Inaugural Lecture - Better by design: statistical planning of experiments

Professor Steven Gilmour's Inaugural Lecture
Better by design: statistical planning of experiments
Thursday 2 February 2017, 18.30 (doors open at 18.15)King's College London, Anatomy Lecture Theatre (K6.29), King's Building, Strand Campus
Abstract
Most statistical methodologies aim to extract the maximum information from data to answer importantquestions. The statistical design of experiments on the other hand, aims to ensure that data collected fromexperiments contain the maximum amount of information to answer the relevant questions. As experimentsbecome more complex and the methods of modelling data more elaborate, designing experiments continuesto throw up new challenges. After a brief review of the history of experimental design, some recentdevelopments will be described, showing how the general principles still apply, but require new methodologyto design experiments in important areas of application.
The following three areas will be highlighted:1. Experiments with many factors of interest, but uncertainty about the appropriate mathematical model. These have recently been used in protein screening.2. Experiments with some factors whose levels are more difficult to set than others, leading to multi-stratum designs and an application in cleaning surgical instruments.3. Designs carried out on social networks, where the usual assumption of treatment-unit additivity is no longer reasonable.
Biography
Steven Gilmour has a BSc in Statistics from Heriot-Watt University (1987) and a PhD in Applied Statistics from the University of Reading (1991). He was Lecturer in Applied Statistics at Reading (1990-99), Reader, then Professor in Statistics at Queen Mary University of London (2000-10) and Professor of Statistics at the University of Southampton (2010-16), as well as having held visiting positions at the University of Wisconsin-Madison and the University of Antwerp, before joining King’s College London in 2016. His research covers the theory, methodology and applications of the statistical design and analysis of experiments.

 The Faculty of Natural & Mathematical Sciences at King’s College London has a Code of Conduct, which we expect participants at our events to abide by. This is intended to ensure an inclusive and productive environment and can be read here