Multilevel models for estimating household effects in longitudinal analysis – Prof. Fiona Steele
Join us for the Data Sciences Speaker Series with Prof. Fiona Steele, Department of Statistics professor at the London School of Economics and Political Science. This talk is co-sponsored by the Data Sciences Institute and the Ontario Regional Centre of the Canadian Statistical Sciences Institute (CANSSI Ontario) , University of Toronto.
Registration HERE
- Date: April 15, 2024
- Time: 11:00am – 12:00pm
- Format: In-person
- Location: Department of Statistical Sciences, 9th floor Seminar Room, 700 University Avenue, Toronto
Talk Title: Multilevel models for estimating household effects in longitudinal analysis
Description:
There is considerable interest among social scientists and epidemiologists in the extent of correlation among coresident adults in their attitudes, behaviours and other outcomes. Household panel studies are an important source of longitudinal data on coresidents, but a major challenge in accounting for household effects is the complex association structure arising from changes in household composition over time. Consequently, most previous research has either considered household effects only at a cross-section, or carried out longitudinal analysis that ignores household effects. In this talk, I will describe a multilevel modelling approach where household random effects may be correlated within clusters of households that share individuals over time, with correlations depending on covariates that characterise the connections between household pairs. The application of this ‘grouped’ random effects model is illustrated in analyses of household and area effects on physical and mental health in the UK. More generally, the proposed approach can be applied in the analysis of repeated measures or multivariate data on clustered individuals, relaxing the traditional assumption of equal within-cluster between-individual correlations.
Biography:
Prof. Fiona Steele’s research interests are in developments of statistical methods that are motivated by social science problems. Her areas of expertise include longitudinal data analysis, multilevel and latent variable modelling, and models for complex covariance structures. She has worked on a range of applications in demography, education, family psychology and health. She has directed several research grants funded by the Economic and Social Research Council (ESRC), including the LEMMA node of the National Centre for Research Methods for which she led the development of a popular online training course on multilevel modelling which has over 30,000 registered users worldwide. Recent grants have been for projects on the interrelationships between housing transitions and fertility in Britain and Australia, and on methods for the analysis of longitudinal dyadic data with applications to intergenerational exchanges of support in Britain. She was elected a Fellow of the British Academy in 2009. Personal website: http://stats.lse.ac.uk/steele/