Mr Rob North is Postgraduate Research Student within Social Sciences: Social Statistics & Demography at the University of Southampton.
I was previously a student at University of Reading where I undertook my Undergraduate and Masters qualifications in Statistics. I then worked in the field of Transport Statistics at the Transport Research Laboratory (TRL) before returning to study for my PhD at Southampton.
My main research interests lie in the areas of Repeated Measures data and Generalized Linear Models and their applications. My Master’s thesis focused on count data and in particular Zero Truncated forms of Poisson and Negative Binomial models. My undergraduate thesis involved applications of Survival Analysis so I have some experience in this area also.
PhD Research Title
Joint Univariate Regression and Association Models for Longitudinal Categorical Data: Applications of the Dependence Ratio
Abstract
The methodological research planned concerns the area of repeated measures but with a categorical response as opposed to the more traditional normal response. The repeated nature of the data means that there are repeated responses for each experimental unit/cluster. Therefore standard univariate methods that assume independence cannot be used and specific methods need to be used to take into account the association within a cluster. A recently proposed method for this is the dependence ratio and it has been shown to offer a number of improvements over previous likelihood based methods that typically used odds ratios, using many different association structures. The research will involve using association structures that have not previously been proposed in relevant application areas such as social mobility and arthritis. It is also planned that part of the research will be dedicated to looking at a Bayesian approach to the dependence ratio technique, something which to our knowledge has not previously been looked at.
Supervisors
Peter W.F.Smith and John W. Mcdonald.
I am currently tutoring second years who are undertaking the STAT2009 module.