the two means are considered to be equivalent. In the multiplicative model, the null hypothesis is that the ratio of the treatment mean to the reference mean is not near one. That is, the ratio is ...