When conducting a measurement system assessment study, practitioners want as precise an estimate of the true repeatability and reproducibility (R&R) as possible so they can correctly decide whether the measurement system is acceptable. They are, however, faced with cost and time constraints that restrict the number of parts and repeated measurements that can be used in the study. By incorporating freely available production measurements (baseline data), they can reduce the number of parts in the study to two or three and still obtain a better estimate of the R&R than they would have otherwise. To avoid bias, they must ensure the baseline data reflect the current manufacturing and measurement processes.
Stevens, N.T., Steiner, S.H., & MacKay, R.J. (2015). Being smart about parts. Quality Progress, Volume 48, Issue 3, 1 March 2015, Pages 32-37.