No measurement is ever exact; there is always an uncertainty associated with it. In the past, international specification standards for the performance of acoustical instruments have not included uncertainties of measurement of the testing laboratory when specifying tolerance limits. However, to demonstrate conclusively whether or not a particular instrument conforms to the specifications of a standard, it is essential that these measurement uncertainties be taken into account. In a recent change of policy, the international committee responsible, IEC TC-29 ‘‘Electroacoustics,’’ has decided that all its future standards must include uncertainties of measurement in the tolerance limits specified. This paper will discuss different types of measurement uncertainty and likely contributions to uncertainty budgets. It also explains the difficulties in accumulating the necessary data for inclusion in IEC 61672, and of incorporating the data into the tolerance limits. The reasoning behind the decision to quote a maximum permitted expanded uncertainty of measurement for the testing laboratory separately for each test will also be discussed. Examples will be presented of the determination of expanded uncertainties of measurement for acoustical and electrical signals and their application in assessments of conformance.