Module leader: D Woods
Please see the full Module Specifications document for background information relating to all of the APTS modules, including how to interpret the information below.
Aims: To introduce the fundamental principles of statistically designed experiments, and other modes of data collection, and highlight their important role in the scientific method. A variety of classical and modern methods will be overviewed, connections between them emphasised. and ongoing research challenges introduced.
Learning outcomes: An understanding of the major different mechanisms for data collection, their similarities and differences. For designed experiments, an appreciation of the impact of the choice of design on the precision and accuracy of the subsequent statistical modelling and inference. An awareness of the challenges presented to data collection methodologies from modern scientific experiments and studies. Familiarity with some of the practical issues in implementing statistically designed experiments.
Prerequisites: Linear and nonlinear/generalised linear modelling. Sampling distributions of parameter estimators, including basic asymptotic results. An understanding of the fundamentals of Bayesian inference. Basic statistical computing, including simple optimisation methods. All these topics, and more, are covered by the APTS Statistical Inference, Statistical Computing and Statistical Modelling modules.
- Modes of data collection, experiments and causality, the impact of design on modelling
- Factorial experiments
- Bayesian design, and design for nonlinear models
- Design of computer and simulation experiments
- Data collection in spatial studies and via sample surveys, and connections to design of experiments
Assessment: Exercises, that will include finding and assessing designs for practically relevant examples.