Intersectionality in statistics

Statistics increasingly need to capture the complexity of inequalities experienced by diverse populations. This paper provides a conceptual and methodological overview of how an intersectional perspective can contribute to this objective by considering how multiple individual and social characteristics interact in complex, overlapping and non-additive ways within broader social and institutional contexts. It examines how intersectional analysis can reveal inequalities and population groups that may remain hidden in aggregate statistics or single-axis analysis, thereby supporting more targeted and evidence-based policymaking. The paper discusses key statistical challenges, including data availability and quality, analytical complexity, interpretation and communication, and ethical and privacy considerations. It reviews a range of complementary approaches, from descriptive and regression-based methods to multilevel modelling, cluster and latent class analysis, machine learning and mixed methods, highlighting their strengths, limitations and appropriate use. The paper provides practical guidance for integrating an intersectional perspective into statistical analysis and production.