Calculation od overall %RSD for intermediate precision test

[alminc]

What is the most correct way to calculate the overall %RSD for data from 2 precision tests conducted by 2 different analysts in this case:

analyst 1, HPLC 1: prepares 6 sample solutions from product batch A, and gets 6 assay results
analyst 2, HPLC 2: prepare his own (new) 6 sample solutions from same product batch A, and gets 6 assay results

From statistical point of view, is this the case of paired data (because of same batch A) or is it unpaired data (two statistically independent sets of data because analyst 2 prepares new sample solutions)?

What calculation is most apropriate to determine the overal %RSD?

Thanks..

[aceto_81]

Hi,

your samples are unpaired.
If you use for example 6 different batches, and then perform analysis by A & B, then you have a paired data set, as sample 1 for analyst A is the same a sample 1 for analyst B, but not the same as sample 2.

For your question about %RSD: you can use your whole results (analyst A & B) and calculate on these 12 results.

Via ANOVA you can also analyze your inter- and intraday precision %RSD.

HTH

Ace