The Ethics of Experimentation in Development Research

Randomized Control Trials (RCTs) are an experimental design used in development economics, in which people are randomly assigned to treatment or control groups in order to evaluate the impact of a certain policy intervention. These types of experiments can generate valuable data and contribute to the formulation of more effective policies. At the same time, they come with a huge responsibility. As a researcher conducting an RCT, you are experimenting with human subjects!

While disciplines like medicine and psychology have a long history of human subject research, it is a relatively new area for economists. Consequently, ethical guidelines and codes of conduct for economic experiments are not as well developed as the ones of their fellow colleagues. That is why the exploding rate of RCTs conducted in developing countries by development economists is especially alarming to me.

Macartan Humphreys, in his paper Reflections on the Ethics of Social Experimentation, summarizes my feeling of discomfort quite well, when he says:

In international development research these interventions can sometimes take the form of researchers from wealthy institutions manipulating citizens from poorer populations to answer questions of little interest to those populations.

Despite the noble intentions of the researchers, all too often development research becomes ‘extractive’ and reproduces already existing power differentials. ‘Extractive’ because often development research seems to benefit more the researcher itself, in terms of his publications and career, than the actual subjects of the research. Additionally, the existing power differential between the countries of the Global North and the Global South is even more confirmed by a white researcher travelling to a country of the Global South and taking the position of an experimenter, who has the right to intervene in local people’s lives. Experimental interventions have real-world consequences, the consequences can be harmless, but often they are not predictable and sometimes they are even adverse.

Consider the following scenario:

What if a Kenyan researcher travelled with a significant budget (according to local standards) to the US, arguing that American democracy is not as sane as it should be and set up an experiment seeking to enhance poor people’s electoral participation?

Probably he would not even get a visa.

I assume at lot of US citizens would be outraged. They would say it is a foreign interference into their domestic politics and undermining their nation’s sovereignty.

The perceptions of what constitutes development research and what crosses the line to foreign interference is context-dependent. The power differential between countries of the Global North and countries of the Global South is the consequence of history. Not always, but often countries of the Global North are tied to countries of the Global South through a history of colonial intervention, which has shaped the perception of the other in people’s minds. Our intention to conduct meaningful research, which has real-world relevance, makes us ignore this uncomfortable past way too easily.

Jacobus Cilliers, Oeindrila Dube and Bilal Siddiqi show in their new paper, which is aptly called The white-man effect, that race and nationality play a significant role, when conducting experiments, as they signal differences in wealth, power and authority.

They played the dictator game in 60 villages in Sierra Leone.

(Just a little reminder: A dictator game is a game in experimental economics with two players. Player 1 is the “dictator” and player 2 the “recipient”. Player 1 receives a sum of money and decides how the money will be shared between the two players and consequently player 2 receives the amount of money given to him by player 1.)

The dictator game was conducted by a team of five members. Four of them were Sierra Leoneans, including the team leader, and the fifth individual varied randomly. In control areas, he was a Sierra Leonean, in treatment areas he was a white foreigner. The team leader was the same person throughout all the games, he played an active role and instructed the players. The fifth person was not allowed to talk to the players and his task was to hand out money at the start of the game.

What did they found out?
The mere presence of a white foreigner affected the generosity of players.

Why?
There are two hypotheses:
1. Players perceive that the white person wants them to give more money and they want to impress him.
2. Players perceive that the white person is testing them for aid allocation and they give less to signal their ‘need’.

RCTs are not a panacea for the problems in this world. Ethical issues surrounding this new experimental approach in development economics have not been thoroughly reviewed and discussed within the profession. I would like to hear your opinion about it, what do you think about experiments in development economics? Under which conditions are they ethically acceptable and when not?

 

Sources:

  • Baele, Stéphane J. (2013): The Ethics of New Development Economics: Is the Experimental Approach to Development Economics morally wrong? In: The Journal of Philosophical Economics 7 (1).
  • Cilliers, Jacobus; Dube, Oeindrila; Siddiqi, Bilal (2015): The white-man effect. How foreigner presence affects behavior in experiments. In: Journal of Economic Behavior & Organization 118, S. 397–414.
  • Humphreys, Macartan (2015): Reflections on the Ethics of Social Experimentation. In: Journal of Globalization and Development 6 (1).
  • Photo by stockvault

3 thoughts on “The Ethics of Experimentation in Development Research

  1. Oh the irony – isn’t it funny that you use a study that implemented RCT and economic experiments, in order to uncover the ethic flaws in economic experiments?

    All joking aside – I would agree on the fact that we need a discourse about ethical guideline in economic experiments, and ultimately a code of conduct. Although, I don’t see them interfering as much with the lives of the “human guinea pigs” as in medical experiments.

    I see some parallels to the debates about colonialist views in other fields of science and how they influence the research. Edward Said’s book Orientalism strongly changed the way orientalists conducted their research and helped the discipline to grow out of its colonial projections.

    Nevertheless, I think that experiments are a very valuable tool.
    First, they are used all over the world, with changing experimenters and probands. Both, World Bank staff in Washington and farmers from remote villages in Kenya have been test persons. So it is not just the First World conducting ethically questionable experiments in the Third World.
    Second, isn’t it better to try to find out what impact my “development tools” have on the recipient on a small scale and in a controlled environment, rather than just implementing the measures without?

    So I agree, we need to think about ethics of experimentation, but that should not keep us from using it, as long as it is appropriate.

    [ Could you maybe give an example of an economic experiment that crossed the line?]

    Like

  2. This topic, randomized controlled experiments (RCTs) in development economics, has been a heated debate recently. I find it quite interesting since it is an integration of economics and other majors like psychology or philosophy. But the challenging point, as you perfectly have illustrated, is that whether it is morally wrong or not.
    It is a difficult question to answer. All economic phenomenon have special costs and benefits and two sides of a coin. In my opinion, it would be better argue how we can take an optimum advantage of it. Basically, as an economist, I recommend we find a way to deal with it optimally. In this regard, I think it is beneficial to have a specific framework or guidelines in order to assess the ethics of experimental method of development economics. More importantly, it is so substantial that scholars find these guidelines not a pressure but an increase in morality. From my perspective, the latter is much more helpful. At the end, they should find it a possibility for improving humanity.

    Like

  3. First of all, thank you both for taking time to write such an elaborate response. I basically agree with what you have said (also on the irony 😉 however there are some aspects where I would like to add some things.

    I also see a difference between medical and economic trials that is why I believe that it is crucial for development economists to develop their “own” guidelines and not simply apply guidelines of other disciplines. I also acknowledge that this is a process and needs time.

    As I already said, experiments generate valuable data and they are conducted all over the world. RCTs in developing countries that I am talking about, are only a small fraction of all experiments taking place. However usually, not always, experiments in developing countries are at least partially guided by foreign nationals, whereas this is usually not the case for experiments in developed countries. I chose to highlight this small fraction of experiments because of its relevance to the field of development economics (and this blog), which is admittedly a narrow scope on experiments.

    The results of RCTs are highly context-dependent, which means there is a possibility that the same policy can have a different impact even in the neighbour village. While you can argue that RCTs are better than not testing at all, I am not sure how reliable the evidence is, when you go from small-scale intervention to large-scale intervention. To say it in more technical terms, while are RCTs are strong on internal validity, they suffer from external validity.

    One type of experiments that I consider risky are “political” RCTs. Especially those related to elections, since in politically instable countries, during times of political tension (like during election), even small-scale interventions can have unexpected real-world consequences, for example influencing the outcome of the election. Furthermore these type of RCTs usually promote the political ideal of democracy, which is an ideal of “Western” society, but does it necessarily mean that all countries have to aspire to become democracies?

    Another “famous” example is Bertrand et al. (2007) “Obtaining a driver’s license in India: An experimental approach to studying corruption”. They offered one of the treatment groups a financial reward if they got a driver’s license within 32 days and thereby assumed the only way participants could do so, is bribing officials. This is what they finally did. This experiment “financed” a driver’s license for people, who did not know how to drive (although they were offered free driving lessons after the study was completed). It also “encouraged” them to bribe. For me the finding of the paper is quite intuitive and does not benefit the participants at all.

    To conclude, my point is not that we should not do RCTs, I do not think that RCTs are per se morally wrong. However, I feel that the accompanying discussion about how to deal ethically with this new responsibility and what is acceptable on moral grounds in terms of experimental design and what not, receives little attention by the profession. As you already said, we should not see ethics as a constraint, but rather as a tool, which will help us to make better experiments in the future.

    Like

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s