Re: [Gretl-users] Serial correlation and omit
On Fri, 9 Jan 2015, Logan Kelly wrote:
Apologies for a very rudimentary question. Is the joint
significance test performed by omit valid in the presence of
serial correlation when (i) the --robust option is used to with
ols and (ii) the --chi-square option is used with omit, e.g.
ols y const x z --robust
omit z --chi-square
When the model is estimated with the --robust flag the "omit" test
is carried out using the robust estimate of the covariance matrix of
the parameter estimates, so it's valid to that extent. Use of the
--chi-square option just means that no degrees of freedom adjustment
is used, but you know what that means.
Allin