Beyond IRBs: Designing Ethical Review Processes for Big Data Research

by Simson L. Garfinkel

Big Data Research is something new, a practice rooted in large-scale data collection that frequently combines aspects of research, experimentation, and entertainment. Yet another mismatch between the worlds of Big Data research and IRB system are the different regulatory regimes experienced by researchers in academia, where work with human subject data is highly regulated, and many corporate researchers, where there are frequently no regulations because the work is not funded by the federal government.
One tempting approach is to erect a fence between big data research and IRBs, and provide data scientists with a new, market–‐based, voluntary regulatory process. While such an approach might be attractive to some, self-regulated data science research is likely to be even less successful than self–‐regulated standards for data collection, advertising, and a variety of other privacy–‐intrusive business practices. Self-regulation is rarely effective in protecting one party when there is a significant power imbalance in a relationship. After all, it was the failure of self–‐regulation in the 1950s and 60s that led to the Belmont Report, the Common Rule, and the current system of regulation for research involving human subjects.