A Random Sample?
In 2019, economists Esther Duflo, Abhijit Banarjee, and Michael Kremer were awarded the Nobel Prize in Economics. Their recognition was driven by a groundbreaking means of testing policy interventions and economic aid: randomized controlled trials (“RCTs”) [1]. Though mainstream in the world of medicine and clinical trials, RCTs are a recent but growing development in social science research and economics.
In an RCT, one group receives an intervention, while another does not and hence serves as the “control” group. As the experiment progresses, researchers analyze differences between the experimental group and the control group. Other than the presence of the intervention, the two groups should be similar: this seeks to ensure that any observed effects are driven by the intervention, rather than other external variables. This active experimental approach allows researchers to test for causation much more effectively than simply analyzing existing data that might be influenced by a variety of unobservable variables. Since the 1990s, researchers have begun to employ RCTs to understand the effects of policy interventions, oftentimes in developing economies.
One example is an ongoing experiment conducted by non-profit GiveDirectly. Over a period of 12 years, GiveDirectly will issue direct cash transfers of $22 per month to randomly selected adults in a village in Kenya [2]. The objective of the experiment is to test the effectiveness of direct cash transfers in alleviating poverty. In addition to large-scale experiments such as GiveDirectly, many governments have also implemented RCTs to analyze smaller-scale policy interventions.
Many acknowledge that valuable insights stand to be gained from these types of RCTs, as it is essential to have an accurate understanding of how to effectively allocate aid and develop policy. Still, some have questioned whether or not RCTs in a social science setting are ethical. First, many contend that RCTs are exploitative, as they render human beings as experimental subjects that are merely utilized for data. Moreover, some have pointed out that it is immoral to issue a benefit to one portion of the population but withhold it from another, particularly in impoverished areas and developing economies. This inevitable inequality might also create discontent and tension within the community in which the experiment is conducted. As the number of individuals participating in RCTs continues to grow, stakeholders must evaluate the extent to which these “human experiments” can be justified.
DISCUSSION QUESTIONS
Is the GiveDirectly cash transfer experiment morally justified
Are random selection effects like hose functioning in this case inherently unfair? Why or why not?
Is there a difference between conducting RCTs in impoverished communities and RCTs conducted to test policy interventions in more affluent nations?
References
[2] NPR, “An Experiment Gives Cash Aid To The Poor. Is That Ethical?”
