Analysis strategies for multiple responses in quality improvement experiments

Quality improvement experiments often aim to find operating condition that achieves the target value for the mean of a process characteristic, and simultaneously minimises the process variability. For this purpose, Taguchi's techniques of analysis based on signal-to-noise ratios and dual response surface methodology are commonly used to achieve this goal. This paper shows how the generalised linear models approach of modelling the 'mean function' and 'variance function' jointly can be used to achieve the goal. Two examples illustrate differences among three approaches.