(a) Discuss the use of the moving range method to estimate the process standard deviation when the data are positively autocorrleated.
(b) Discuss the use of the sample variance s2 with positively autocorrelated data. Specifically, if the observations at lag i have autocorrelation i, is s2 still an unbiased estimator for 2?
(c) Does your answer in part (b) imply that s2 would really be a good way (in practice) to estimate 2 in constructing a control chart for autocorrelated data?