led by Dr Natalie Bennett, University of Sheffield. This session is run in partnership with White Rose DTP.
Multilevel models allow researchers to analyse data with a clustered structure—for example, pupils nested within schools or individuals within neighbourhoods. Recently, a variation of multilevel modelling has been developed to study intersectional inequalities in individual outcomes. The Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) approach nests individuals within their intersectional strata—that is, their unique combination of sociodemographic identity categories, such as gender, age, ethnicity, and social class. This method holds great potential for uncovering and understanding intersectional inequalities, where multiple social identities interact in complex ways to shape societal (dis)advantages.
This one-day training course will provide a brief introduction to multilevel modelling, followed by an overview of the intersectional MAIHDA approach. The course will cover the basics of two-level random-intercept multilevel models, how to apply this model within the MAIHDA framework, key statistics generated by the approach, examples from the literature, and guidance on visualizing the results.
Topics
1. Overview of multilevel modelling
2. The two-level random-intercept model
3. Intersectionality: theory and practice
4. Ways of statistically identifying intersectional inequalities: dummy variables and interactions
5. The MAIHDA approach
6. Visualising results from MAIHDA
7. Conceptual and practical challenges with MAIHDA
8. Extensions of MAIHDA
Format
The course will consist of a mix of lectures and hands-on practical sessions applying the taught methods to real datasets (the lectures are software independent). Practical sessions will follow lectures, giving participants the chance to replicate the presented analyses and to consolidate their knowledge. The practicals are offered in participants’ choice of R or Stata and are self-directed: participants complete the practicals at their own pace. At the end of each practical session the instructors demo the different software.
Pre-requisites
We assume no prior knowledge of multilevel modelling. However, participants should be familiar with estimating and interpreting linear regression models, including the writing and interpretation of model equations, hypothesis testing and model selection, and the use and interpretation of dummy variables and interaction terms. We assume you are already users of R or Stata and so have these softwares already installed and know the basics. Participants may wish to have a look at this published MAIHDA tutorial paper prior to attending, although it is not required: Evans, Leckie, Subramanian, Bell, Merlo (2024) A tutorial for conducting intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA). SSM-Population Health, online.
You will need a laptop with up-to-date versions of R or Stata installed, depending on your preference.