Design and Analysis of Experiments

16, 18, 20, 23 and 25 September 2024

researcher doing an experiment

This course covers the principles and practice of designing experiments or trials, and the analysis of data from them. The course covers the following topics:

choice of experimental units;
importance of randomisation, and the practicalities;
replication and sample size;
blocking and matching;
commonly used designs, including completely randomised designs, randomised block and matched pair designs, Latin square designs;
treatments, including factorial structures;
analysis of data from designed experiments;
analysis of variance and covariance;
special designs, including incomplete block designs, split-plot designs, and fractional factorial designs;
transformations of data;
practical and ethical issues arising in the conduct of experiments.

Course presenter:

The presenter is Associate Professor Graham Hepworth, Consultant for the Statistical Consulting Centre in the School of Mathematics & Statistics. Graham has had extensive experience over three decades in design and analysis of experiments, in fields such as forestry, horticultural science, animal studies, medicine (randomised controlled trials), industry and the social sciences.

Course structure:

This online course runs over five mornings (Mon/Wed/Fri/Mon/Wed). Each day will commence at 9:00 a.m. and finish at 12.30 p.m. The sessions will mix lecture presentations with practical exercises. The statistical package Minitab will be used in the course. However, the course will not be package-centred, and no prior experience with Minitab is necessary.

Who should take this course?

The course is suitable for anyone involved in the design and analysis of research experiments. Applications include randomised trials in medicine or the social sciences, designed experiments in the biological sciences, studies of processes in engineering, as well as many possibilities in other disciplines.

Prerequisites:

Participants should have some understanding of statistics at an introductory level. For example, it would help to know something about hypothesis tests and confidence intervals, and, preferably, analysis of variance. The course "Statistics for Research Workers" conducted by the Statistical Consulting Centre would be ideal preparation.