Re: [Gretl-users] small bug
On Wed, 5 Jul 2006, James Wells wrote:
Hooray for the QLR test recently added. There is a small
cosmetic bug. If I run a regression, then do a QLR test, then
do some other test, the model window doesn't have a break
between the QLR and the next test. For example:
QLR test for structural break -
Null hypothesis: no structural break
Test statistic: F(5, 31) = 2.71302
(0 percent critical value = 0)RESET test for specification -
Null hypothesis: specification is adequate
Test statistic: F(2, 34) = 1.17283
with p-value = P(F(2, 34) > 1.17283) = 0.321698
Aside from the line-break bug there's also a more substantive
bug here: "(0 percent critical value = 0)". Both are fixed in
CVS now.
I'll put up a new Windows snapshot shortly.
Allin Cottrell