This article reviewed the non-traditional estimators of the measurand, received from empirical distributions. First of all it is important for distributions with associated with the theoretical models differed from the normal distribution. One-component and multi-component estimators for family of trapezoidal distributions are considered. As the best estimator of the distribution center for the whole family can be used two-component estimator, which determined as a linear function of sample mean and sample midrange. Proposed new estimator expand the list of measurand estimators and ways of uncertainty estimation of the type A.
measurement result, uncertainty estimation of type A