Statalist The Stata Listserver


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

Re: st: How do I test if coefficients are sign different from each other in grouped data?


From   Austin Nichols <[email protected]>
To   [email protected]
Subject   Re: st: How do I test if coefficients are sign different from each other in grouped data?
Date   Fri, 10 Feb 2006 05:37:35 -0500

You indicate that you have included all categories and all
interactions, which is not possible, even with the -nocons- option
(though you could produce such a model after the fact: -findit
devcon-).  You should always show your (excerpted) output if you want
a complete response to your question.  Assuming all your indicator
(dummy) variables and interactions have names starting with "size" or
"order", and their coefficients measure the difference from some base
category, you might want to use a  command like
. testparm order* size*
for a joint test of all possible differences=0.  Other more subtle
tests are possible with a series of  commands of the form
. test order1=order2
. test order1size2=order2size2, accum
etc.

But note you only get one test using these commands "for free" after
you estimate your model--any subsequent tests of this form will have
incorrect size, as.you will have entered the world of multiple
hypothesis testing.

F�rst�r du? Ska jag f�rs�ka ber�tta det p� svenska?  Jag tror
inte--din engelska �r mycket b�ttre �n min svensk.

On 2/10/06, lena Lindahl <[email protected]> wrote:
> Y=X+(ORDER1-3) + (SIZE1-3) + (ORDER1-3)*(SIZE1-3) + X*(ORDER1-3)+
> X*(SIZE1-3) + X*(ORDER1-3)*(SIZE1-3)
> Now I want to test if there are any significant birth-order and
> family-size effects.

*
*   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/



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