3 ms·
>If you have a skewed distribution you might not be able to use the central limit theorem at all. The CLT only requires finite variance. Skew can be infinite a
by _dain_ 2y ago
>If you have a skewed distribution you might not be able to use the central limit theorem at all.
The CLT only requires finite variance. Skew can be infinite and you still get convergence to normality ... eventually. Finite skew gives you 1/sqrt(N) convergence.