[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Update to -outreg2-

From   Roy Wada <>
To   <>
Subject   st: Update to -outreg2-
Date   Thu, 10 Apr 2008 22:34:01 -0700

Okay, let's try this again:

Thanks to Kit Baum, an update to -outreg2- is available from SSC (since 
last month, actually). In the rare case that someone has not figured this 
out, please download the latest version from SSC instead of contacting 
the help desk at the StataCorp.

New options have been added while I was in there. You should read the 
following if you use -outreg2-.

(1) reporting marginal effects with -mfx- option

. sysuse auto, clear
. logit foreign price rep head length
. mfx
. outreg2 using myfile, mfx see
For multiple outcomes, you might want to consider using -mfx2, replace- 
by Richard Williams or -margeff, at(mean) replace- by Tamas Bartus. They 
are also available from SSC for your downloading pleasure.

(2) omitting standard error with -nose- option. Only the estimated 
coefficients are reported for each variable.
(3) selectively reporting variable-specific statistics with -stat(coef se 
tstat pval ci aster)- option. The stats will be reported in the order specified. 
The usual table would be duplicated by the following:

. outreg2 using myfile, stat(coef se)

The -nose- option mentioned above would be duplicated by the following:

. outreg2 using myfile, stat(coef)

Please note that -stat( )- option currently does not work with -eform- or 
-beta- options. (It would be reasonable to expect them to transform the 
estimated coefficient, yet some people may want to report both the 
raw and the transformed coefficients. I'm guessing it needs to be done 
both ways.)

(4) Medical/health tables that goes sideways. A table can be made to go 
sideways by building it up one column at a time.

. outreg2 using myfile, stat(coef) ctitle(coef) replace
. outreg2 using myfile, stat(se) ctitle(se)
. outreg2 using myfile, stat(aster) ctitle(aster) see

The -stat(aster)- would place asterisks in its own column next to the 
reported numbers, which would allow you to left-justify the asterisks 
and have them hugging the numbers, or leave them as they are. You 
would be doing this in MS Word or Excel.

The current help file -help outreg2- is a little skimpy on these issues. 
I wish someone would write a compiler/interface for writing Stata help 
file, possibly basing it on one of TeX engines.

Or go the other way around and use TeX as the basis for the Stata 
help file, which in its current markup form can be difficult to read on 
the screen, but this obviously wouldn't work unless someone makes it 


Pack up or back up–use SkyDrive to transfer files or keep extra copies. Learn how.
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index