Bioequivalence and Bioavailability Forum

Dear Reddy!

❝ Is there any guidelines for the acceptance criteria of calibration curve correlation coefficient (r, r2)?

I don’t think so. Both values depend on the number of data points (in linear regression r 0.9877 (n=3) carries the same information as r 0.729 (n=6)).
Rules like r>0.995 or any other number are just stupid.
See this old thread at David’s list (although it deals with the estimation of the elimination phase).
See http://www.fda.gov/cder/guidance/4252fnl.pdf FDA’s Guideline about the acceptance of calibration and QCs.

❝ what is the difference between r and r2?

r = √r²

❝ Which value (r or r2) we have to consider when the weighing factor is 1/x2?

Doesn’t matter.

---

Edit: Link corrected for FDA’s new site. [Helmut]