The Censorship You've Never Heard Of
The licensing of speech and the press is profoundly unconstitutional. Other methods of controlling speech, such as after-the-fact penalties, are unconstitutional only when they are used in dangerous ways. But licensing—the requirement that one get permission before speaking or publishing—is always dangerous and therefore predictably unconstitutional. This method of control is itself a threat.
Licensing of speech is a means of wholesale suppression: When the government relies on other methods, it works through individualized court proceedings; but when it relies on licensing, it simply bars entire categories of speech, until individuals get the government’s permission. Second, and equally serious, licensing is profoundly overbroad and disproportionate. In order to prevent harm in some instances, a licensing system imposes prior review in all instances, and it thereby directly suppresses or at least chills much entirely innocent speech. Third, and most disturbing, licensing instills in the minds of the people that they have no authority on their own to share information and ideas. Speech is the means by which they ordinarily hold government to account. But licensing requires them to get permission every time they speak, thereby forcing them to acknowledge the government’s sovereignty over their words. Licensing of speech thus substitutes control for freedom and inverts the relationship of individuals to their government.
The Inquisition licensed printing in Italy, with notorious consequences for Galileo. The Star Chamber required such licensing in 16th- and 17th-century England. Even acting companies in Shakespeare’s time needed a license to recite plays on stage. In repudiation of such things, the First Amendment prohibited all licensing of speech and the press. So it would be astonishing if the government of the United States had restored this old method of suppression.
Yet it has done just that.
Over the past four decades the licensing of speech and the press has returned on a massive scale. Although the suppression is less overt than that of the Inquisition or the Star Chamber, the outcome is no less dangerous. Since the 1970s, the government has overseen the establishment of bodies called Institutional Review Boards, and these “IRBs” have suppressed vast amounts of talking, printing, and publishing—even mere reading and analyzing—for hundreds of thousands of Americans. This is utterly unconstitutional, and in stifling research and its publication, it has proved deadly.
Although IRBs are unfamiliar to the general public, they exist at every university and most other research institutions. Their primary role is to review and license “human-subjects research,” whether funded by the government or not.1 The main government agency behind IRBs is the Department of Health and Human Services. Together with other agencies, it relies on two documents to lay the foundation for IRBs: the Belmont Report, which sets out basic principles, and the Common Rule, which provides more detailed licensing standards. The government also authorizes the IRBs at each institution to add their own requirements. The result is not only a federal baseline for the licensing but also a federal authorization for more severe local standards.
A “human subject” is defined by the government as “a living individual about whom an investigator…conducting research obtains (1st) Data through intervention or interaction with the individual, or (2nd) Identifiable private information.” The government defines “research” as a “systematic investigation” designed to produce “generalizable knowledge”—meaning a general statement or theory.
IRBs comprise faculty, administrators, and at least one community member, all of whom are to be chosen for their “sensitivity to…community attitudes.” These are the persons who regulate and license speech, and who accuse, judge, and punish violators. All IRB deliberations are carried out in secret and researchers cannot appeal decisions.
Like so many harmful government initiatives, IRBs were established to prevent harm. In the mid-20th century, a series of government-run research studies, such as the Tuskegee syphilis study, became notorious for their inhumane treatment of participants. Although the Public Health Service had been studying latent syphilis among black men at Tuskegee for decades, it did not offer or even tell the men about penicillin when it became available after World War II. The drug would have helped at least some of them, but the Public Health Service wanted to preserve its opportunity to study untreated subjects.
In 1972, when the Tuskegee study came under scrutiny, there was public outrage. The government should have acknowledged that the leading examples of ethical problems in science came from the combination of government and doctors in research. The government, however, responded to its own misconduct by declaring that all human-subjects research, even if private and nonmedical, was risky and needed licensing.
Although the government’s primary goal in establishing IRBs was to prevent harm, the suppression of speech was almost inexorable. Under the First Amendment, the government cannot license speech, and even after the fact it cannot penalize speech for causing minor mental distress. But the government had other considerations in mind. First, it had a scientific understanding of research, in which scholars produce knowledge in the form of general statements or theories, which then need to be published so they can be tested. Therefore, when it took aim at research, it defined it as an attempt to produce “generalizable knowledge”—thus defining research as a type of speech. Second, it adopted a medical view of harm, in which any discomfort, even the slightest and most ephemeral, justified government action. The government therefore targeted the full range of speech harms, including the mere anxiety caused by a probing question or a candid publication. Third, the government was not content to punish the harm after the fact, but was determined to prevent it beforehand through licensing. The best of intentions thus produced government control of speech, reaching the most sensitive harms, by means of licensing. It was an unconstitutional hat trick.
Here is how the licensing works. In order for a professor to get government research funds, his university first must assure HHS that all human-subjects research at the institution will comply with the principles of the Belmont Report (or some equivalent document)—including the underlying principle that all human-subjects research requires a prior license.2 The university must also assure the government that it will impose the licensing on federally funded research in compliance with the standards of the Common Rule.
HHS and the other agencies, however, use their conditions to impose the licensing on all human-subjects research regardless of its source of funding. They do this by requiring each institution that receives federal dollars to agree that “all of its activities related to human-subjects research, regardless of the source of support,” will be guided by the Belmont or equivalent principles.3 In addition, the agencies invite the institutions to commit themselves to apply the Common Rule to all human-subjects research done by their personnel. Although this commitment is supposedly voluntary, it has not always been so in practice, and most institutions simply assume that they are obliged to require IRB licensing for all human-subjects research.
By using its conditions to make the licensing pervasive, the federal government has been able to cultivate the idea that a research institution would be violating its duty of care if it failed to impose the licensing. Institutions therefore feel obliged not merely by federal pressures, but more concretely by state tort law, to apply the licensing rules to all research conducted by their personnel, even when the research is not federally funded. Indeed, the fear of being sued under state law has become the primary force behind the licensing.
To top it off, many states have added a third mode of coercion: statutes requiring the use of IRBs. The statutes sometimes impose IRBs on academics and sometimes even on the general public, but they all rely on statutory command. Thus, one way or another—whether by condition, tort doctrine, or statute—the licensing comes with the force of law.
Defenders of the licensing system claim that Institutional Review Boards merely ensure sound research practices. Although these defenders generally do not deny that the government defines research in terms of speech, they insist that IRBs, as a practical matter, do not constrain speech. But is this true? Consider the IRB process. When a faculty member or student seeks to do research on human subjects—to be precise, when she pursues a theory or “generalizable knowledge”—she must first prepare a research proposal and submit it to the IRB at her institution. The board then will review the proposal and follow one of several options. It will permit her to do her work; it will deny her permission; or, most typically, it will deny her permission until she agrees to make changes—changes that usually will limit her acquisition or sharing of information. In particular, the board can forbid her from talking to human subjects, reading about them, observing them, or publishing about them. It can do this even if she merely wants to interview government officials, distribute social-science surveys, or read public records. The board also can directly modify her research: It can rewrite her questions, adjust her academic method, and otherwise alter her inquiry.
In setting standards for this censorship, the government imposes a sort of Victorian prudishness. Little nowadays is not openly discussed, but HHS cautions IRBs: “Stress and feelings of guilt or embarrassment may arise simply from thinking or talking about one’s own behavior or attitudes on sensitive topics such as drug use, sexual preferences, selfishness, and violence.” In other words, when scholars study addiction, desire, or other aspects of human nature, IRBs should use their censorship power to protect individuals from academic questions that might cause them discomfort. And this is only about the use of words to acquire information; the censorship is even more stringent when it gets to what scholars can publish.
That such requirements violate free speech is obvious. But the suppression goes beyond what is conventionally understood as speech. For example, IRBs often require the destruction of data. When scholars publish, they have a moral and academic duty to preserve their data so that the accuracy of their conclusions can be tested. When scholars don’t preserve data, one might reasonably suspect fraud. Nonetheless, review boards require scholars to disaggregate and otherwise destroy their data—whether by requiring that personal “identifiers” be separated from other information or by requiring scholars to get rid of all their data, usually after three years. Either way, such destruction means that many studies can never be challenged for fraud or even merely for misunderstanding the data.
As if this were not bad enough, some review boards actually order scholars and students to stop analyzing data. The University of Michigan, for example, states explicitly that without IRB permission, “no identifiable data may be…analyzed.” In other words, review boards not only license the acquisition and publication of information; they not only require the destruction of knowledge; they also bar thinking.
Faculty and students face devastating punishments for nonconformity—most pervasively because, if an individual fails to comply, his IRB will identify him as “uncooperative” and will review his future work with greater severity. The licensing thus serves a double function. The current licensing is the means of dictating what can be said and published and what cannot. At the same time, the threat of future licensing is the means of enforcing the current licensing.
If a scholar begins an inquiry or prepares a paper for publication without permission, the IRB can tell him to stop. If he submits such a paper to a journal, the board can demand that he withdraw it. This has happened at the University of Illinois, George Washington University, and the University of Pittsburgh. Review boards increasingly compare faculty publications with IRB records to ferret out which scholars have failed to get permission. Review boards can even declare a scholar unfit to do any future research, and they sometimes cooperate to suppress scholarship or harass scholars into quitting.
As a result of the coercion, many faculty and students retreat from inquiry and publication that might annoy their IRB. Entire areas of scholarship have therefore declined, and much important inquiry is never even begun, let alone published. In a 2002 essay in the journal Academe, Margaret Blanchard, a distinguished scholar of journalism, explained her decision to switch fields of study: “I am…leaving the contemporary period behind. It is much safer in the nineteenth century.” Only living persons are “human subjects,” she pointed out, and therefore “you do not have to worry about the IRB when you work in the nineteenth century.” Blanchard added: “I have seen students alter research projects to avoid IRB contact. I have seen some give up projects because of the red tape involved. I have heard words such as ‘thought control’ used far too often.” She concluded: “A better formula for stultifying research is beyond contemplation.”
Many academics are scared to complain at all. Cary Nelson—a journalist who wrote about IRBs in an article entitled “Can E.T. Phone Home? The Brave New World of University Surveillance”—found that “everyone” he talked to (except one whose case was already public) “requested anonymity.” The faculty members wanted anonymity because they were “afraid of their IRB,” and “the IRB members all requested anonymity” because they were “afraid of the federal regulators.”The licensing system thus not only stifles academic inquiry and publication; it also suppresses complaints about its own suppression.
All sorts of excuses are advanced for the IRB licensing, but none of them passes constitutional muster. One concerns conduct. From this perspective, because the regulations refer to “research,” they concern conduct rather than speech, and therefore may be licensed. Yet when the government defines human subjects in terms of “data” and “information,” it is focusing not on conduct, but on the acquisition and sharing of knowledge. And when it defines “research” as a “systematic investigation” designed to produce “generalizable knowledge,” it is expressly taking aim at theories or general statements. Indeed, it is targeting the theories that, from a scientific perspective, require publication. Recognizing the implications, IRBs assume they must license whatever a researcher “plans to publish.” For example, the University of Michigan IRBs state that “publication” is an “indicator” of “generalizable knowledge.”4
This licensing is clearly unlawful, and it is a reminder that the government should never be able to mask its licensing of words by tying the words to conduct. For example, although a state can license driving, it cannot license speaking while driving; nor can it license driving on the basis of what will be said while driving; for it still would be licensing speech. If the government could get away with such excuses, it could always get cover for licensing words.5 In fact, the 16th- and 17th-century English licensing laws illustrate this danger. Rather than impose licensing generally on words, they imposed licensing on printing—that is, on conduct in the use of a new technology. But the laws assumed that licensers would focus on the words that would be printed, and the English licensing laws thus are a perfect example of what the First Amendment prohibited.
There are other supposed justifications for the licensing. One rests on the assumption that the First Amendment centrally protects political speech and thus does not fully protect scientific or academic speech. Scientific and academic speech, however, are foundational. The inquiry done by scientists and other academics has long been understood to illustrate the importance of freedom of speech and the press. Even before the Athenians forced Socrates to drink hemlock, they condemned Anaxagoras for speculating about the physical character of the sun and the moon. Two millennia later, Galileo revealed not only the arrangement of the heavens but also, much closer to Earth, the danger of licensing scientific or other academic publication. And when the Continental Congress in 1774 explained “the importance of the liberty of the press,” it began by observing liberty’s significance for “the advancement of truth, science, morality, and arts in general.”
A final justification of IRB licensing claims that a constraint on speech can be justified by a “compelling” government interest. Undoubtedly, there is a government interest in preventing harms, including those arising from research. But this does not settle the matter, for the IRB regulations focus not so much on harm as on speech. Consider their perverse application. Even when an academic engages in dangerous physical interactions, he needs no permission, as long as he is not aiming for “generalizable knowledge”—the sort of knowledge that academics consider publishable. In contrast, when he pursues entirely harmless activities, he needs permission only if he is aiming for what might be publishable. Thus, not only on the surface of the regulations, but also in reality, the guiding principle as to what requires a license is publication rather than harm.
More generally, the constitutional inquiry cannot rest on a generic claim that there is a compelling government interest in preventing research harms. Instead, the inquiry must concentrate on whether there is a government interest that justifies using the licensing of speech as a means of preventing research harms. To be sure, a compelling government interest can justify after-the-fact penalties on speech; perhaps it even can justify the licensing of moving pictures and other expressive conduct. But it cannot justify the licensing of words.
The licensing of speech causes far more harm than it prevents. Especially in a society based on freedom of speech, the curtailment of this freedom is literally deadening. IRBs annually censor at least tens of thousands of research proposals. In addition, because of the “chilling effect,” an uncertain but large number of projects are never begun or are abandoned. All of this is worrisome because at least some of the suppressed research might otherwise have had profound effects in saving and improving lives. Imagine that each year a single transformative project gets suppressed on account of the censorship. If one adds up the costs over decades, it becomes clear that the consequences for human life are sobering.
An example of the lethal effects came to light in 2006 in Michigan. When catheters are introduced near the heart in intensive-care units, they often cause fatal bloodstream infections. Some researchers from Johns Hopkins therefore attempted to test whether the infections could be reduced by instituting simple checklists—requiring, for example, that doctors wash their hands before inserting such catheters. The hunch proved correct. Already in the initial study, the checklist approach was highly successful. According to the New York Times: “Within three months, the rate of bloodstream infections from these I.V. lines fell by two-thirds….Over 18 months, the program saved more than 1,500 lives and nearly $200 million.” The benefits of publishing the study were even greater, for there were perhaps 80,000 catheter-related bloodstream infections in America every year, resulting in up to 28,000 deaths, with annual costs of more than $2 billion.
But the key point here is the role of the IRB licensing. The researchers dutifully got IRB permission before they began their research. But after the study was well underway, HHS concluded that the review board had not met HHS’s licensing standards. According to HHS, the board should have required the researchers to handle not only the patients but also the doctors as human subjects. On account of the IRB’s failure to comply with HHS’s licensing standards, HHS ordered the Johns Hopkins doctors to shut down the ongoing research. Fortunately, HHS acted too late to stop the research altogether or to prevent its publication. As a result, tens and perhaps even hundreds of thousands of lives worldwide have been saved.
Most biomedical research on human subjects examines only a limited number of persons in order to explore theories designed to help much larger numbers in the general public. Thus, almost inevitably, the harms prevented by IRBs are relatively small compared with the harms they cause. In addition, biomedical research tends to become the foundation for yet more such research, which then benefits even larger numbers. Consequently, whereas the harms prevented by IRBs expand only arithmetically, the harms caused by them tend to expand exponentially.
It is not enough, however, to think of the harm in medical terms, for the harm is also political. Academic studies of human subjects traditionally were uncensored, and through their radical critique of government, they did much to shape the establishment of government health services. This source of information about health-care and its delivery, however, is now subject to the licensing overseen by HHS. In other words, the very government department that regulates health-care also imposes licensing on much of the academic study of health-care. The results can be observed in the studies that draw information at a personal level from doctors, nurses, administrators, patients, and their families. Such studies are essential for judging any health-care policy, but they have been largely stultified, for they cannot be started, let alone published, without IRB permission. HHS thus regulates the public, and is judged by the public, on a scientific record that HHS has largely crushed under its censorship.
The licensing even bars academics and students from interviewing most persons in government, let alone publishing the interviews, without first getting IRB permission. For example, almost all HHS officials, other than the secretary of HHS, are protected by the HHS regulations on IRBs. So too are members of IRBs. As a result, empirical academic critiques of HHS decision-making, and even of the IRB licensing, are practically impossible.
The licensing imposed by HHS is the most widespread and systematic assault on freedom of speech and the press in the nation’s history. McCarthyism was more overtly political, but IRB licensing is more pervasive and methodical, and its consequences are far more lethal.
To grasp the constitutional danger, try the following thought experiment. Imagine an NRB—that is, a Newspaper Review Board. Imagine that the government used conditions and regulations to require not universities, but newspapers, to establish review boards. Imagine, moreover, that a journalist had to get permission from an NRB before beginning any investigation designed to produce publishable knowledge. A journalist would need permission before beginning his labors—even before merely reading, observing, talking, or interviewing—if his goal were to get publishable information. In reviewing his proposal, the NRB would license what he would do, and especially what he would say, to limit the risks to the persons studied, observed, or interviewed. The NRB would be especially careful to limit the risk that the journalist might ask insensitive, stressful, or embarrassing questions and thereby might cause emotional, financial, or legal harm. Indeed, the NRB would prevent him from publishing the names of persons who might suffer anxiety or other emotional harm as a result of publication. All of this would be outrageously unconstitutional. And if it is unconstitutional for journalists, it also is unconstitutional for academics and students.
The First Amendment recognizes that America is an experiment in freedom and that this depends on the freedom of speech and the press. The amendment therefore establishes this freedom as a constitutional right, and on this foundation Americans can resist the current licensing. Socrates died for his freedom. Galileo went to prison for his freedom. All that Americans need is a judge who is willing to enforce the Constitution.
1 Although the government also uses IRBs in more specialized ways—for example, to review trials of new drugs and devices under the auspices of the Food and Drug Administration—such specialized use is not at issue here. Instead, this essay concerns the use of IRBs under the laws and regulations that generally govern human-subjects research.
2 In theory a university can commit to other principles, if approved by the government, but there is no evidence that any other principles have ever received government approval.
3Federalwide Assurance (FWA) for the Protection of Human Subjects, OMB No. 0990-0278 (approved for use through June 30, 2014).
4 University of Michigan, Human Research Protection Program, Office of the Vice President for Research, Activities Subject to the HRPP (July 2011), Operations Manual—Part 4, Section V.C, table 4.
5 Because the point here only concerns the licensing of words, there are no implications for the licensing of conduct; nor for after-the-fact penalties on speech or conduct. For example, no doubts are raised here about penalties on using a phone while driving.