Interval estimations in metrology
arXiv:1402.0248 · doi:10.1088/0026-1394/51/3/191
Abstract
This paper investigates interval estimation for a measurand that is known to be positive. Both the Neyman and Bayesian procedures are considered and the difference between the two, not always perceived, is discussed in detail. A solution is proposed to a paradox originated by the frequentist assessment of the long-run success rate of Bayesian intervals.
11 pages, 5 figures, submitted to Metrologia