Mean or geometric mean

Donnenberg, Albert donnenbergad at MSX.UPMC.EDU
Wed Mar 25 13:23:23 EST 1998

I have a question on basic statistics, and I hope its not too stupid: 
I´m wondering what to use: "mean" or "geometric mean" as a relative 
fluorescence measure; and how about the "median"??  


The idea behind using any of these descriptive statistics is to find one
provides an accurate summary of your data.

If you are using a log amp (or log transformed data acquired with a
linear amp, as in the Coulter XL) then the geometric mean (i.e. log
mean) is appropriate, providing that your population is relatively
symmetrically distributed about the geometric mean.  If your population
is very skewed, an average fluorescence (arithmetic or geometric)
provides a misleading description and should be avoided.  You will know
that this is the case if the mean is very different from the median.
The median conveys less quantitative information (only that half of the
events are dimmer and half are brighter than the median), but for that
very reason, is less likely to be misleading.  When the fluorescence
pattern is bi- or multi-modal or has a tail or a shoulder of biological
interest, one has to resort to other methods to adequately describe the
data (e.g. the algorithms used for cell cycle analysis).

Hope this helps,

Albert Donnenberg
University of Pittsburgh

More information about the Cytometry mailing list