RE: st: nlcom and delta method

["Dr. Martin Weiss" <martin.weiss1@gmx.de>]

<>

Maybe calculate the two models and -suest- them, which allows you to -test-
them jointly.


HTH
Martin


-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Beatrice Crozza
Sent: Dienstag, 3. August 2010 20:08
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: nlcom and delta method

Dear Maarten,

thank you very much for your help.

It works perfectly!

The only thing is that when I calculate one non linear combination and
I save it with the post option and then I calculate another no linear
combination, after a different maximum likelihood, I cannot compare
the two values with the test command.

Do you have any idea of how to solve this problem, or there is no solution?
Because in order to calculate the test by myself I should also need
the covariance.

Thanks a lot.

Bea

2010/8/3 Maarten buis <maartenbuis@yahoo.co.uk>:
> --- On Tue, 3/8/10, Beatrice Crozza wrote:
>> I did a maximum likelihood, using delta method, to get the
>> values.
>>
>> So, for example, one result of my maximum likelihood using
>> delta method is this one:
>>
>>
>> . nlcom (a: 1/(1+exp(-[ar]_b[_cons])))
>>
>> Now I wan to test a no linear combinations, using the
>> coefficient of a , however the usual command for the
>> coefficient: _b[a] doesn't work.
>
> if you want to access _b[a] then the command that created
> a must be an estimation command, and by default -nlcom-
> is not an estimation command. You can turn -nlcom- into
> an estimation command by specifying the -post- option.
> However now the results of your previous maximum
> likelihood command are overwritten. So this won't work if
> you want to use parameters from your maximum likelihood
> command. In that case it is probably best to do the entire
> opperation in one -nlcom- command. In that case it is
> often convenient to store a string containing the command
> used to create a (n.b. not the value but the command) in
> a local. Say that your second non-linear combination is
> a / [ar]_b[x], then I would do that as follows:
>
> local a "( 1/(1+exp(-[ar]_b[_cons])) )"
> nlcom (b: `a'/[ar]_b[x])
>
> Hope this helps,
> Maarten
>
> --------------------------
> Maarten L. Buis
> Institut fuer Soziologie
> Universitaet Tuebingen
> Wilhelmstrasse 36
> 72074 Tuebingen
> Germany
>
> http://www.maartenbuis.nl
> --------------------------
>
>
>
>
>
> *
> *   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/
>

*
*   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/


*
*   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/