Application of linear models and generalised linear models to practical problems and implementation in R. Formulating statistical problems. Choosing an appropriate method of analysis, model-building, verification of assumptions. Working with real-life data. Presentation of results in an appropriate format.Application of the above through tackling a variety of real-life statistical problems.
Design of Experiments: principles of experimental design; planning of experiments; comparative experiments; and selected topics from common designs: completely randomised, randomised blocks, factorial experiments; fixed and random effects; associated analyses - analysis of variance. Observational studies v. experiments: problems of bias, confounding, difficulty of causal interpretation; planning observational studies; analysis: matching, adjusting for confounding variables; cohort studies; case-control studies.
Sampling target and sampled populations, finite populations, simple random sampling and selected topics from stratification and cluster sampling, ratio and regression estimators, randomised response methods.
Skills needed for writing for scientific reports, papers and dissertation. Having identified the area of work and been assigned a project advisor, the student will undertake a conceptual review, a literature search, and a critical analysis of previous work. In consultation with the project advisor the student will write a research review and proposal for a subsequent research project.
Simple and multiple regression: estimation of model parameters, tests, confidence and prediction intervals, residual and diagnostic plots. Practical forecasting. Time plot. Trend-and-seasonal models. Exponential smoothing. Holt's linear trend model and Holt-Winters seasonal forecasting. Autoregressive models. Box-Jenkins ARIMA forecasting.