Individuals do not respond uniformly to treatments, such as events or interventions. Sociologists routinely partition samples into subgroups to explore how the effects of treatments vary by selected covariates, such as race and gender, based on theoretical priors. Emerging machine learning methods based on decision trees allow researchers to explore sources of variation that they may not have previously considered. In this talk, Brand describes a range of approaches to study effect heterogeneity, including tree-based machine learning. Assessing a central topic in social inequality, college effects on socioeconomic outcomes, she compares what we learn from covariate and propensity-score-based partitioning approaches to recursive partitioning based on causal trees.
