This PhD project is offered on a self-funding basis. It is open to applicants with funding or those applying to funding sources. Details of tuition fees can be found at www.uea.ac.uk/pgresearch/pgrfees.
A bench fee is also payable on top of the tuition fee to cover specialist equipment or laboratory costs required for the research. The amount charged annually will vary considerably depending on the nature of the project and applicants should contact the primary supervisor for further information about the fee associated with the project.
Description
Meta-analysis is an extremely popular statistical method for combining results from several studies with the intention of obtaining more accurate estimates and more powerful tests. There are many very good texts that describe meta-analysis methods and applications [Borenstein et al. (2011), Hartung et al. (2011) and Cumming (2013)]. The popularity of the use of meta-analytic methods continues to rise exponentially. Consequently, providing improved statistical methods which do not suffer from the problems faced by many existing methods is extremely important, Kulinskaya et al. (2013).
The standard fixed effect models and random effects models (REMs) for meta-analysis have been widely criticised, the former for being over-simplistic in assuming equal study effects, and the latter for assuming that inter-study variation can all be captured in one random observation per study. Thus despite the widespread usage of such models, it is important to develop somewhat more complex models that describe the within and between study variation more realistically.
This project aims to do this by introducing flexible 2-level fixed and random effects models. At level 1, these models allow for a large (comparable with within-studies sample sizes) unknown random number of clusters within each study. For example in a large epidemiological meta-study with individual studies measuring some continuous variable, each study itself could be carried out on an unknown number of clusters of individuals with the same genotypes. At level 2, an additional random variance component may be added to account for between-study variation, resulting in a more flexible ‘cluster REM’.
This project will contribute to the UEA-based part of the methodology stream of the ESRC-funded Business and Local Government) BLG centre. Developed methods will be used for analysis of the local government and business data held at the BLG.
Nr of positions available : 1
Research Fields
Computer science - Other
Career Stage
Early stage researcher or 0-4 yrs (Post graduate)
Research Profiles
First Stage Researcher (R1)