An introduction to Multilevel Modelling for Intersectionality Research: The MAIHDA approach

This training session is open to all ESRC and non-ESRC funded PhD and MA Social Research students within the WRDTP’s seven partner universities. PGRs from all seven interdisciplinary Pathways are welcome to attend.

Multilevel models allow researchers to use data that has a clustered structure – such as pupils nested within schools, or individuals within neighbourhoods. Recently, a version of multilevel models has been developed for the study of intersectional inequalities in individual outcomes. With the Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA), individuals are nested within their intersectional strata – that is their combination of various sociodemographic identity categories –for instance gender, age, ethnicity and social class.

This method has great potential for uncovering and understanding intersectional inequalities, where a combination of social identities combine in complex ways to produce societal (dis)advantage.

This training course will provide a short introduction to multilevel modelling followed by an introduction to the intersectional MAIHDA approach. This will include an understanding of the simplest two-level “random intercept” multilevel models, how to use this model to apply the MAIHDA approach, some key statistics that can be produced from that approach, and give advice on how to visualize the results that are produced.

Students will gain:

Learn the concepts underlying multilevel modelling

Learn how to fit a two-level “random intercept” multilevel model, in Stata or R

Learn the conceptual underpinning of the MAIHDA approach, incorporating intersectional theory to a multilevel modelling framework

Learn how to fit the two models that make up the MAIHDA approach, again in Stata or R

Learn about various potential extensions of the MAIHDA approach, with resources signposted.

The course will include interactive lecture sessions, and then computer practical exercises that can be undertaken in either Stata or R.

Guidance for attendees:
No pre-reading required. However students might wish to have a look at this paper prior to attending:
Evans, Leckie, Subramanian, Bell, Merlo (2024) A tutorial for conducting intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA). SSM-Population Health, online. https://doi.org/10.1016/j.ssmph.2024.101664
Computers will be provided, but students wishing to use their own laptops should ensure they have R or Stata installed on their computer. If using R, we recommend installing RStudio additionally. R and Rstudio are free, whilst students may be able to install Stata through an institutional licence.

Andrew Bell
Andrew Bell is a Senior Lecturer in Quantitative Social Sciences at the Sheffield Methods Institute, University of Sheffield. He in an interdisciplinary researcher, with work spanning social epidemiology, geography, political science, economics and sociology. His methodological interests are around multilevel modelling, including longitudinal analysis. He currently holds an ESRC grant developing the MAIHDA approach within the social sciences.
George Leckie
George Leckie is Professor of Social Statistics and Co-Director of the Centre for Multilevel Modelling at the School of Education, University of Bristol. His methodological interests are in the development, application and dissemination of multilevel, longitudinal, and dyadic data models to analyse educational and other administrative and survey data. He has worked on the MAIHDA approach on a number of grants funded by the ESRC and Swedish Research Council.