sample kurtosis and sample skewness

i have noticed that the cfa book has a bit of a different formula for sample skewness and sample kurtosis than the schweser notes. When it comes to the 1/n, and book has n/(n-1)(n-2). please help, which one should i use or do i just need to understand the values, because the LOS doesn’t say compute.


You should never be asked to compute kurtosis or skewness for the exam. I’m not sure what formula you’re referring to because they differ between distributions, but you can be more than confident that it won’t be tested quantitatively.

phew, thank you…

Trust the CFA text when in doubt regarding which formula to use. Typically (n-1) degrees of freedom may be used to sample statistics whereas n degrees of freedom would be used to caculate population stastics. And check the learning outcome statements (LOS) listed in the CFA text. If it says the word calculate or compute when referring to the statistics such as kurtosis and skew then you need to calculate them. However, I just looked at the LOS for LI and I’m pretty sure you don’t need to calculate those.

Yeah you wont need to compute them anyway but The 2 formulas are correct but one for small samples n/(n-1)(n-2) and one for large samples (1/n)