SPECIFICATION OF THE ASSUMPTIONS FOR THE TYPE-A EVALUATION OF STANDARD UNCERTAINTY

摘要

The formulation of Type A uncertainty in the GUM does not include the necessary assumptions. The paper reviews the common set of four assumptions and, in particular, the replacing the identity of distributions (i. d.) by a more general assumption of equivalence of observations. This assumption shows several advantages including its more natural relation to the measurement conditions. Possibilities of validating the assumptions, by an analysis of experiment and using the statistical tests, are briefly discussed.