[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

OLS & Tobit consistence [WAS: Re: st: RE: -log- in Stata to bypass commands that resulted in errors?]

From   Nick Winter <[email protected]>
To   [email protected]
Subject   OLS & Tobit consistence [WAS: Re: st: RE: -log- in Stata to bypass commands that resulted in errors?]
Date   Fri, 31 Oct 2008 13:42:55 -0400

Is this a homework assignment?

simanti banerjee wrote:
I have a few qustions regarding postestimation commands in stata.

1. I have performed two sets of trgressions, one with OLS and the other with Tobit. q1 is there a way to show that the estimators are consistent? Also, can we compare the consistrency properties of both? q2. Both sets of regressions are run on orthogonalised variables. Does this information help in any way to show that the estimators are consistent directly from properties of matrix algebra without using stata commands?
Please help.



Thu, 30 Oct 2008 Nick Cox wrote :
 >I don't think this is possible directly, at least not for users.
 >I think it would be feasible to write something, e.g. in Mata, that did
 >much of that retrospectively, but it seems not to have tempted anyone as
 >a project.
 >The biggest reason, in my experience, is that you want to edit a log not
 >just for mistakes, but also for repetitions, digressions, and things
 >that were legal but not interesting or useful
 >In any really serious project, you should use -log- and -cmdlog-
 >simultaneously. Then with an eye on the log, edit the command log so
 >that it contains only commands that were legal and useful. Then run the
 >command log as a do file and get the log you now wish you had.
 >[email protected]
 >Mingfeng Lin
 >I'm just curious if there is a way (or maybe an add-on) in Stata to
 >create a log file for only the correct commands? -log using- or -#
 >review- will capture all commands that have been typed in the window,
 >including those resulting in errors.  If there is a way for the log to
 >automatically filter out those commands, it would be really nice,
 >especially replicating what we do in an interative manner with Stata.
 >*  For searches and help try:

Rediff Shopping <>

Nicholas Winter                                 434.924.6994 t
Assistant Professor                             434.924.3359 f
Department of Politics                  [email protected] e
University of Virginia w
PO Box 400787, 100 Cabell Hall
Charlottesville, VA 22904

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index