Curtosis normal distribution

According to what I read from here , the excess of the normal distribution should be around 3. However, when I use the matlab kurtosis function, I could not verify this:

data1 = randn(1,20000);
v1 = kurtosis(data1)

The kurtos of the normal distribution seems to be around 0. I was wondering what was wrong with him. Thank!

EDIT I am using MATLAB 2012b.

+5
source share
1 answer

If this were so, this would be convincing evidence that he calculated the excess of excesses which is defined as an extreme minus three.

However, my MATLAB does not actually do this:

MATLAB>> data1 = randn(1,20000);
MATLAB>> kurtosis(data1)

ans =

    2.9825
+8
source

All Articles