Linear Models

Course outline:
This course gives an introduction to statistical modelling and the theory of linear statistical models. The student is introduced to the principles of experimental design, statistical software and practical data analysis through weekly computer practicals and the exposure to many data sets. The course has three sections:
Regrerssion: The multivariate normal distribution; quadratic forms; the linear model; maximum likelihood; estimates of parameters in the linear model; the Gauss-Markov theorem; variable selection procedures; analysis of residuals.
Design and analysis of experiments: Introduction to the basic design principles, basic experimental designs (completely randomised design, the randomised block design, latin square design,) factorial experiments, analysis of variance, the problem of multiple comparisons, power and sample size calculations, introduction to random effects and repeated measures.
Nonparametric statistics: Introduction to nonparametric tests and methods, including Mann-Whitney U, Kruskal Wallis, Friedman and randomisation tests.