Design and analysis of randomized experiments: robust and efficient inference with minimal modeling assumptions
Observational studies: critical view of the key assumptions of unmeasured confounding and overlap of covariates
Natural experiments: instrumental variable, difference-in-differences, synthetic controls, and panel data
Causal mechanisms: heterogeneous treatment effect, principal stratification, and mediation analysis
Data integration: combining multiple sources with heterogeneous quality, and meta-analysis
Survey sampling, missing data, and measurement error
Fundamental properties of linear model: dimensionality, misspecification, interpretation
Correlated data analysis: clustered data, network analysis, causal inference with interference
Intersection of Frequentist and Bayesian statistics: Professor Carl Morris called it FB (Morris' football)
Applied statistics in social sciences and biometrical studies