Statistical Quality Control Procedures

Taguchi's emphasis on instilling quality at the design stage has been important in a number of situations and deserves careful attention. His specific robust design approach has yielded good results in many instances, but recent studies indicate that more classical designs can serve the same purposes-sometimes, at least, doing so more efficiently, albeit less understandably for the technically unsophisticated. While some details of Taguchi's method as practiced in manufacturing may not be applicable in white-collar, public-sector, service-oriented activities, some of the Taguchi philosophy might be and might, for example, help to make the design of existing or future data collection, data analysis, and decision-making processes robust with respect to such factors as measurement biases (due, perhaps, to poor calibration or laboratory-tolaboratory variation), highly variable measurements, and missing data. Irrespective of incorporating a formal Taguchi approach, significant benefits can often accrue from improved calibration and sample size calculation, and from the use of better statistical techniques.