If x
contains missings and these are not removed, the skewness
is NA
. Otherwise, write $x_i$ for the non-missing elements of x
,
$n$ for their number, $\mu$ for their mean, $s$ for
their standard deviation, and
$m_r = \sum_i (x_i - \mu)^r / n$
for the sample moments of order $r$.
Joanes and Gill (1998) discuss three methods for estimating kurtosis:
[object Object],[object Object],[object Object]
Only $G_2$ (corresponding to type = 2
) is unbiased under
normality.