n1 and n2
values were randomly sampled with replacement
from the data set. To model a ‘‘difference in location’’
parameter, a specific effect size (ie, fraction or multiple
of the data set’s standard deviation) was added to n2
for each iteration of an experiment. The added effect
sizes to model a treatment in shift in location parameters
were d = .2s, .5s, and .8s, which Cohen defined as
small, medium, and large, respectively,