Estimation in multiple measurements

A recent paper by Nam et al. (Accred Qual Assur 14:43–47, 2009) considers the issue of repeated measurements of a same measurand. The paper is a followup of two previous papers (Choi et al. in Accred Qual Assur 8:13–15, 2003, 8:205–207, 2003) on the same topic. The first two papers show a well-known fact, i.e., that systematic effects are not reduced by repeating measurements, and give formulae for the uncertainty of the estimate of the average from repeated measurements in the case that the differences between individual results are not (Choi et al. in Accred Qual Assur 8:13–15, 2003) or are (Choi et al. in Accred Qual Assur 8:205–207, 2003) statistically significant, respectively. The last paper addresses a more interesting issue, that is, a difference in the estimate average for non-linear models depending on how raw data is processed in the model. In the present paper, I re-discuss the whole issue of multiple measurements, trying to clarify the various aspects of the topic.