By Richard A. Nielsen, Massachusetts Institute of Technology
* This memo was prepared as part of the “Ethics and Research in the Middle East” symposium
The hour was late at the Al-Azhar mosque in Cairo, but the students kept up a continuous stream of questions and the young teacher, perhaps eager to increase his following, was entertaining them all. I sat silently in the second row of the study circle, but my blond hair loudly announced my presence. As the class adjourned, a young man asked me, “Is your family Muslim?” When I said they were not, he replied, “That must be very difficult. God bless you.” I panicked slightly. Of course he would assume that I was Muslim, because why else would someone listen to hours of discussion about the intricacies of Islamic law? Should I tell him that I was not – and that I was there primarily to learn about scholarly networks at Al-Azhar? Should I have announced myself beforehand? Would he be angry if I had? Pleased? Curious? Skeptical? Would he assume that I was with the CIA?
I left as quickly as possible.
In a very different kind of study circle a year later, senior political scientists debated the ethics of my proposal to randomize the framing of requests for fatwas from Muslim clerics online. What if my experiment resulted in clerics advocating violence? Could someone be harmed or killed as a result of my research? Was it ethical to deceive clerics by representing my request as a genuine religious question? Was it ethical to experiment with religion at all?
I left chastened, with no plans to run the experiment.
My experiences highlight ethical dilemmas that arise because of researcher intervention in a religious space, whether in the service of experimental or ethnographic research. In most areas of political science research, scholars largely agree that ethical intervention requires providing participants ample chance to give informed consent and weighing the benefits of research against the costs. When studying the intersection of politics and religion, meeting either of these criteria can be challenging or impossible. If ethical research on religion and politics in the Middle East faces substantial ethical barriers then scholars may focus on other topics and regions, leaving important holes in our collective knowledge.
Here, I offer some thoughts on the ethics of research intervention in religious settings, based partly on a paper I wrote for the “Ethics in Comparative Politics Experiments Conference” convened by Scott Desposato at the University of California, San Diego in May 2013. I argue that some types of interventions will be difficult or impossible to implement ethically when studying religion and politics in the Middle East. However, we as scholars should do the hard work of creatively identifying ways to carry out research on religion and politics rather than throwing up our hands.
Ethical research interventions in religious settings are often hard
The papers at the UCSD ethics conference catalogued a litany of ethical challenges that experimental research faces in comparative politics, most of which are likely to be present in any study that attempts to experimentally study religion and politics in the Middle East.
For example, a number of participants noted the challenges of operating in authoritarian regimes in which experimental research might provide tools for autocrats to make repression more extensive and efficient. Should researchers seek local approval when conducting experiments in authoritarian regimes? Not seeking approval might decrease the complicity of the researcher in regime politics, but is possibly illegal and increases the risk that researchers will conduct culturally inappropriate research. When ethics review boards outside of the United States have very different standards or practices than U.S. institutions, should they be followed or ignored? What if they permit research that U.S. institutions find questionable? What if such boards do not exist at all? Given the relatively large number of authoritarian regimes in the Middle East, these issues should concern scholars of the region, regardless of methodology and topic.
Others at the conference voiced concerns about whether experiments and surveys of government officials were so time-consuming that they decreased constituency service. After all, an hour-long survey administered to a hundred officials takes up over two workweeks of public time; the benefits of proposed research should be weighed against this cost. These considerations might apply for scholars whose research takes the time of bureaucrats in the Middle East, whether through surveys, interviews, or experiments.
The conference additionally addressed more ethical questions: Does payment for participation in a study create economic inequality and preclude informed consent? Can researchers examine the data produced by ethically questionable experiments carried out by NGOs or governments? Can they partner with NGOs or governments to randomize the rollout of intrusive programs? In most cases, potential answers were framed in terms of informed consent and risk-benefit calculations. Unfortunately, as I describe in the next section, it is on precisely these foundational questions that ethical interventions in religious settings face additional challenges.
Ethical research interventions in religious settings are sometimes impossible
The ability of experiments to produce credible causal estimates comes for the control that researchers exercise over the characteristics of the subjects. Medical trials are credible because researchers control which subjects take the active drug or a placebo and by randomizing this assignment. If scholars want to study the political effects of religiosity with experimental interventions, this requires researcher control over participant religiosity. I offer two theoretical arguments for why this is likely to be unethical in many instances (developed more fully here).
The first is an argument that the ambiguity of religious harms and benefits makes it difficult for researchers to do risk-benefit calculations because the risk and benefits are subjective:
(a) Religion makes untestable claims about harms and benefits.
(b) People with different beliefs can view identical treatments as clear harm or clear benefit.
(c) All treatments that meaningfully change religiosity will be a clear harm to someone.
(d) Such a claim of harm is not demonstrably false by (a).
(e) Therefore the harm should be taken seriously.
(f) These harms are often large enough (if taken seriously) to outweigh any benefits of the research.
My second argument is that any study of the effects of manipulating religious beliefs that uses informed consent will be on a sample of participants to which our theories largely do not apply (those participants willing to subject their religious convictions to a coin flip):
(a) People have a right to determine their own religious beliefs.
(b) This requires informed consent of potential consequences of a treatment intended to change religiosity.
(c) The effect of changing the religiosity in populations who consent to having their religiosity determined by a coin flip is rarely the quantity of interest.
(d) Therefore we cannot experimentally learn about the effects of religion in most populations.
Taken together, these arguments suggest that it will be difficult to use experiments to study the political effects of religion because researchers cannot weigh the benefits of the research against the subjective costs, because people disagree about whether changes in religiosity constitute a cost; and the types of subjects who would consent to have their religiosity randomly manipulated are generally not representative of the population we would like to study. In short, the two axioms of ethical research – informed consent and risk-benefit analysis – lead to the conclusion that experiments manipulating religion may generally be unethical.
My fear is that these challenges will deter young scholars from working on issues of religion and politics. If experiments are privileged and professional rewards accrue to those who use them, young political scientists may naturally gravitate toward topics and questions for which conducting experiments is not fraught with ethical complications. Given the importance of religion in the politics of the Middle East and elsewhere, this would be a real loss.
However, challenges are not for experimentalists alone. Ethnographic styles of research face at least some of the ethical challenges above when undertaken in religious contexts. Ethnographers are not attempting to change the religiosity of others so those arguments do not apply, but certainly issues of informed consent can be problematic as my opening anecdote highlights.
My remedy is to propose that scholars should entertain unethical research designs – not in order to implement them (please do not!) but because pondering an unethical design can often lead to ideas about how it could be made ethical. If scholars are too quick to reject some approach as unethical, a breakthrough that allows the research to proceed will never occur. This work may be more difficult, but ultimately the importance of understanding the role of religion in Middle East politics demands that we face the inherent ethical hurdles head on rather than diverting our best efforts to more tractable topics.
Richard A. Nielsen is an assistant professor of political science at the Massachusetts Institute of Technology.