. We needed to log transform in order to justify using a t-test on these data - otherwise we would have violated the test's assumptions. However, log-transforming changed the question we were asking about the data. Scientifically, we want to know if there is a difference in survival times between small and large cell cancers. a) What did the test we ran actually ask about our data? b) Given your answer to a), were we still addressing the scientific question we set out to answer when we ran our t-test on log-transformed data?