Science and Subpoenas
On Friday, Inside Higher Ed sadly reported on an effort underway by the state superintendent of Arizona to use the subpoena powers of a federal court to obtain data from education research scholars at the University of Arizona and Arizona State University. The research in question examines the state’s approach to educating children who are not native speakers of English. If the pending cross-motion of the state superintendent to compel production of experts’ source data prevails, the outcome of that determination could compromise the privacy protection promised to research participants and the confidentiality of the data.
Today, a federal judge has an opportunity to issue an order that prohibits the release of any information that would directly or indirectly disclose the identities of teachers who participated in important education research studies. Or the court could strike a blow to science serving the public interest by ordering identifiable information to be disclosed and thereby making teachers who voluntarily participated in these studies vulnerable to harm (e.g., loss of jobs, public criticism). In addition, the court ruling may result in a broader and longer-term impact by chilling the willingness of others to participate in research or stifling researchers from undertaking some of the studies most necessary to policy and public decision making.
In the human sciences — whether in biomedical or social and behavioral science — advancements to knowledge depend on individuals’ willingness to participate in research. In science, we have long recognized that responsible conduct of research and quality science require the ethical treatment of those willing to be studied. Across ethics codes in the human sciences, including the Ethical Standards of the American Educational Research Association, informed consent and protecting the confidentiality of identifiable information are bedrock to the conduct of research.
In 1979, the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research set forth Ethical Principles and Guidelines (widely known as the Belmont Report) that emphasized the centrality of informed consent and protection of research participants from increased risk of harm due to their participation in research. The Belmont Report led to passage in 1981 of the Code of Federal Regulations for the Protection of Human Subjects (45CFR46), which delegates review of research and implementation of the regulations to institutional review boards (IRBs) at universities and other institutions.
Under 45CFR46, institutions and their IRBs are provided with some degree of discretion in specifying human research protection plans, but have the affirmative responsibility of ensuring that the research procedures set forth by the research scientists protect the privacy of subjects and the confidentiality of data. Discharging their responsibilities, the IRBs at the University of Arizona and Arizona State University approved the research procedures set forth for these studies to protect the identities of participating teachers and any information that could disclose their identities. Following these protocols, faculty members made commitments to the teachers participating in their research, and teachers gave their consent to participate with the understanding that the data in these studies would not be used in any way that could reveal their identities. These protocols reflect a commitment of researchers to their research participants; by virtue of their review and approval, the institutions’ IRBs also affirmed their commitment to the specified plans.
With these education research studies now completed and reports from them independently peer-reviewed and made accessible by the Civil Rights Project at the University of California at Los Angeles, two of the investigators have agreed to appear as expert witnesses before the U.S. District Court of Arizona. As with any scientific expert appearing before the courts, these education researchers can be cross-examined with respect to their own expertise or about the studies that they seek to present. Serving as expert witnesses, however, does not waive in any sense the agreements that researchers made to research participants or the responsibility of the universities and their IRBs to ensure that these commitments are honored.
In an open society, we should engender a commitment among scientists to disseminate their findings on important matters of societal dialogue and debate. Scientific discoveries that are relevant to high profile or controversial issues — whether global warming, disease transmission, efficacy of first responders in disasters, recovery from traumatic events — need to be part of the public discourse where sound science can contribute to policy and practice. There is no more important an arena where science can matter than education, where schools and communities are struggling with limited information and equally limited resources to help guide effective interventions and outcomes, including in this case those related to instruction for English language learners. There is every reason to support and encourage scholars who undertake quality research to publicly present it without fear of undermining how ethically responsible research is done in all of human science or putting their research subjects at direct risk.
Looking ahead, there are steps that institutions and their faculty members or researchers can take to reduce the likelihood of disclosure — including forced disclosure through legal proceedings. These steps range from how and where identifiable data are retained to strengthened data protection plans through obtaining Certificates of Confidentiality from federal agencies, in particular from the National Institutes of Health. These certificates are intended to protect the privacy of research subjects by protecting institutions and investigators from being forced to release sensitive information that could be used to identify participants, including education research addressing sensitive topics (e.g., truancy). Certificates of Confidentiality are intended to encourage recruitment of subjects on sensitive issues by enhancing privacy protections and reducing concerns about participation. Although these certificates are more visible to researchers working in areas where criminal or civil legal proceedings are more likely (e.g., medical malpractice research, studies of drug treatment programs), little is known about the strengths or durability of these protections. On less sensitive topics — as with many studies in education — the research may not qualify at all for Certificates of Confidentiality, even if sought.
Parties in a dispute, as in the Flores v. State of Arizona et al., the suit that has led to the demands for information about the education research conducted in Arizona, certainly have a right to present and scrutinize the research and findings presented before the court. To review the methodology, the data analyses, and the findings — much of which seems to be provided in the reports on the UCLA website — does not, however, require knowing the identity of the research participants. In one of the studies, a survey of some 800 teachers, the investigators themselves do not know who the anonymous respondents are. By forcing disclosure of information about specific schools or school districts, the identities of now anonymous participants could be deductively disclosed.
In the social sciences and very much evident in education research, there is a history of data access under restricted licensing agreements that allow other bona fide scientists to reanalyze data or undertake additional analyses adhering to the original confidentiality agreements made with research participants. As a condition of these licenses, there are very stiff penalties for violation. It would seem that such steps would not be necessary in the pending case because the court could satisfy itself that the expert testimony is reliable based on the factors established by the Supreme Court in Daubert v. Merrill Dow Pharmaceuticals, Inc. without needing to know the names of the participants. If there are questions unaddressed that are important to the Court, it would be far better to ask the researchers to provide further analyses or to have other qualified scholars undertake analyses under restricted conditions than run the risk of compromising teachers or intimidating researchers by compelling disclosure of the identities of human subjects.
We are at a time in our history when science is sorely needed to advance the public good in education and elsewhere. These universities should stand firm in protecting the integrity of science and backing their IRBs that approved of the research protocols guaranteeing that subject identities would not be revealed. Were the universities to do otherwise, they would need to scrutinize whether breaches of confidentiality constitute adverse events according to their own guidelines. For academic researchers or their universities to face this challenge when researchers and their studies could be scrutinized without revealing participants’ identities is a sad moment indeed. In an era when good evidence in education can help to make wise decisions, we can only hope that the U.S. District Court for the District of Arizona will defend the interests of these researchers and those who serve science and society through their participation in research.
Felice J. Levine is executive director of the American Educational Research Association. She is associate editor of the Journal of Empirical Research on Human Research Ethics.