New approaches for generalised linear multilevel models with missing covariate values
Event Details
Event Contact
Event Description
- Full Description:
- The talk will look at different ways in which missing data have been handled and will discuss the shortcomings of many popular methods.
Recent work implementing a fully Bayesian model, that extends the Rubin et al multiple imputation procedure and provides fully efficient and consistent estimators, will be discussed. One novel feature is that interaction and polynomial terms involving missing data are handled properly. Examples will be given and a possible application to modelling growth mixture models will be suggested.
Professor Harvey Goldstein is a chartered statistician, is currently joint editor of the Royal Statistical Society's Journal, Series A, has been a member of the Society's Council and was awarded the Society's Guy medal on silver in 1998. He was elected a member of the International Statistical Institute in 1987, and a fellow of the British Academy in 1996 and awarded an honorary doctorate by the Open University in 2001.
Recently he has been involved in a collaboration with Bill Browne, James Carpenter and Mike Kenward in developing methods for handling missing data values and measurement errors using Bayesian modelling and latent normal transformations.
He uses statistical modelling techniques in the construction and analysis of educational tests the implications of which he explored in a series of papers since 1977. He has been involved in a number of longitudinal studies of 'value added indicators' for comparing institutions and the use of such indicators for school improvement purposes in regards to 'educational (school) effectiveness'.
Research interests also include the methodology of multilevel modelling, supervised the production of software package (MLwiN) and is the author of 'Multilevel Statistical Models': (2011) 4th Edition, Wiley, Chichester.
Directions to UQ
Event Tools
Share This Event
Print
Email
Share
Rate This Event
Tweet This Event
Calendar Tools
Featured Calendars
Subscribe via RSS