To estimate repeatability, each operator measures each part at least twice. To estimate reproducibility, at least two operators must measure the parts. It is important for operators to measure the parts in random order, and the selected parts should represent the possible range of measurements.