Page 1 of 1
Eviews 8 panel data structure (Pooled, fixed)
Posted: Wed Apr 22, 2015 8:09 am
by acemeister
1
Re: Eviews 8 panel data structure (Pooled, fixed)
Posted: Wed Apr 22, 2015 8:12 am
by EViews Gareth
That format is exactly how EViews prefers it. Just drag that file on to EViews and it will open as a panel.
Well, it would, but you have a data mistake. Row 298 has the wrong date (I'm guessing).
Re: Eviews 8 panel data structure (Pooled, fixed)
Posted: Wed Apr 22, 2015 8:58 am
by acemeister
wow thanks for the quick answer! :)
I still have some questions regarding the model. In the attached file, I did a test run.
I got quite different results with Eviewss and Gretl.
The results in Gretl are postive signficant, for the "Outside" variable (which was the expected result according to the theory), and corrected for standard erros.
In Eviews the results are negative for the "Outside" variable, also I was not able to correct for standard erros in Eviews, and when trying to include my manual dummies for time (year 1-7, excluding year 8, because of dummie-trao) in pooled model it said "near singluar matrix).
Thanks alot!
Re: Eviews 8 panel data structure (Pooled, fixed)
Posted: Wed Apr 22, 2015 9:06 am
by acemeister
The attached file!
Re: Eviews 8 panel data structure (Pooled, fixed)
Posted: Wed Apr 22, 2015 9:21 am
by EViews Gareth
I can't comment on Gretl, but I just ran that model in Stata and SAS, and they both match the EViews output.
Re: Eviews 8 panel data structure (Pooled, fixed)
Posted: Wed Jun 03, 2015 6:09 am
by acemeister
I can't comment on Gretl, but I just ran that model in Stata and SAS, and they both match the EViews output.
Gareth, I wonder if you could still comment on why does the R-squared become so high in models like this, is it becuase of the fixed effects on the cross sections?
Thanks
Ace
Re: Eviews 8 panel data structure (Pooled, fixed)
Posted: Wed Jun 03, 2015 7:03 am
by EViews Glenn
For the R2 that EViews computes, most probably yes.