My Blog

In the course Ethical Data Futures, I studied some cased which refers to ethical implications. One of these impressed me the most. I’d like to share my ideas here.

The case is that Facebook users were shown a “social message” at the top of their News Feed. The social message encouraged them to vote and displayed a clickable button reading “I Voted” as well as a counter indicating how many other Facebook users had previously reported voting, and—most importantly— showed up to six randomly selected thumbnail profile pictures of the user’s Facebook friends who had reported having already voted (by clicking the “I Voted” button). In addition, a further half a million users were shown no message at all (the control group). A validation study for this experiment showed that users who received the social message were 0.39% more likely to turn out to vote than users who received the informational message or no message at all. When you consider social contagion (subjects influencing their friends, and friends of friends, both on- and offline), the effects were still greater.

What is worth discussing, in this case, is its ethical problems. Firstly, whether there is some potential negative impact on the society and country after changing election results. In the context of elections, according to the research(Slough, T., 2019), even expert researchers and their local partners can fail to correctly predict the outcome of their interventions. It can imply that they do not have sufficient information to anticipate harm. Indeed, Baele (2013: p. 28) suggests that “artificial interventions in tension-ridden political episodes such as elections may provoke unpredictable chain reactions.” Experiments on elections can generate social harm by changing who wins office. Also, Whitfield (2019: p. 7) argues more broadly for political research ethics that respect the “self-determination of communities.

In addition, I think there are also ethical issues involved: whether to ask for the consent of the participants. Like most field experiments, electoral experiments are generally conducted without the consent of subjects or non-subjects who may be affected by the intervention. We need to think about how experimentalists might seek consent in new ways or provide additional protection to subjects and non-subjects in their absence. Further development of and debate about these alternatives is important to address issues of consent.

Reference List

Slough, T., 2019. The ethics of electoral experimentation: design-based recommendations. New York: Columbia University. Working Paper. http://taraslough. com/assets/pdf/eee. pdf.

Baele, S., 2013. The ethics of new development economics: Is the experimental approach to development economics morally wrong?. Journal of Philosophical Economics, 7.

Whitfield, G., 2019. TRENDS: Toward a separate ethics of political field experiments. Political Research Quarterly, 72(3), pp.527-538.