Data Bias Meets Cognitive Bias: The Impact of Conveying Data Uncertainty in Human Rights Decision-Making

Tamy Guberek


Having spent years studying the production of silence and absences in archives, datasets and other collections of information about these kinds of events (Guberek & Hedstrom, 2017; Lum et al., 2010; Guzmán et al., 2012; Roth et al., 2011), I am motivated by the implication of these silences when records and data are used in downstream analysis, knowledge production and communication. Information about violence are collected at high risk under repressive governments, recorded during the fog of war by watchdog organizations, or produced via administrative reporting requirements in various institutions. Missing data may be intentional (concealment, censorship, etc.) or inadvertent. Missing information is rarely randomly distributed and complete data is almost never attainable. One of most consistent recommendations by scholars concerned with the consequences of data bias is that data providers should communicate data limitations to information consumers (Root, 2015; Price & Ball, 2015; Greenhill, 2010; Lum & Isaac, 2016; Cohen & Hoover Green, 2012). The underlying assumption is that doing so will enhance the credibility of the information and the information provider, ultimately leading to “better” decisions. At the same time, while human rights scholars recommend that data providers transparently communicate data limitations to data consumers, the literature from cognitive science and science communication suggests that expressing data uncertainty could be ineffective, unproductive, or neutral at best (Kahneman, 2011, Lupia, 2013, Joslyn & LeClerc, 2013). In fact, there is little empirical understanding of how presenting the process of record production and their limits is received by users, and how that presentation interacts with cognitive biases. This is the research I am undertaking for my Ph.D. dissertation, which uses experimental and qualitative methods to explore the impact of conveying data uncertainty on decision-makers. I propose to present the conceptual frame, methods and progress of the research at AERI 2017. While this research is inter-disciplinary in nature, I believe it is relevant for the archival research community, as it is motivated in large part form lessons on administrative record making and other forms of documentation as evidence of abuses. The study’s findings will shed light on the implications of silences in records in downstream analysis and communication – relevant for archival, human rights and science communication fields, and also on fields researching evidence-based decision-making.

Cohen, D. K., & Green, A. H. (2012). Dueling incentives: Sexual violence in Liberia and the politics of human rights advocacy. Journal of Peace Research, 49(3), 445-458.
Guberek, T., & Hedstrom, M. (2017). On or Off the Record?: Detecting Patterns of Silence about Death in Guatemala’s National Police Archive. Archival Science, 17(1), 27-54.
Kahneman, D. (2011). Thinking, fast and slow. New York: Farrar, Straus and Giroux.
Lupia, A. (2013). Communicating science in politicized environments. Proceedings of the National Academy of Sciences of the United States of America, 110, 14048-14054.
Root, B. (2015). Numbers are Only Human : Lessons for Human Rights Practitioners from the Quantitative Literacy Movement. In P. A. a. S. Knuckey (Ed.), The Transformation of Human Rights Fact-Finding: Oxford University Press.



Tamy Guberek is a Ph.D. Candidate at the University of Michigan’s School of Information, where she researches the production, analysis and consumption of information about violence. She earned her M.A. in World History at Columbia University and her M.Sc. in International History at the London School of Economics. For over a decade prior to graduate school, she worked as a researcher and advisor to non-governmental organizations and state agencies, conducting surveys and analyzing multiple sources of information on human rights questions in various countries around the world.

Leave a Reply

Your email address will not be published. Required fields are marked *