STATS 768 describes longitudinal data analysis. This is important for the modelling of repeated measurements, and part of the course will focus on their use in the analysis of epidemiological data. The course will begin with data exploration techniques and within-person summary statistics reflecting changes over time as well as data visualization.
Progressing to the exploration and applied regression modelling of longitudinal and clustered data, with a focus on the health sciences and use of clinical and cohort data. Including generalized estimating equation (GEE), generalized linear mixed models (GLMMs) plus sandwich estimators, survival analyses, estimating variation in slopes of mixed models, marginal structural models for causal inference. Emphasis will be placed on algorithms, likelihoods, drop-outs with the meaningfulness of missing data.
Students will be taught to program in both R and SAS and to be able to interpret the resulting output. The skills developed in this course are particularly useful for those wishing to have a career involving the analysis of epidemiological and clinical trial data collected in the health sciences, census information collected by the NZ government, data collected for biological experiments or become a biostatistician or global health analyst.