Human + Machine Decision-Making
Predictive algorithms are being used increasingly to help address complex social policy issues—who is likely to commit a crime if released from jail, who is likely to have a heart attack soon, or who will perform well in a job? My research focuses on the human aspects of these tools, by developing descriptive and normative theories of (1) when people want to use algorithms, (2) when people want algorithms to be used, and (3) how our minds affect what algorithms learn. This work is often conducted in high-stakes field settings and attempts to grapple with the many institutional and personal influences at play.
When To Use Algorithms
How core psychological mechanisms like image concerns, automaticity, and motivated reasoning impact algorithmic decision-making.
EXAMPLE WORK:
Automating Automaticity: How the Context of Human Choice Affects the Extent of Algorithmic Bias w/ A. Agan, J. Ludwig, and S. Mullainathan
Media coverage: The Pie, econimate
Predictably Bad Investments: Evidence from Venture Capitalists
Media coverage: Bloomberg, Chicago Booth Review, Forbes, Hacker News, Institutional Investor, Marginal Revolution, MarketWatch, The Telegraph
Discriminatory Discretion: Theory and evidence from use of pretrial algorithms
Upcoming Work
Various lab and field projects on economic, political, and psychological mechanisms impacting algorithmic decision-making.