Abstract:
Quantitative intersectionality research is gaining momentum in the social and health sciences as scholars increasingly recognize that individuals' lived experiences and outcomes are shaped by complex, inequitable relationships between social systems and their multiple intersecting social identities. However, the standard approach for investigating intersectional inequalities—regression with interactions—struggles to capture high-dimensional interactions between social identities as theorized by intersectionality scholars. This limitation may lead to mischaracterizing inequalities, inefficiently targeting marginalized groups, and misallocating resources, all of which negatively impact both individuals and society.
Intersectional MAIHDA (Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy) is a new multilevel modelling-based approach for investigating intersectional inequalities, developed in social epidemiology. Proponents argue that MAIHDA provides improved predictions of intersection mean outcomes, more effectively distinguishes the contributions of sociodemographic characteristics from unmodeled individual-level factors, better separates the additive and multiplicative effects of these characteristics, and is able to identify high-dimensional interactions between sociodemographic factors.
In this talk, I will provide an overview of intersectional MAIHDA through an illustrative application to the study of intersectional sociodemographic inequalities in student achievement. This includes presentation of the underlying multilevel models and assumptions, an exploration of the statistical properties of different predictions of intersection mean outcomes, and a discussion of future directions and applications.