UNCERTAINTY LIMITS FOR CERTIFIED VALUES OF STANDARD REFERENCE MATERIALS

Uncertainty limits usually employ the 95% confidence limits and should include allowances for the uncertainty of known sources of systematic errors of the measurements. But in many cases, when an interlaboratory comparison study is used for certification purposes of standard reference materials, only the 95% confidence limits are considered and given as the uncertainty limit. In this work, we shall show the results as the number of mean values (n) reported vary and only the 95% confidence limits are considered, using as an example the intercomparison study on IAEA/SL-3/lake sediment sample. And finally, we shall compare the results if only the 95% confidence limits are considered with those of the standard deviation and propose when n is large that the standard deviation should be employed instead of the 95% confidence limits as the uncertainty limits, if the other systematic and random errors are not included. © 1990 Springer-Verlag.