Displaying 1 result from an estimated 1 matches for "nonmarginal".
2008 Feb 26
3
OLS standard errors
...errors of the coefficients again
summary(reg)
The diagonal elements of the variance/covariance matrices should be the
standard errors of the coefficients. Both are identical when computed by
hand. However, they differ from the standard errors reported in
summary(reg). The difference of 1% seems nonmarginal. Should I have
multiplied var(e)*solve(t(z)%*%z) by n and divided by n-1? But even if I do
this, I still observe a difference. Can anybody help me out what the source
of this difference is?
Cheers,
Daniel
-------------------------
cuncta stricte discussurus