[Mr_Quality]

Escuse me if I'm getting carried away, but how's this:

Gage Reproducibility & Repeatability Study
To determine the sources of variability (based on ANOVA)
To be performed at a single level of interest, say 400 ppm

Factors:

Replicate (take 3 measurements per time point per operator)

Day (perform measurements over different days, you can decide if you want to space the days apart over say weeks and you can decide if you want to recalibrate each time)

Operator (have 2 or 3 different people take the measurements to see if there is a technique component)

This could be done for the probe and again for a reagent based kit for comparison.



Regression Study:
To test for linearity, proportional and constant bias relative to another method.
To be performed with several samples (I’d recommend at least 10) that span the entire linear range of the device or reagent test.

Assay the samples by the probe and assay the same samples using either a reagent based test or a reference method like ICP/AA.