This is a free one-day workshop sponsored by the MRC IEU, MRC grant MR/N027485/1 and ESRC grant ES/R010285/1

Overview:

This workshop will introduce new methods for modelling within-individual (or within-group) variation in repeatedly-measured outcomes. As well as talks introducing a variety of case studies, there will be practical sessions to demonstrate how to use multilevel models to analyse within-individual (or group) variation (a.k.a. mixed effects location scale models), with an extension to joint models relating differences in within-individual (or group) variation to an individual- (or group-) level outcome.

These models can apply to within-individual variation – if the dataset consists of observations repeatedly-measured within each person – but also to any scenario in which observations are repeatedly-made within a group of interest: for example exploring variation within clinics, schools, neighbourhoods, and other clusters. For instance, these techniques have been used to investigate: the factors related to within-individual variation in blood pressure, including association with later biomarkers of cardiovascular disease (Barrett et al., 2019); within-school variability in pupil achievement (Leckie et al., 2014); within-individual variation in weight gain in pregnancy (Goldstein et al, 2017); within-interviewer variability in survey responses (Brunton-Smith et al., 2017); the factors associated with within-individual variation in mood as assessed via ecological momentary assessment (Hedeker et al., 2008); etc.

As part of this workshop, we will demonstrate how to fit this general class of model in a Bayesian framework using R, Stan and WinBUGS. Participants are asked to bring laptops for use in the practical sessions, with the relevant software installed (further instructions and advice regarding software installation will be given once attendance is confirmed).

Speakers & panel discussants:

Jessica Barrett (MRC Biostatistics Unit, University of Cambridge)

Ian Brunton-Smith (Department of Sociology, University of Surrey)

Harvey Goldstein (Centre for Multilevel Modelling, University of Bristol)

George Leckie (Centre for Multilevel Modelling, University of Bristol)

Graciela Muniz-Terrera (Centre for Clinical Brain Sciences and Centre for Dementia Prevention, University of Edinburgh)

Richard Parker (MRC Integrative Epidemiology Unit, University of Bristol)

Kate Tilling (MRC Integrative Epidemiology Unit, University of Bristol)

Who can attend?

The workshop will be most relevant to anyone with an interest in modelling within-individual (or group) variation, including medical researchers, epidemiologists, social scientists, psychologists, etc. Knowledge and experience of using multilevel models to analyse clustered data is essential.

To book a place: please email Angharad Davies (angharad.davies@bristol.ac.uk)

Queries: Further details regarding timings for the day and a final programme will be provided in early October 2019. Please contact Richard Parker (richard.parker@bristol.ac.uk) in the meantime with any queries.

References:

Barrett JK, Huille R, Parker RMA, Yano Y, Griswold M. (2019) DOI:10.1002/sim.8074

Brunton-Smith I, Sturgis P, Leckie G. (2017) DOI:10.1111/rssa.12205

Goldstein H, Leckie G, Charlton C, Tilling K, Browne WJ. (2017) DOI:10.1177/0962280217706728

Hedeker D, Mermelstein RJ, Demirtas H. (2008) DOI:10.1111/j.1541-0420.2007.00924.x

Leckie G, French R, Charlton C, Browne W. (2014) DOI:10.3102/1076998614546494