In a two sample clinical trial, one of the outcome measures was highly skewed. To test the difference between the levels of this measure in the two groups of patients, possible approaches include: (a)...

In a two sample clinical trial, one of the outcome measures was highly skewed. To test the difference between the levels of this measure in the two groups of patients, possible approaches include:

(a) a standard t test using the observations;

(b) a Normal approximation if the sample is large;

(c) transforming the data to a Normal distribution and using a t test;

(d) a sign test;

(e) the standard error of the difference between two proportions.