Statalist


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

st: re: lincom


From   Kit Baum <baum@bc.edu>
To   statalist@hsphsun2.harvard.edu
Subject   st: re: lincom
Date   Thu, 18 Jun 2009 11:20:38 -0400

<>
Martin said

It is useful indeed, but there is no extended discussion of -lincom- of the sort requested by Nikolaos in there.

I don't know what such an extended discussion would contain. After any estimation command you have a coefficient vector b=e(b) and a precision matrix e(V) [VCE]. If you want to examine any linear combination of coefficients, alpha' * b, calculating that combination is simple algebra which you could do yourself. Calculating its standard error, using e(V) and the standard rules for the variance of a random variable, is also something you could do yourself if you had e(V) at hand (-estat vce-). There is nothing fancy going on here. If I want to form the lincom of 2 * beta_3 - 4 * beta_5, I can calculate that expression with display, and I can calculate its s.e. using three elements of the VCE (the 3,3, 5,5, and 5,3 element). What more is there to say about lincom? It's just a convenience command.


Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
An Introduction to Stata Programming | http://www.stata-press.com/books/isp.html
   An Introduction to Modern Econometrics Using Stata  |   http://www.stata-press.com/books/imeus.html



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index