The Ethics of Randomized Experiments in International Development

The Ethics of Randomized Experiments in International Development

In 2009, a team of researchers set out to determine whether financial assistance could help low income households in Tangiers, Morocco, achieve improved access to water mains. In what is known as a randomized control trial (RCT), the researchers selected a random subset of low income households, and provided them with the necessary assistance to secure domestic water access. While the intervention was effective, the experiment led to increased water bills — bills that poor households could not necessarily afford. Financial consequences were endured by individuals enrolled in the experiment long after the researchers left. Subsequently, the study became the subject of ethical concern. 

Since the early 2000s, randomized control trials have become the gold standard for conducting research in the field of International Development. RCTs were first pioneered in the medical field, and involve the random allocation of treatment to one portion of a population, and the withholding of treatment from another. After a certain amount of time, researchers compare outcomes in treated and untreated groups to see whether a certain intervention has had a significant effect and could be used to inform policy. RCTs have been used to assess a variety of development-related interventions, from the provision of microcredit to alleviate poverty to the use of deworming to improve child education outcomes. 

While in the past, development economists were preoccupied with formulating grand, abstract theories of development, RCTs have motivated researchers in the field to seek answers to smaller, more immediate questions. Can bed nets help fight malaria? Can textbooks improve test scores? And can cash transfers alleviate poverty? In 2019, the Nobel Prize for economics went to three researchers who pioneered the use of RCTs in Development Economics for their “experimental approach to alleviating global poverty”. Today, RCTs continue to dominate research in international development, informing government policy, aid work, and funding decisions. Although RCTs have helped shift the field of international development from a problems-orientated to a solutions-oriented field, the ethical considerations that surround many of these experiments are rarely discussed. 

With many RCTs, ethical concerns are a byproduct of experimental integrity. Any RCT must restrict a potentially beneficial treatment to a subset of a population in order to determine the efficacy of an intervention. Randomized experiments often take several years to carry out in order to determine the long-term effects of a particular intervention. If positive results arise halfway through an RCT, the intervention cannot be applied to the rest of the population, as this would render the RCT inconclusive. Is it really ethical to withhold an intervention from certain individuals for the sake of experimental integrity in cases like these, where researchers already possess the resources and certainty to administer a beneficial intervention? 

For many researchers, ethical concerns like this are outweighed by the potential for an RCT to positively inform policy or funding decisions. For example, randomized experiments examining the effects of subsidies on the provision of malaria bed nets have helped spur millions in funding towards organizations that provide free bed nets. Yet, not all RCTs have cascading effects like this. In fact, most RCTs aren’t generalizable. Many RCTs involve the study of extremely specific populations, and the findings often aren’t replicable beyond specific political, social and institutional contexts. Some critics estimate that over 90% of RCTs lack such external validity. If this is the case, then there is cause to wonder whether the empirical benefits of experimental integrity really outweigh potential ethical concerns. 

Several RCTs, despite their utilitarian goals, may knowingly or unknowingly inflict harm. A 2019 RCT entitled “Is Your Pain My Gain?” saw researchers provide a subset of Kenyan villages with cash transfers to see whether the psychological well-being of those in neighbouring villages would be harmed. In 2007, researchers in Zambia assigned different prices to the same health product to see if higher prices would incentivize usage. While both experiments intended to inform policy, experimenting with psychological well-being and access to healthcare provokes an almost instinctive sense of ethical discomfort. In other cases, it is not the RCT itself that provokes ethical discomfort, but rather, the precedent set by the experiment. For example, in 2009, researchers conducted a randomized study during Mozambique’s presidential elections to see whether information campaigns could increase voter turnout. While their intervention was successful, the prospect of foreign entities altering local elections under the guise of experimentation strikes a similar sense of ethical concern. 

Ethical discomforts are often heightened when we consider power imbalances between foreign researchers and low-income subjects. One critic has argued that many RCTs carried out in low-income countries would likely not be carried out in the West; and that the vulnerable economic statuses of these experimental subjects render them expendable. While many RCTs do undergo ethical assessments, these assessments are most commonly carried out by the Institutional Review Board (IRB) of whichever country a researcher is from, rather than the country in which the experiment is taking place. As a result, local stakeholders are frequently excluded from voicing their concerns or consulting foreign researchers during the experimentation process. 

Linked to this, many RCTs fail to obtain informed consent from participants. In medical research, informed consent is achieved when a researcher describes to a patient the experiment they are carrying out, as well as the potential risks involved. Most RCTs conducted for development research rarely achieve this level of informed consent, since they require participants to be unaware of the intervention at hand. Yet, given that not everyone wants to be experimented on, informed consent should become a standard prerequisite for this type of research. A lack of informed consent cements individuals as subjects for experimentation rather than human beings with agency. Additionally, without informed consent, individuals cannot alert researchers ahead of time if they believe that a particular intervention might cause unintentional harm. 

While RCTs certainly can be useful tools for informing development research, researchers should take more time to consider the ethical concerns associated with such experiments and ask whether an RCT is always the best way to learn something. Researchers should avoid RCTs that exploit individual vulnerabilities (financial, physical or emotional), make greater efforts to attain informed consent, and include more comprehensive discussions of ethical concerns in their research work. Even though cost-benefit analyses can be difficult to arrive at, researchers can work with local scholars, stakeholders and IRBs to assess the potential consequences that an experiment may inflict, and avoid studies where the prospect of unintentional harm outweighs the potential empirical benefits. 

Edited by Sarah St-Pierre.

Photo credits: “Glasses notebook and laptop” by Dan Dimmock, published July 2017, licensed under Unsplash. No changes were made.

Leave a Reply

%d bloggers like this: