Research Interest

  • Political Methodology ▾
  • International Political Economy ▾

Contact Information


Work in progress

    Experimental Evaluation of Computer-Assisted Human Decision-Making: Application to Pretrial Risk Assessment Instrument.
    Joint with Kosuke Imai, Zhichao Jiang, D. James Greiner, and Ryan Halen.

    Despite an increasing reliance on computerized decision making in our day-to-day lives, human beings still make highly consequential decisions. As frequently seen in business, healthcare, and public policy, recommendations produced by statistical models and machine learning algorithms are provided to human decision-makers in order to guide their decisions. The prevalence of such computer-assisted human decision making calls for the development of a methodological framework to evaluate its impact. Using the concept of principal stratification from the causal inference literature, we develop a statistical methodology for experimentally evaluating the causal impacts of machine recommendations on human decisions. We also show how to examine whether machine recommendations improve the fairness of human decisions. We apply the proposed methodology to the randomized evaluation of a pretrial risk assessment instrument (PRAI) in the criminal justice system. Judges use the PRAI when deciding which arrested individuals should be released and, for those ordered released, the corresponding bail amounts and release conditions. We analyze how the PRAI influences judges' decisions and impacts their gender and racial fairness.

    A Flexible Bayesian Ideal Point Estimation Method for Correlated Multidimensional Ideal Points.
    Joint with Johan Lim and Jong Hee Park.

    We present a flexible Bayesian method for multidimensional ideal point estimation that aims to uncover correlated ideal points from multidimensional space. We first report that existing IRT and NOMINATE–based methods have difficulties in recovering multidimensional ideal points. The new method (1) substitutes the ℓ2 norm (Euclidean distance) of existing models with the ℓ1 norm (Manhattan distance) and (2) employs a multivariate slice sampling method. We apply the proposed method to the UNGA roll–call voting data.


  • polnet

    R package for the Latent Space Network Model (LSM) and bipartite Link Community Model (biLCM) developed by Kim and Kunisky (forthcoming in Political Analysis).