The role of data transformation in taguchi analysis

This paper reports on the usefulness of data transformation in a proper application of the ‘Taguchi method’ for off‐line quality control. With the joint aim of establishing a meaningful measure of performance variability and reducing the risk of incorrectly assuming no interactions in the performance model, a simple method of choosing appropriate data transformations is presented. This provides the basis for a suggested general procedure for carrying out Taguchi analyses. The technique is assessed by considering how it performs on data taken from previously published applications of the Taguchi method.