Network and School Variations in Adolescents’ Health Behaviour and Educational Attainment: A Multilevel Analysis of UK data
Keywords – Social Network analysis; Multilevel Modelling; Health Behaviour; Educational Analysis; Variation; Social Statistics; Spatial Models; Peer Effects; Social determinants of health behaviour.
Project Summary - Previous multilevel studies of pupils in schools suggest that much of the inequality in pupils’ health and educational outcomes is between schools. However, part of the variation in these outcomes may be due to friendship networks (peers). The proposed research will formulate and compare advanced statistical models for data on pupils in networks in schools to answer three key, interrelated, research questions: (i) To what extent are friendship networks associated with pupils’ inequalities in health and educational outcomes? (ii) Are these inequalities in part explained by friendship networks? (iii) do these associations differ by network composition, geographical location, or time?
The student will apply and compare school and network multilevel models to several UK datasets. New applications of Multiple Membership Multiple Classification (MMMC) and Conditional Auto-Regressive (CAR) models will be made for both cross-sectional and longitudinal friendship networks in the context of health and educational inequalities. The research will also consider inequalities by network composition, including gender network differences. Previous work by Moore on Welsh data investigated the link between friendships and the co-evolution of health behaviour. This project will formulate MMMC and CAR models to be compared with this previous work.
The project will use advanced quantitative methods – including non-hierarchical and non- linear multilevel statistical models, and involving the use of advanced statistical software, in the context of real public health data. The student will also undertake interpretation and presentation of the results of these analyses, for both academic and non-academic audiences from a variety of disciplines including Public Health. The student will benefit from doing research across the disciplines of health and social statistics, and from the interdisciplinary expertise of the supervisors in statistical modelling, social epidemiology, and intervention design. The student will present their work at international conferences from the primary discipline of each of the supervisors, such as the Society for Social Medicine Conference and the sunbelt and EUSN social networks conferences.
The project will use advanced quantitative methods – that is, linear and non-linear multilevel and spatial statistical models in the context of real public health data, and the PhD scholar will benefit from both advanced statistical training (via the UK-wide APTS programme – ww.apts.ac.uk) and interpreting the results in a public health and policy context. The enhanced doctoral training provided by the interdisciplinary team will also involve the use of advanced statistical software. The student will also undertake interpretation and presentation of the results of these analyses, for both academic and non-academic audiences. The student will also benefit from presenting at conferences from the primary discipline of each of the supervisors, such as the Society for Social Medicine Conference, the GEOMED spatial statistics meeting, and the Sunbelt and EUSN social networks conferences - which typically include sessions on the statistical analysis of social network data, including those organised by Tranmer.
Project Team –
1st Supervisor: Professor Mark Tranmer, School of Social & Political Sciences, University of Glasgow. mark.tranmer@glasgow.ac.uk
2nd Supervisor: Dr Duncan Lee, School of Mathematics & Statistics, University of Glasgow. duncan.lee@glasgow.ac.uk
3rd Supervisor: Professor Laurence Moore, Social and Public Health Sciences Unit (SPHSU), University of Glasgow. laurence.moore@glasgow.ac.uk
The student will be based at the Gilmorehill Campus (West End), University of Glasgow, and will also have the opportunity to visit and work at the Social and Public Health Sciences Unit (SPHSU) in the city centre.