Regression output: Models in rows instead of columns

[Jan Leicht]

Hi everyone,


I loop multiple time-series regressions with > 1000 dependent variables but only roughly 40 independent ones for a forecasting analysis. Certainly, I group these models so that it will not be one single table output. However, the standard output formats employed by outreg, outreg2, or esttab will always display the statistics of each model in a column.

-

Model 1	Model 2	Model n
Cons	Cons	Cons
(SE)	(SE)	(SE)
Coef	Coef	Coef
(SE)	(SE)	(SE)
R2	R2	R2
N	N	N

-

The only option that seems specifiable in this respect depends on whether I merge (--> in several columns) or I append (--> models in 1 column) the different regression models.
Given that I have such a high number of different models, I want each model to be displayed in a row with the statistics and independent variables as columns.

-

Model 1	Cons	Coef	R2	N
	(SE)	(SE)
Model 2	Cons	Coef	R2	N
	(SE)	(SE)
Model n	Cons	Coef	R2	N
	(SE)	(SE)

-

Hopefully, the tables clarify what I mean. I thought that someone else has run across this issue, but did not find any related thread or any other hint in the extensive help files (and I tried a lot with the different commands' options). Is there any possibility to do this in STATA?

Many thanks for your time and hints!

Jan

[Maria Boutchkova]

Hi,

I create my own output matrix and copy in it the contents of e(b) and then use frmttable to make it look nice. You can add more columns for e(r2) and e(N). See my earlier question for example code that appends each consecutive regression results below the previous.

http://www.statalist.org/forums/foru...able-in-a-loop

Maria