Working Papers
- "Inference at the data's edge: Gaussian processes for estimation and inference in the face of extrapolation uncertainty" (with Chad Hazlett and Doeun Kim) [Minor revision, Political Analysis] [arXiv] [software]
- "Non-existent outcomes in research on inequality: A causal approach." (with Ian Lundberg) [R&R, Sociological Methods & Research] [arXiv] [software]
- "When help seems optional: Why climate refugees face greater skepticism than war refugees." (with Jieun S. Park and Margaret Peters; co-first author with Park) [in draft]
- "Causal inference without controls: Gaussian Processes for estimating treatment effects in Interrupted Time Series." [Job Market Paper]
Abstract
Many inferential tasks involve fitting models with observed data and predicting outcomes at new covariate values, requiring interpolation or extrapolation. Conventional methods select a single best-fitting model, discarding fits that were similarly plausible in-sample but would yield sharply different predictions out-of-sample. Their uncertainty estimates thus ignore ``extrapolation uncertainty." Gaussian Processes (GPs) offer a principled alternative. Rather than committing to one conditional expectation function, GPs deliver a posterior distribution over outcomes at any covariate value. This posterior effectively retains the range of models consistent with the data, widening intervals where extrapolation magnifies divergence. In this way, GPs address extrapolation uncertainty, helping to tame the “dangers of extreme counterfactuals” \citep{king2006dangers}. The approach requires specifying (i) a covariance function linking outcome similarity to covariate similarity, and (ii) Gaussian noise around the conditional expectation. We provide an accessible introduction to GPs with emphasis on this property, together with a simple, automated procedure for hyperparameter selection implemented in the R package \texttt{gpss}. We illustrate the value of GPs for capturing counterfactual uncertainty in three settings: (i) treatment effect estimation with poor overlap, (ii) interrupted time series requiring extrapolation beyond pre-intervention data, and (iii) regression discontinuity designs where estimates hinge on boundary behavior.Abstract
Scholars of social stratification often study exposures that shape life outcomes. But some outcomes (such as wage) only exist for some people (such as those who are employed). We show how a common practice—dropping cases with non-existent outcomes—can obscure causal effects when a treatment affects both outcome existence and outcome values. The effects of both beneficial and harmful treatments can be underestimated. Drawing on existing approaches for principal stratification, we show how to study (1) the average effect on whether an outcome exists and (2) the average effect on the outcome among the latent subgroup whose outcome would exist in either treatment condition. To extend our approach to the selection-on-observables settings common in applied research, we develop a framework involving regression and simulation to enable principal stratification estimates that adjust for measured confounders. We illustrate through an empirical example about the effects of parenthood on labor market outcomes.Abstract
Climate-induced migration presents a pressing global challenge, with millions projected to face displacement, yet public attitudes toward climate refugees remain understudied. Existing research treats refugees as monolithic or exclusively war-driven, obscuring how displacement causes shape public perceptions. We argue that climate refugees receive systematically less public support than war refugees due to heightened suspicion about their economic motivations—a skepticism particularly pronounced in developed countries where robust disaster management systems create expectations that climate impacts should be handled domestically. Drawing on cross-national survey data from 28 countries and two original survey experiments in the United States, we test how motivational suspicion shapes refugee preferences. We first find that this climate refugee disadvantage concentrates in developed countries, where suspicion about economic motivations predicts the support gap. Second, experimentally priming such suspicion disproportionately erodes climate refugee support. Third, displacement cause generates stronger attitudinal differences than regional origin, with climate refugees facing skepticism about displacement legitimacy regardless of demographics. These findings extend group-specific attitude theories to climate-displaced populations and suggest that building public support requires addressing legitimacy concerns beyond humanitarian appeals.Research in Progress
- "The effects of COVID-19 Stay-at-home orders on gun purchasing." (with Jack Kappelman, Tanvi Shinkre, Ryan Baxter-King, Haotian (Barney) Chen)
- "Drivers of attitudes toward Asian Americans: Disentangling stereotyping from racial identity." (with Jieun S. Park and Jessica Hyunjeong Lee)
- "Modeling and assessing controlled direct effects with the regression-with-residuals method." (with Ian Lundberg)