The Estimation of Slope When the Errors are Autocorrelated
SUMMARY In this paper an investigation has been made of the problem of constructing systematic and randomized designs for the estimation of slope when the errors are autocorrelated. Some results are obtained for general stationary processes and specific designs given for moving average and autoregressive processes. Comparison is made with randomized block designs. The theory is illustrated on a time-series obtained from a uniformity trial on a batch production process.