Hi everyone,

I have a very basic knowledge of statistics so apologies if the question is not formulated in appropriate terms. I really hope to get some help from you on this.

I have data comprising a response variable and three predictor variables (a,b,c). I would like to assess the precision of the testing method in a statistical way.
I am thinking of extracting from the data specimens with a constant var a and var b (c not constant) and based on this extracted data I could assess if coeff of variation for the response variable is similar to coeff of variation for var c. If the CoVs were similar I could say my test method gives precise results.

Are there any other ways of approaching it in a more statistical way? Could I somehow use ANOVA to assess it instead?

Any help much appreciated!

Thank you,
Hanna