The logarithmic (log) transformation is a simple yet controversial ste
p in the analysis of positive continuous data measured on an interval
scale. Situations where a log transformation is indicated will be revi
ewed. This paper contends that the log transformation should not be cl
assed with other transformations as it has particular advantages. Prob
lems with using the data themselves to decide whether or not to transf
orm will be discussed. It is recommended that log transformed analyses
should frequently be preferred to untransformed analyses and that car
eful consideration should be given to use of a log transformation at t
he protocol design stage.