faculty chair of the IRB at a well-known university informed me that if a research project is, by her lights, poorly designed, any amount of risk—even the harm associated with wasting a subject's time—is too great. Her campus's IRB had rejected, and would continue to reject, applications because of unsound methodology. I fear that this practice is not a rare exception.

   • • Federal regulations—at least Berkeley's version—license IRBs to inquire into not only the risks to research subjects, but the social “significance” of research. I am not sure why. Perhaps the more socially “significant” the work, the greater the range of justifiable risk (“no pain, no gain”). Such a trade-off might be reduced to a Hand-like risk-benefit formula to determine a research project is worthy of support. Again, I can understand such an approach in biomedical research, but it makes no sense in most social science research. Although the IRB review aims to protect research subjects, the process can attract, and has attracted, third parties who effectively intervene, formally or informally, to stifle research. Such third parties are likely to be research “objects,” rather than subjects, who seek to squelch research that might portray them in an unflattering light. Cigarette companies, businesses, and university administrations come to mind here. These scenarios are not far-fetched hypotheticals—they are real problems.Footnote ³

   • • Yet another problem is that, as with many censorship regimes, the IRB review process is highly decentralized. For example, each University of California campus has its own IRB (some have two: one for the social sciences and one for the medical school). This promises convenience—I can walk from Boalt Hall to Dwinelle Hall and hand-deliver my proposal to the IRB office staff and have face-to-face contact with them, rather than send it to an anonymous committee at a P.O. Box at the president's office in Oakland. But this convenience comes at a price. In the name of flexibility, accessibility, and efficiency, IRBs at the UCs (and in many places) are staffed by faculty who serve for limited rotating terms. There are at least three problems with this arrangement. First, continual rotation of board members leads to inconsistency and lack of institutional memory. Second, rotation and voluntary staffing leads to self-selection; self-appointed protectors of ethics gravitate to and remain on these committees (and their permanent staffs) while others quickly come and go. Third, and most troubling, IRB service seems to confirm the power of “role theory”: create the position of censor, appoint a person to that position, and there is a good chance that she will act as a censor. Even our liberal, productive, research-oriented colleagues are not immune to this tendency! Furthermore, like all censorship boards, IRBs are subject to moral panics. At times they can overreact to hot-button issues of the moment, and more generally they can produce a chilling effect that turns ultimately to ice.

   • • Such problems, which frustrate individual researchers as they seek to get their research underway, also have important institutional and “upstream” consequences. Faced with the costly delays and changes that follow from IRB-related challenges, researchers may abandon their projects and settle on more easily approved forms of research. This is the case for some of my graduate students and junior colleagues at Berkeley, and I have heard of similar responses by others elsewhere. But more generally, I fear that such problems affect scholarship in several less-obvious ways. These problems may affect courses offered, they may privilege certain research methods at others' expense, and they may affect the priorities of funding agencies.

IRBs as Lawless

There may now be a generation of scholars who take this regime of censorship for granted. If so, entire research areas and methodologies are in jeopardy, insofar as the difficulties of obtaining IRB approval affect research priorities for funding agencies and universities' willingness to support researchers. This is not idle speculation: I have for years been attending panels on IRBs at meetings of professional associations, talking to colleagues who have served on IRBs, and reading the occasional article about IRBs. Furthermore, I have held the hands of two generations of graduate students who have had to navigate the minefields of IRB review. One overarching theme has consistently appeared in such encounters: reasonableness—everyone involved in administering IRBs professes to being “reasonable.” Yet in exercising their reason, responsible people have rejected research proposals for lacking sufficient promised social significance; reached quite divergent conclusions, for example, when defining the scope of “public official” exceptions—even within the same university; and varied widely in their propensities to consider far-fetched hypotheticals when assessing the possibility of harm to subjects.Footnote ⁴

What is my point in recounting all this? That IRBs can be tyrannical? Certainly. That we should be alarmed at their expanding authority? Yes. That IRBs should be reined in? Yes, all these things—but something more as well. If you are like me, you regard your campus IRB as a nuisance, to be evaded when possible, and otherwise to be approached deferentially, hat in hand, like an IRS auditor. But let me reflect more seriously on IRBs in light of the concerns with which I began this address.

We, all of us, are scholars of law and society—and, as I argued at the outset, our tradition is anchored in a jurisprudentially robust conception of law. Despite this, it appears that few of us have brought our sociolegal instincts to bear on the pathologies of a censorship regime that operates in plain sight right under our noses, and that encroaches upon our research lives. Yet this is an open regime in which we can and do participate. As often as not, the enemy is in plain sight, and it is here, not “over there”; it is “us,” not “them.”

When held up to the standards for legality, IRBs fall short in a many respects. They substantially violate our understandings of legality and the rule of law. Lon Fuller, a leading American legal philosopher and an active supporter of this Association at its founding, wrote extensively about the conditions of legality. In his view, inherent in the concept of law is an aspiration for justice and morality. That is, the law itself makes minimal moral claims, and the extent to which a law falls short of those claims is the extent to which it fails to count as law. In a famous passage of his classic, The Morality of Law, he writes:

Fuller then observes, “A total failure of any one of these eight directions does not simply result in a bad system of law; it results in something that is not properly called a legal system at all” (1965:40). After elaborating these ideas, he draws on sociologist Georg Simmel's observation that successful governance depends on reciprocity between government and citizen. “Respect for constituted authority,”Reference FullerFuller reminds us, “must not be confused with fidelity to law” (1965:41). If Fuller and Simmel are correct, we may grant IRBs respect not because of their “fidelity to law,” but because of their position of power. IRBs violate several of Fuller's conditions for law. Accordingly, I suggest, we cannot accord them the fidelity that law requires.

These failures are of the same sort that Judge Marvin Reference FrankelFrankel once identified with criminal sentencing (1972). Thirty-five years ago, Judge Frankel wrote about vast unchecked discretion, disparities, inequalities, inconsistencies—pervasive arbitrariness in the American sentencing system. He did not simply point out that some judges abused their authority, or that the sentencing process is erratic. Rather, he argued that criminal sentencing—in conception as well as in application—was flagrantly lawless, a great “wasteland in the law.” Yet Judge Frankel did not throw up his hands. Instead, he reminded his audience:

Judge Frankel's statement about criminal sentencing in the 1970s applies today, with equal force, to the process of the ethical review of research on human subjects.Footnote ⁵

I appreciate that my views may be more firm and fixed than those of many readers. Clearly, I have not given an utterly disinterested, nuanced, or complete account of the issues—this has not been my objective. Rather, I have sought to identify important issues that have simmered for years within our organization and within the social sciences generally. These issues may now be coming to a boil. I certainly hope so. My intention here is to turn up the heat by highlighting this growing problem as it applies to our field. I urge colleagues to discuss and debate it, through panels at LSA meetings and in other professional associations, but most of all through activities at our home institutions.

I believe there is hope. As more legal scholars begin to assess IRB activities, and as complaints about IRBs are increasingly cast in terms of rights violations, the level of concern seems to be rising (Reference HamburgerHamburger 2004; Northwestern University Law Review 2007). Perhaps this follows the dynamic posited by Stuart Reference ScheingoldScheingold (1973) some time ago, and more recently by Michael Reference McCannMcCann (1994) and still others in this Association: while the law itself may make little difference, litigation—or the threat thereof, or the successful invocation of the “myth of rights”—can make a difference by catalyzing social mobilization. Whatever the reasons, I am pleased to see this growing awareness of a serious problem that affects us all.

What is to be done? I do not have a specific blueprint. But much can easily be done. As I have suggested, and as Pogo used to remind us, “We have met the enemy, and it is us.” IRBs at most American universities are not foreign regimes foisted on us; they are administered by us and by our peers. Join them; subvert them—or at least curtail them. Serve on them and do all you possibly can to facilitate the research of your colleagues rather than act as a censor. Remember that federal regulations require only that federally funded research be reviewed by IRBs; urge your university to adopt and then implement a policy that spares researchers who are not federally funded from having to run the IRB gauntlet. Most of the research conducted by this Association's members is not federally funded and thus could easily be free of IRB review. At present, only a few universities—the University of Chicago comes to mind—have adopted this policy, but so far to my knowledge none of them has acted on it yet. Press for your university to opt out.

Finally, respond to opponents who would question your sensitivity to ethics in research, and demand that they document the nature and extent of the harms (see, e.g., Reference MuellerMueller 2006). Of course research curtailed eliminates risk of harm to an absolute zero. But press and ask about estimates of how much actual harm to subjects in social science research has been prevented by IRB actions. And ask for documentation. Of course, some harms have occurred, and others have been prevented by IRBs; one would expect all but the most ineffective bureaucracy to have some successes after 30 years and countless dollars devoted to the task. But ask whether the harms are substantial enough and frequent enough to warrant an immense bureaucratic apparatus that engages in aggressive proactive review and oversight.

Indeed, appreciate that some harms to research subjects may be desirable—isn't that what some critical and some action research is about? Many of us assign the movie Erin Brockovich or Barbara Reference EhrenreichEhrenreich's book Nickel and Dimed (2001) to our classes. Others of us are inspired by the examples of Nancy Scheper-Hughes, Pierre Bourdieu, and countless other committed social researchers. Why shouldn't we be able to conduct similar work ourselves?

Ask how risk-averse this process should be. Ask whether there is another, less-obtrusive alternative. Ask whether a reactive, tort-like regime for civil remedies following demonstration of harm might not be more appropriate than an elaborate proactive regulatory system. It may be debatable whether current laws on libel, slander, and copyright are sufficiently effective, but even if not, there is, so far as I know, no groundswell for a proactive regulatory regime to replace them. One reason is that principles related to free expression, proportionality, and the presumption of lawfulness are values engrained in American law. The American legal tradition is to set boundaries and permit freedom within them, sanctioning only those who abuse their freedom by transgressing these boundaries. IRBs, in contrast, police the interior of those boundaries in order to preempt transgression.

There is a better way to deal with abuses of human research subjects, one that is familiar to the American legal tradition, easily implemented, and able to provide redress. I hope many of us will invoke it as we critically shape the discussion in the coming crisis of IRBs.