Evaluation of Measurement Systems with a Small Number of Observers

The standard R&R study for evaluation of measurement systems assumes that participating observers constitute a random sample. Often there are only a few observers, all of them included in the study. An alternative measure for the gauge R&R is proposed for this situation, and it is shown that this may improve the perception of the quality of the measurement system markedly, especially with only a few observers. Finally it is shown that a simple estimator can be used, with a bias limited to just a few percent.