Institute for Improvement of Quality and Productivity, University of Waterloo, Ontario, Canada
G1.15, G Block, University of Waikato
Reducing the variation of process outputs is a key activity in improving the quality and reducing the cost of high volume manufactured items. A fundamental strategy used in these efforts is to identify the inputs to the manufacturing process whose variation is transmitted through the process to the outputs. If this transmitted variation is a major component of the overall output variation, then the source must be better controlled or its effect mitigated in order to make progress.
In this talk, I will highlight some statistical issues that arise in an ongoing attempt to reduce variation in wheel alignment characteristics of trucks. The issues are driven by the choice of metric to measure variation in the process. The usual choice, based on the standard deviation, leads to some challenging communication problems. A second set of issues is generated by the measurement system, its contribution to the problem and its control.