Skill level











Cold Tolerance



Classical generalized linear models assume that marginal effects are homogeneous in the population given the observed covariates. Researchers can never be sure a priori if that assumption is adequate. Recent literature in statistics and political science have proposed models that use Dirichlet process priors to deal with the possibility of latent heterogeneity in the covariate effects. In this paper, we extend and generalize those approaches and propose a hierarchical Dirichlet process of generalized linear models in which the latent heterogeneity can depend on context-level features. Such a model is important in comparative analyses when the data comes from different countries and the latent heterogeneity can be a function of country-level features. We provide a Gibbs sampler for the general model, a special Gibbs sampler for gaussian outcome variables, and a Hamiltonian Monte Carlo within Gibbs to handle discrete outcome variables. We demonstrate the importance of accounting for latent heterogeneity with a Monte Carlo exercise and with two applications that replicate recent scholarly work. We show how Simpson’s paradox can emerge in the empirical analysis if latent heterogeneity is ignored and how the proposed model can be used to estimate heterogeneity in the effect of covariates.

Current Research Projects

Modeling Latent Effect Heterogeneity

This project seeks to develop semi-parametric Bayesian regression models to estimate latent heterogeneity in the effect of treatment variables and/or observed covariates.

Perceptions, Political Behavior, Polarization, and Welfare Attitudes (with Rob Franzese)

This project examines how socioeconomic positions of the individuals impact their perceptions about the mechanisms of reproduction of social inequalities, and how those perceptions affect their political behaviour. It seeks to understand how socioeconomic-position-dependent perceptions about the socioeconomic environment make some individuals but not others more susceptible to certain populist ‘otherizing’ appeals.

Information, Income Distribution, and Policy Preferences

Using nation-wide surveys and surveys experiments, this project investigates how different social groups process information about inequality and income distribution, how information affect their perceptions, and how those perceptions affect policy preferences.

Unsupervised Learning Methods and FMM for Election Forensics (with Walter Mebane)

This is a NFS-funded project with Walter Mebane to develop positive models to detect fraud in elections. The project uses unsupervised learning methods to estimate probability of fraud using election data.


  • (773) 702 8317
  • Division of the Social Sciences
    University of Chicago
    5730 S. Woodlawn Ave., Room 406
    Chicago, IL 60637, USA
  • DM Me