Abstract
Planning needs to be done before a process capability study is conducted. Some of the important issues in planning are the concept of what constitutes a rational sample and how the data should be collected to assure that the sample selected from the process is representative of the population as a whole. The primary goal of this research is to propose guidelines on how one can design some process capability studies. One specific problem discussed is how to allocate a fixed number of observations to be taken among different levels of a two stage random effects nested design. Four different criteria are studied for designing data collection schemes to be used while conducting a process capability study. Formulations are developed for these four criteria namely, minimizing the variance of the sum of the estimates of the variance components, minimizing the variability of the estimate of the total variance, minimizing the average of the variances of the estimates of the variance components and maximizing the power of the F test. The solution methodology is discussed and the results are presented. Some general guidelines, which will help an experimenter in planning process capability studies are developed using these results.
Mantri, Nitin Ramesh (1994). Designing data collection schemes for process capability analysis. Master's thesis, Texas A&M University. Available electronically from
https : / /hdl .handle .net /1969 .1 /ETD -TAMU -1994 -THESIS -M2934.