Disentangling the Heterogeneity in Minority Stress: A Data Mining Approach

Abstract

Minority stress theory posits that discrimination contributes to increased distress among individuals with minoritized identities, including individuals who identify as lesbian, gay, and bisexual. Theory-driven studies have increasingly shown that intersectionality is critical to inform research on distress and discrimination among sexual minorities. Extant literature, however, has often examined discrimination and distress as homogeneous constructs. Overlooking nuances in specific forms of discrimination and distress may mask the potential effects of intersectionality in investigating sociodemographic factors that intersect with minoritization in sexual minority populations. Additionally, empirically deriving intersections among a broad set of sociodemographic factors in terms of their influences on discrimination and distress can offer new insights for population subgroups. The current study identified sociodemographic correlates (e.g., age cohort, race, gender identity, sexual identity, education, urbanicity, and poverty) related to patterns of specific distress and discrimination. These patterns are each related to subgroups within sexual minorities defined by intersections of these sociodemographic correlates. We applied a novel data mining method—the multivariate conditional inference tree—to a nationally representative sample of US adults of sexual minorities (N = 1518) from 2016 to 2018. We identified three groups—adults, adults in poverty, and bisexual and other sexual minority adults—that showed elevated patterns of minority stress. Theoretically, we found minority stress among sexual minority adults in the USA to be heterogeneous. Methodologically, we demonstrated that conditional inference tree modeling is an important tool to better position intersectionality studies moving forward to inform policy and intervention development.