Standard % recovery of HS-GC

[Mac2]

I'm wondering if there is scientific or statistical justification for setting acceptance criteria for % recovery.
I am developing a HS-GC method using a standard prep'd in duplicate. I run standard 2 three times bracketed by standard 1's.
I am quite often getting results outside of 95 - 105% which is our current spec. I want to increase the spec to 90 - 110% but need some solid justification other than just my results.
Thanks.

[chromatographer1]

There is no more solid evidence than your results and your results are scientific.

Assuming you are not causing the variation by something you are doing your results indicate that as you are presently executing your procedure your range of results should be wider than 95-105%

Quoting some book reference will change your results in the future? Your specifications are based on your evidence and experimental results. No other justification is applicable.

best wishes,

Rod

[Peter Apps]

I presume by "recovery" that you mean "accuracy" i.e. how close your result is to the known composition of the test sample.

How accurate do your results with real samples need to be ?. If they have to be within 5%, then setting a recovery criterion of 10 - 110% is not going to get you where you need to be.

Your being unable to get within 5% suggests that there is considerable room for improvement in the method of the way that you are making up your test samples.

Peter