RSSNI - Talk by Professor Fiona Steele, 22nd November, 2023, at 2pm, GMT.

Date: Wednesday 22 November 2023, 2.00PM
Location: Online & in Queen's University Belfast, UK.
Online MS Teams Link

And QUB Campus Map link,662793,en.pdf

Local Group Meeting

Share this event

This is the 2nd talk in RSSNI's Autumn schedule of seminars.
RSSNI - Talk by  Professor Fiona Steele,  22nd November, 2023 at 2pm GMT
RSSNI is pleased to announce our second talk of the Autumn session which will be given
on Wednesday, 22nd of  November @ 2pm GMT, online using MS Teams. In addition, one may attend the talk in person in the Mathematics & Physics Teaching Centre, Queen's University of Belfast (link), room 005, ground floor.

Speaker: Professor Fiona Steele, Department of Statistics, London School of Economics.

Title:   Marginal and random effects models for estimating household effects in longitudinal analysis. 

[This is joint work with Paul Clarke (ISER, University of Essex), Jouni Kuha (Departments of Statistics & Methodology, LSE) and Siliang Zhang (School of Statistics, East China Normal University).]
There is considerable interest among social scientists and epidemiologists in the extent of correlation among coresident adults in their attitudes, behaviours and other outcomes. However, in spite of the availability of longitudinal data on coresidents from household panel studies, most previous research has either considered household effects only at a cross-section, or carried out longitudinal analysis that ignores household effects. A major challenge in estimating household effects in longitudinal studies is the complex association structure due to changes in household composition over time. This presentation considers two approaches to the problem: marginal models where the changing correlation structure between individuals is modelled directly, and multilevel models with household random effects that may be correlated for households that share individuals over time. An advantage of the random effects approach is that higher-level clustering, for example within areas, can be accommodated. Both models are applied in analyses of UK panel data.

All welcome !

Please note that we are recording the talks now and they will appear in the RSS's U-tube channel shortly.

Gilbert MacKenzie  &
Felicity Lamrock
Professor Fiona Steele, Department of Statistics, London School of Economics, London, UK.
Gilbert MacKenzie for RSS Northern Ireland Local Group