If the pooled variance of certain type of instruments is already known, then the particular instrument received for calibration may be rejected if its sample variance is more than the prescribed limits. For the purpose one may think of Fisher’s test again; say for example the ratio of the sample variance to the pooled variance of the group of instruments may not exceed the tabulated value for the known degrees of freedom at 5% level of significance. Attention is drawn to the ratio of the mean variances of proving rings D and E, which are 5:1. Had there been already some fixed criterion existed it should have been possible to reject the proving ring on the basis of its variance. For example, according to criterion proposed above, the proving ring E could be rejected.
Source : measurement uncertainty
