To add my two cents, the "official" version of outreg available from
SSC (3.1.7 of 2002-05-14) is USELESS to me, since it does not support
LaTeX output, and I do not use MS Word for scholarly work. I
complained to John Gallup for a long time that it would be very easy
to add a "tex" option, and he did, to version 4.0.0 of 2001-02-26:
but that never became the public version of outreg, and in fact the
3.1.7 version has added options lacking from 4.0.0. I further hacked
4.0.0 to add support for ivreg2.
The latter issue highlights the reason why estout is a superior
approach: outreg has to be programmed to handle everything, and a
thousand questions on Statalist have read "outreg does not work
after ..." or "outreg works after this command, but does not give me
access to..."
I am interested in fully automated table production (not quite as
ambitious as Phil Schumm's approach, presented last summer in Boston,
but...) so I want an output routine that works after EVERY estimation
command and allows me to add ANY statistic to ANY set of estimates.
The estout / estadd / eret2 suite does that, and it leverages Stata's
own -estimates- technology to do so. If I write a new estimation
command (e.g. ivreg2), it will work immediately with estout since it
produces fodder for -estimates-. It will not work with outreg --
original or revived -- until the author has modified estout to
recognize its peculiarities. In that regard, Roy has signed up for
full time employment if he is going to make outreg as capable as estout.
Wrappers are easy. Stata itself has many examples of very simple
commands that can be made much more complex by adding options. I dare
say that Ben Jann could write a version of estout that allows for
several schemes--perhaps those most commonly used in preparing
estimation tables--that would allow taking default values for almost
all options, and come down to a syntax as easy as that of outreg.
outreg was a marvelous bit of programming in its day, but today I
don't have to write my own relational database routines; I just call
mySQL. I don't have to write my own character parser; I just use
perl. Likewise, Ben did not have to write code which handles all of
the -estimates- details, because the fine people of Stata have
already done that.
Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html
On Sep 15, 2005, at 2:33 AM, statalist-digest wrote:
As a user of both outreg and estout, I'd have to say "none" is not
exactly correct. estout is a terrific program, it is quite
powerful and
very flexible (and Ben deserves lots of credit for creating it). That
said, as is the case with many/most powerful and very flexible
programs,
the syntax is not concise (in the sense that there is too much to
memorise). =20
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
