By Anonymous on Wednesday, May 2, 2001 - 06:09 pm:
Can someone tell me a quick way to calculate %RSD from two standards using their peak areas and standard weights in Excel?
By Anonymous on Thursday, May 3, 2001 - 05:14 am:
Std. Dev / mean * 100 =%RSD
By Anonymous on Thursday, May 3, 2001 - 05:16 am:
The std dev and/or %RSD means nothing when calculated from only 2 standards.
By M Emerick on Thursday, May 3, 2001 - 05:23 am:
Divide the weight of std by it's area count and put these two numbers in two seperat cells. Then use the insert function key for another cell and place in the average of the two. In another cell go back to the insert function and insert stdev. Then in one more cell divide the stdev by the avg cell and multiply by 100%.
By DR on Friday, May 4, 2001 - 08:59 am:
NOOOOOO!!!!! Now you've weighted it; biased it lower by tossing an average in as a 3rd #!
Get 3 standard responses, then calculate (100*STDEV/AVERAGE) or just STDEV/AVERAGE, formatted as a %. I don't know how or why, but Excel will return a value for STDEV where n=2. It really should not.
By tom jupille on Friday, May 4, 2001 - 10:51 am:
Uhh - I'll argue with that.
Fire up Excel and get STDEV for the two values 2 and 3; the returned value is 0.707. Now get STDEV for the three values 2, 2.5, and 3; the returned value is 0.5.
In fact, the formula for sample standard deviation lets you compute sigma from two values. FWIW, my old Naval Ordinance Statistics Manual (ca 1960) has a table of factors for estimating sigma from the range of a small number of experiments (down to 2).
Having said that, I concur with anonymous#2 that the result is not very useful. If I wanted to be cynical, I'd say that the major utility is cosmetic (it sounds more professional to say "the relative standard deviation was 7%" rather than "We ran it twice and got 19 the first time and 21 the second").
-- Tom Jupille / LC Resources
By Anonymous on Wednesday, May 9, 2001 - 07:16 am:
IS THIS PROCEDURE IS ACCEPTABLE TO FDA AND REGULAROTARY BODY
By Anonymous on Wednesday, May 9, 2001 - 07:50 am:
IF you validate that the spread sheet does the calculation properly, then I don't know that they would have a problem with it.
FYI, I do these using excell as well as turbochrome, as well as a hand-held calculator, and get the same answers.
good luck
By M Emerick on Wednesday, May 9, 2001 - 09:29 am:
We have submitted validations with the same calculations and have had no problems.
Posting is currently disabled in this topic. Contact your discussion moderator for more information.