Linearity of impurities!

[koko]

Hello experts!

I have a doubt, and I hope that you can help me! I am validating a method for impurities, in the accuracy experiment I get a r2= 0.95, the 100% of the concentration of the sample is 0.1 ug/ml. Can I considered this value of r2 as a linear or I have to change the method?

[tom jupille]

No single/simple answer to that one. r2 does not *prove* linearity, it merely gives an indication of the degree of scatter. Comparing the r2 values from a linear fit to those from a quadratic fit will give you a bit more insight (but not much). There are more meaningful statistical tests that can be applied. Download and read the articles in this series for more detailed information:

http://www.lcresources.com/sandbox/amla ... stics.html

[koko]

Tom!
Thank you, the link was very usefull. Now I have more tools for a successfull validation.