“…Following calls to incorporate intersectionality into quantitative analysis (Bauer, 2014;Bowleg, 2012), a wide array of methodological proposals have appeared (Bowleg and Bauer, 2016;Bright et al, 2016;Evans et al, 2017;Gustafsson et al, 2016;Jackson, 2017;Jackson et al, 2016;Wemrell et al, 2017;Yette and Ahern, 2018) In this Issue, Bauer and Scheim present companion papers that call for an application of causal mediation analysis, VanderWeele's 3-way decomposition (Vanderweele, 2013) to study the mediating and interactive role of perceived discrimination in producing differences in psychological distress across groups defined by multiple social characteristics, race and transgender identity. While a formal framework for using causal mediation analysis and other decompositions to unpack intersectional disparities has already been outlined (Jackson, 2017), Bauer and Scheim add to this literature by developing new measures of perceived discrimination (Scheim and Bauer, 2019) and using the 3-way decomposition to study its contribution to intersectional disparities (Bauer and Scheim, 2019) through differential exposure and differential effects (sometimes in other literature referred to as differential vulnerability or susceptibility (Diderichsen et al, 2018)). In this commentary, we aim to: 1) review more nuanced causal interpretations with social characteristics; 2) review the shared features, strengths, and weaknesses of various forms of causal decomposition analysis for health disparities research; 3) examine how best to operationalize and distinguish differential effects from differential exposure under the goal of reducing disparities; 4) highlight critical identifiability issues when explanatory variables (mediators) such as perceived discrimination represent different constructs across social groups and suggest some ways forward.…”