Experiments lie at the heart of the Scientific Method, and good statistical design is critical to the success of any experiment or experimental programme, be it for furthering scientific knowledge, market research or process optimisation. In order to be able to design a good experiment and get the most out of the resulting data, a basic knowledge of statistics is required. While statistical packages do help, and are often essential for carrying out the calculations, without the underlying statistical knowledge the user may easily be misled and produce poor designs or perform inappropriate analyses.
After an introduction to the necessary concepts and techniques of probability and statistics, and their application to real-life problems, participants will be taken through the principles and practice of experimental design and analysis. Starting with single- and two-sample experiments, the course will move on to multifactor and factorial designs, including fractional factorials, latin-square and incomplete block designs. Screening designs, requiring the minimum number of observations, to response surface designs, which enable the optimal response to be explored, will be covered. Experiments in which the responses are not Normally distributed will be analysed and participants will be introduced to the basic ideas of statistical process control. Concepts will be illustrated throughout using practical examples from science, engineering and management. The analysis of experimental data, from simple graphical analyses to analysis of variance and regression modelling will be dealt with, including the important topics of error analysis and transformations. Candidates will be taken through most of the analyses on spreadsheets, but specialised experimental design software will be used for the more complex designs.
Scientists, engineers and metallurgists who are involved in experimental work or in process optimisation, or who wish to gain an understanding of statistical design and analysis of experiments. The course will also be useful for those from other areas who wish to know more about the field. No previous knowledge of statistics will be assumed, and although some mathematical notation will be used, no mathematical knowledge beyond first-year university level will be assumed. The course material is at NQF level 7.
The course Outline is detailed in attached brochure.