What Data Can Do: A Typology of Mechanisms

Abstract

This article offers an analytical framework for understanding the effects of data on the social world. Specifically, I ask: What happens when new data —digital or not—is introduced in a given context? Drawing on a mix of historical and contemporary examples, I provide a typology of 5 mechanisms: tracking, homogenizing, triaging, nudging, and valuating. I then demonstrate how this framework can change how we understand two empirical cases involving data-driven software programs, one in Web journalism and the other in criminal justice. Based on this analysis, the article makes three main points. First, at a time of rapid technological development, we should pay particular attention to what is not changing with digitization. Second, we need further theoretical integration in the rapidly growing field of critical data studies. Third, I suggest that the umbrella concept of “data” should be broken down into smaller and more manageable components depending on the mechanisms scholars are interested in studying.