Design and Analysis of Experiments
A course for research scientists involved in designing their own experiments
8 - 10 November 2004
· Have you ever been challenged about the validity of the design and analysis aspects of your study?
· Have you ever conducted experiments where the resulting data led unexpectedly to non-significant results?
· Are you concerned that your experiments may not be based on sound statistical principles?
· Have you ever worried whether your experimental designs are truly cost-effective?
· Do you feel you could benefit from efficient approaches to experimental data analysis methods?
If the answer to one or more of these questions is YES, and you have the necessary pre-requisites (see below) then this course is for you.
Course aims and approach
Our aim in this course is to ensure that you gain intuition and confidence in the statistical aspects of sound experimental design. The underlying concepts and principles will be discussed, together with associated data analysis methods. You will learn to use resources efficiently at the planning stage of your experiment, by giving due attention to the expected outcomes and ensuring that the results have a high degree of precision.
The course will include presentations, discussions and computer-based practical sessions. A choice of software (Genstat, MINITAB, SAS and SPSS) will be available for practical work. You will have the opportunity to analyse data from a variety of different types of experiments. Much emphasis will be placed on interpretation of computer output to help answer questions posed by research objectives.
Statistical principles of good experimental design
Playing "TOMATO" - an experimental simulation game to illustrate design concepts, unbalanced data structures and factorial treatment structures.
Understanding hierarchical data structures
Designing experiments in small blocks
Dealing with data structures that involve both classification factors and variates.
Participants should have knowledge of basic statistics (estimation, confidence intervals and hypothesis testing) and be familiar with, or have a working knowledge of, the analysis of variance.
How to apply
Please send an e-mail to [log in to unmask] before 18th October with your full name, job title, direct telephone number, the name and address of the organisation you work for, an address for invoicing and a purchase order number if necessary. Please also let us know of statistics and/or spreadsheet packages you currently use. The full fee for attendance is £690.
course webpage http://www.ssc.rdg.ac.uk/courses/sctt04a.html
Statistician, Statistical Services Centre
University of Reading