The Role of Screen-Level Data Builders in Constructing Data Sets and Predictive Policing Technologies

• In Event: 30821 - Open Refereed Roundtables
  In Refereed Roundtable (90 minute): Table 12

Abstract

Scholars have explored the role of discretion in the development and evaluation of public sector data sets and information technologies. However, the effect of mid-level bureaucrats on public sector data practices requires further study. Through conducting a content analysis of responses to public records requests we made to 50 state Departments of Corrections seeking prison entry data, we show that discrepancies with criminal legal system data are primarily the result of screen-level bureaucrat decision-making. In this way, data work becomes political theater as researchers must navigate and frame data requests to best suit the bureaucracy of a particular state. We provide context on how the ecosystems of big data for the public sector operate, offer a perspective of how bureaucracy affects the form and function of big data, and highlight the strengths and limitations of government data sets for future sociological studies and technology design. Ultimately, our research raises serious questions about the efficacy of predictive policing technologies (especially those that rely on big data), the usability of big data in crafting emerging technologies, and data transparency.

Authors

• Christine Head, University of California-Irvine

• Amelia Roskin-Frazee, University of California-Irvine