GENERALISED LINEAR MODELS (2462)ORGANISER: Dr James Carpenter TIMETABLE SLOT: 9 January 2012 to 8 February 2012 (9:00am Monday to 12:30pm Wednesday) AIM To equip students with the necessary skills to understand the principles and analyse data using Generalised Linear Models. OBJECTIVES By the end of this module students should be able to:
- demonstrate an understanding of the theoretical basis of Generalised Linear Models;
- use Generalised Linear Models for analysis of discrete data;
- present results in a form suitable for publication in a medical journal.
CONSTITUENCY This module is intended for people with both mathematical (up to first year undergraduate level) and statistical background (undergraduate degree level in joint mathematics/statistics for example) intending to pursue a career in medical statistics. A knowledge of linear regression, analysis of variance, likelihood theory and simple methods of analysing quantitative and categorical data is essential.
CONCEPTUAL OUTLINE Topics to be covered are:
Formalisation of Generalised Linear Models, log likelihood and deviance, comparison of nested regression models, logistic regression for binary data, application of logistic regression to prospective and case-control studies, conditional logistic regression for matched case-control studies, Poisson regression for rates, log-linear models for contingency tables, the fitting of Generalised Linear Models, Pearson, normalised and deviance residuals, overdispersion. Specific topics will include comparison of proportions, allowance for matching, cross-over trials, missing data, quantifying repeatability and agreement and measurement error. A particular focus will be on interpretation of regression parameters and strategies of analysis. TEACHING STRATEGY Learning will generally be based on relevant practicals following lectures. Some sections of each module will involve the use of computers. Assignments will also be given as part of the practical work. Approximately half of the contact time will be spent in the form of practicals. Time will be allocated for private study sessions.
LEARNING TIME The module is made up of 150 Notional Learning Hours – 50 hours contact time, 60 hours directed self-study, 10 hours self-directed learning, and 30 hours assessment, review and revision.
ASSESSMENT Students will carry out one assessment, each consisting of an analysis of data together with submission of a short report.
|