Social Experimentation in an Unjust World

There is a resurgence of interest in social experimentation as a means of promoting social progress, including progress in justice. In this chapter, we first advance an argument in favor of social experimentation drawing on its capacity to resolve uncertainty both about how to achieve socially valuable goals and about which goals are worth pursuing. We then identify four challenges: the information problem (experiments may not yield relevant information), the selection bias problem (potentially informative experiments may not be undertaken), the uptake problem (the information generated by experiments may not be put to good use), and the risk problem (experiments may carry unacceptable risks). Finally, we argue that certain injustices can exacerbate all four problems, rendering social experimentation a less reliable path to progress, and, in cases of severe injustice, perhaps even a regressive force. The upshot is not that we should abandon social experimentation, but that we should temper our expectations and focus on constructing conditions under which experimentation is more likely to be progressive. Specifically, to render social experimentation a more reliable engine for social progress of any sort, we must remedy or mitigate the injustices that diminish its value.