It is often easier for investigators to specify treatment effects in percentage change in the mean and variability in coefficients of variation. We show that with this specification questions about ...