The utility of quantitative methods for intersectionality: a real world example using the Born in Bradford’s Better Start (BiBBS) cohort dataset

Intersectionality is a critical theory highlighting how social categories such as race and sex interact to create unique experiences of disadvantage and marginalisation, and how this reflects wider structures of power and privilege. There is ongoing discussion regarding suitable methods for incorporating intersectionality into quantitative research, with the exploration of methods in varied research contexts and with different datasets identified as key to finding the best approaches.

This study added to this research through applying quantitative methods to real-world data, to assess their utility for intersectionality. The highly diverse Born in Bradford’s Better Start (BiBBS) birth cohort dataset was used as a case-study. A logistic regression analysis acted as a baseline for comparison with five intersectional methods; four identified in previous research and one novel method identified by the author. Each was applied to the BiBBS data, to explore a real-world research question:

“Which sociodemographic and lifestyle factors are associated with engagement with the Better Start Bradford early years interventions?”

Strengths, limitations, and suitability of the methods when applied in practice were assessed, and were considered alongside the impact of the methods on the analysis outcomes and interpretations, to judge the utility of each for intersectional research. Tensions were evident between theory and practice in the application of the framework, and no single method was judged to be ideal for quantitative intersectional research when applied in this case-study. Multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) was the most effective method in the study context, while decision tree analysis showed promise for larger datasets, and latent class analysis (LCA) allowed a more nuanced understanding of the study population. Suggestions were given for future research, and a more flexible and pragmatic approach to quantitative intersectional analysis recommended.