Comparing a single case to a control sample: differences in distribution versus differences in means