I have a set of 70 individuals and 90 continuous variables measured for those individuals, and I run univariate models to test for the association to an outcome variable. Is it necessary to always perform multiple testing correction even in cases like these?
Please limit answers to the question in place.