Estimating Heterogeneity in Sociological Effects: From Interactions to Machine Learning

• In Event: 2204 - New Approaches to Understanding and Addressing Inequality

Abstract

Variation in sociological effects across subpopulations of interest is ubiquitous. Sociologists routinely partition their samples into subgroups to explore how the effects of particular events or interventions vary, or treatments, often by variables like race and gender. Causal methodologists also explore how effects vary by selection into the treatment. In both cases, the key subpopulations are determined by the researcher based on theoretical priors. Developing machine-learning techniques, however, allow researchers, to explore sources of variation they may not have previously considered or envisaged, i.e. to explore data-driven treatment effect heterogeneity. In this paper, we analyze an important topic in the stratification literature, the effects of higher education on unemployment and low wage work, with well-defined theoretical guidelines as to effect heterogeneity of interest, and compare what we learn from conventional interaction and propensity methods to machine learning methods. We encourage researchers to follow similar practices in their work on variation in sociological effects, and offer simple yet powerful tools by which to do so.

Authors

• Jennie E. Brand, University of California - Los Angeles

• Jiahui Xu, University of California-Los Angeles

• Bernard Koch, University of California, Los Angeles

• Pablo Geraldo, University of California-Los Angeles