Do you mean for example, reduction in number of viable cells after some
E.g., number of bacteria per ml in two or more treatment groups?
We usually do a transformation of all plate counts into log form and
the geometric mean and deviation.
It is not sensible to use the arithmetic data for computing the means.
I learned this the hard way, by making a public fool of myself.
(See letters to the editor ASM News circa 1980 or so-- I don't recall
citation.) I wrote what in retrospect was an idiotic letter on this
issue. Others promptly
responded to my published letter politely (more politely than I
deserved) educating me to the fact that plate count data is not normally
distributed and therefore it is not appropriate
to use parametric statistics on such data without first doing the log
Hope this helps.
Oh, if the stat software indicates that data are not normally
distributed, even after the
log transformation, then I use a non parametric test on the transformed
"Rottjakob, Dave" wrote:
> What is the best way to calculate log reduction?
>> Dave Rottjakob
>>dave.rottjakob at apptecls.com>> ---