Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: nlcom produces different standard errors


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: nlcom produces different standard errors
Date   Sun, 5 Dec 2010 17:22:03 -0600

The latter part does not account for sampling variability in
-thisisit- scalar. If you made it a -local- instead of a -scalar-, it
would be substituted symbolically, and -nlcom- would know that it
needs to use it as a part of s.e. calculation. Your last -nlcom- is
essentially

lincom (normalden(-.914485))*_b[x]

On Sun, Dec 5, 2010 at 4:10 PM, Tunga Kantarcı <tungakantarci@gmail.com> wrote:
> In the code below the two nlcom commands produce different standard
> errors. I pasted the output below. Why is this happening? I expected
> exactly the same standard error produced from the two nlcom commands.
>
>
>
>
>
>
> probit y x
> nlcom normalden(_b[_cons]+_b[x]*0.4)*_b[x]
> scalar thisisit = _b[_cons]+_b[x]*0.4
> nlcom normalden(thisisit)*_b[x]
>
>
> . probit y x
>
> Iteration 0:   log likelihood = -1083.6381
> Iteration 1:   log likelihood = -1024.6386
> Iteration 2:   log likelihood = -1024.5509
> Iteration 3:   log likelihood = -1024.5509
>
> Probit regression                                 Number of obs   =       2708
>                                                  LR chi2(1)      =     118.17
>                                                  Prob > chi2     =     0.0000
> Log likelihood = -1024.5509                       Pseudo R2       =     0.0545
>
> ------------------------------------------------------------------------------
>           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
> -------------+----------------------------------------------------------------
>           x |   3.180135   .3136994    10.14   0.000     2.565296    3.794975
>       _cons |  -2.186539   .1132633   -19.30   0.000    -2.408531   -1.964547
> ------------------------------------------------------------------------------
>
> . nlcom normalden(_b[_cons]+_b[x]*0.4)*_b[x]
>
>       _nl_1:  normalden(_b[_cons]+_b[x]*0.4)*_b[x]
>
> ------------------------------------------------------------------------------
>           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
> -------------+----------------------------------------------------------------
>       _nl_1 |   .8351398   .0978799     8.53   0.000     .6432987    1.026981
> ------------------------------------------------------------------------------
>
> . scalar thisisit = _b[_cons]+_b[x]*0.4
>
> . nlcom normalden(thisisit)*_b[x]
>
>       _nl_1:  normalden(thisisit)*_b[x]
>
> ------------------------------------------------------------------------------
>           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
> -------------+----------------------------------------------------------------
>       _nl_1 |   .8351398    .082381    10.14   0.000     .6736759    .9966036
> ------------------------------------------------------------------------------


-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.

*
*   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   |   Site index