Hi Raphael, On Monday 12 January 2009 10:04:03 Raphael Wimmer wrote: > When entering the numbers "3 7 7 19" in fields A1-A4 of a new spreadsheet, > selecting them and selecting "Column statistics" from the context menu, a > new spreadsheet pops up. > It contains (among others) two fields for Standard Deviation and Variance. > LabPlot calculates the variance as 48 and the StdDev as 6.9... which is > incorrect. The correct variance is 36, the correct StdDev is 6. > This error can be reproduced with other combinations of numbers. > > It seems that LabPlot always divides the sum of squares by n-1 instead of > n.
LabPlot's behaviour is consistent with other programs like OOCalc, Mathematica and QtiPlot. STDEV() and VAR() in OOCalc and StandardDeviation[] and Variance[] in Mathematica both reproduce what LabPlot calculated. As explained on [1], all these programs compute the "bias-corrected sample variance" and it's square root. On that page it is also mentioned that using "variance" for "bias-corrected variance" is consistent with widespread inconsistent and ambiguous terminology. Too bad, but that's the way it is. I think it would be more confusing if LabPlot would deviate from the terminology of all other programs. Cheers Frank [1] http://mathworld.wolfram.com/StandardDeviation.html
signature.asc
Description: This is a digitally signed message part.
