Usually, you use mean (SD) for normally distributed data and median (IQR) for non-normal. However, for certain variables, I think it could seem odd to use median, e.g. for certain continuous variables like blood pressure (mmHg), drug dose (mg), and oxygen supply (l/min) despite a non-normal distribution (also AFTER log transformation). But is it ok? could someone provide me with references?

More Michael Dan Arvig's questions See All
Similar questions and discussions