Statalist


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

RE: st: RE: Graphing xttobit predictive values


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: RE: Graphing xttobit predictive values
Date   Fri, 20 Mar 2009 12:13:34 -0000

I would like to underline Johannes' very proper comment: Apply with
caution. 

As a program author (although not of anything mentioned here) I would
say: You are free to modify my programs but those modified programs are
then yours and I am not then responsible for what they do or do not do. 

Other authors can speak for themselves but I imagine that a similar
attitude is common. 

Nick 
n.j.cox@durham.ac.uk 

Johannes Geyer

Apply with caution: You can make -postgr3- work with xttobit (I have no 
experience with xtlogit but it might also work) - you have to include a 
line in a subprogram that is called by -postgr3- and allow explicitly
for 
the command -xttobit-  in -_pebase.ado-  and/or -_pebase2.ado- (simply 
copy the line for -tobit- and replace it by xttobit in these ado-files).

However, I guess the authors might have had a reason not to include
these 
random effects models in their program.

Nick Cox 

> My impression first time, and this time too, is that it's most
unlikely
> that a canned Stata routine exists for this task. I guess that others
> who work in similar territory had vague thoughts along similar lines. 
> 
> I don't think Excel is _needed_ for this graph. I would expect that
you
> could do it in Stata, but you'd need to write the code yourself. 

Erick Guerrero, M.A., Ph.D. Candidate
 
> I am trying to get a bar graph that draws from a regression 
> that has predicted values on the Y axis and bars on the x 
> axis representing two categorical variables i.e. 
> organizations offering  methadone tx.and comparing these 
> organizations based on wave number (1995 vs. 2005) holding 
> all other covariates constant at their mean. 
> 
> I used the xi3 command followed by postgr3, but the postgr3 
> command did not work for xttobit or xtlogit, random effects.
> 
> Is there a command like postgr3 that works with xttobit or 
> xtlogit holding all other covariates at their mean and that 
> graphs different categorical variates in the same graph?
> 
> Do I need to use excel to graph these relationships also?

*
*   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