The Asilomar Convention: Revisiting Research Ethics and Learning Science
Last week, a group from a diverse set of higher education institutions released a first draft of the Asilomar Convention, a document attempting to guide ethical decision-making in advancing higher education research. The Convention was composed by researchers, administrators, privacy advocates and others, in an effort to launch a dialogue about how best to advance the science of learning in the digital age. I was one of the folks chiming in on the conversation, and what follows is my own sense of what we were aiming for; others might have a different perspective.
The key document guiding contemporary research ethics is the Belmont Report, which was released in the wake of revelations about, among other things, the Tuskegee experiments where African-American men were not properly informed about or treated for syphilis in the course of a research study. The Belmont Report of 1979 calls for research guided by three principles. The first is respect for persons, including protecting their autonomy. The second is beneficence, which is that research projects should maximize the possibility of advancing the public good while minimizing private harm. Finally, the third principle is justice, which means that the risks of research should not be borne by one group, especially when the benefits might accrue to another. Today, anyone who wants to conduct research on human subjects in a university setting needs to complete a short online course learning about the Belmont Report, its principles, and its history.
There is much quite laudable about the Belmont Report, but it was written in an analog age to guide biomedical research, and in certain ways it provides insufficient guidance to researchers studying online learning in a digital age. So we tried to build on the Belmont Report with a short document that would reaffirm those principals but also begin to speak directly to issues in our own context.
One of the present challenges of digital learning research is that the research institutions who have historically had the most resources available for advancing science (Harvard, MIT, Stanford, CMU), are not the same institutions who are teaching the vast majority of online learners, which are in a variety of State universities, community colleges, and for-profit institutions. Many of us who are interested in online learning research would like to see greater opportunities for sharing and collaboration, in particular through sharing data, technology and practices between institutions. Sharing and openness, particular with data, is good for science, but most of the risks of online research have to do with the risks of exposure of people's personal information, and sharing and merging datasets increases those risks. The thing that that could help us advance science also can increase an individual's risk of personal exposure.
One central tenet of the Belmont Report is that researchers need to weigh the costs and benefits of research, so it's hard to figure out whether society is best served by having scientists merge and share data to potentially advance research and potentially increase the risks of individual disclosure, or whether we ought to restrict sharing, slow science, and also slow developments that can increase exposure. What makes these calculations even harder is that the risks and benefits are quite speculative. I think there are few cases of breakthrough research from these kind of data-sharing arrangements, nor are there clear cases of serious harm from the revelations of learner data collected by social scientists.
So the Asilomar Convention is the first step in a conversation to wrestle with these issues, of opportunity, risk, equity, and learning science in a digital age. We primarily affirmed the principles of the Belmont Report, while also suggesting new contexts for these principles to be re-interpreted. In the weeks and months ahead, I hope the community of researchers and others interested in learning science research will continue to discuss these tensions, and work to find ways that allow us to conduct research, advance science, and protect the autonomy of learners upon whom our research depends.
The full text of the Asilomar Convention is here, and reprinted below.
The Asilomar Convention for Learning Research in Higher Education
June 13, 2014
Individuals, nations, and international agencies of all kinds increasingly rely on the promise of education to improve the human condition. Contemporary technology has created unprecedented opportunities to create radical improvements in learning and educational achievement, but also conditions under which information about learners is collected continuously and often invisibly. For these reasons, collection and aggregation of evidence to pursue learning research must proceed in ways that respect the privacy, dignity, and discretion of learners.
Virtually all modern societies have strong traditions for protecting individuals in their interactions with large organizations, especially for purposes of scientific research, yet digital media present problems for the inheritors of those traditions. Norms of individual consent, privacy, and autonomy, for example, must be more vigilantly protected as the environments in which their holders reside are transformed by technology. Because the risks associated with data exposure are growing simultaneously with the promise of building new knowledge, researchers and educational organizations must be accountable for how they pursue learning inquiry. This convention reaffirms enduring commitments to ethical conduct, and to the protection of public trust in the institutions of higher education.
The convention affirms two tenets for learning research:
I. Advance the science of learning for the improvement of higher education.
The science of learning can improve higher education and should proceed through open, participatory, and transparent processes of data collection and analysis that provide empirical evidence for knowledge claims.
Maximizing the benefits of learning research requires the sharing of data, discovery, and technology among a community of researchers and educational organizations committed, and accountable to, principles of ethical inquiry held in common.
Six principles should inform the collection, storage, distribution and analysis of data derived from human engagement with learning resources. The principles are stated here at a level of generality to assist learners, scientists, and interested citizens in understanding the ethical issues associated with research on human learning.These principles are informed by the 1973 Code of Fair Information Practices, and by the Belmont Report of 1979. These principles will not always produce unambiguous solutions to particular questions, nor should they. Ethical decisions must always be informed by the particularities of their situation.
Respect for the rights and dignity of learners
Data collection, retention, use, and sharing practices must be made transparent to learners, and findings made publicly available, with essential protections for the privacy of individuals. Respect for the rights and dignity of learners requires responsible governance by institutional repositories and users of learner data to ensure security, integrity, and accountability. Researchers and institutions should be especially vigilant with regard to the collection and use of identifiable learner data, including considerations of the appropriate form and degree of consent.
Individuals and organizations conducting learning research have an obligation to maximize possible benefits while minimizing possible harms. In every research endeavor, investigators must consider potential unintended consequences of their inquiry and misuse of research findings. Additionally, the results of research should be made publicly available in the interest of building general knowledge.
Research practices and policies should enable the use of learning data in the service of providing benefit for all learners. More specifically, research practices and policies should enable the use of learning data in the service of reducing inequalities in learning opportunity and educational attainment.
Learning and scientific inquiry are public goods essential for well-functioning democracies. Learning and scientific inquiry are sustained through transparent, participatory processes for the scrutiny of claims. Whenever possible, individuals and organizations conducting learning research have an obligation to provide access to data, analytic techniques, and research results in the service of learning improvement and scientific progress.
The humanity of learning
Insight, judgment, and discretion are essential to learning. Digital technologies can enhance, do not replace, and should never be allowed to erode the relationships that make learning a humane enterprise.
In a rapidly evolving field there can be no last word on ethical practice. Ethically responsible learner research requires ongoing and broadly inclusive discussion of best practices and comparable standards among researchers, learners, and educational institutions