RE: st: RE: predictnl for effects and se of interaction terms in probit model

["shuang zhang" <shuangshuangmm@hotmail.com>]

Dear Maarten,

Thanks for you kindly reply!
Actually, I have read the websits you gave to me, and I really writing my 
command like that. I have tried and failed. Can you give me some 
suggestions?

Interaction terms in my probit model are n1*m, t1*m, n1*m, n2*m
My command to estimate the interaction term as following:

probit poverty0 n1 t1 n2 t2 m n1m t1m  n2m t2m
qui sum n1 if e(sample)
local meann1=r(mean)
qui sum t1 if e(sample)
local meant1=r(mean)
qui sum n2 if e(sample)
local meann2=r(mean)
qui sum t2 if e(sample)
local meant2=r(mean)
qui sum market if e(sample)
local meanm=r(mean)
local xb _b[n1]*`meann1' + _b[t1]*`meant1'+ _b[n2]*`meann2'+ 
_b[t2]*`meant2'+ 
_b[n1m]*`meann1'*`meanm'+_b[t1m]*`meant1'*`meanm'+_b[n2m]*`meann2'*`meanm'+_b[t2m]*`meant2'*`meanm'_b[m]*`meanm'+ 
_b[_cons]
predictnl dydn1m=normden(`xb')*(-(`xb'))*(_b[n1]+_b[n1m]*`meanm')*(_b[m]+ 
_b[n1m]*`meanm'+_b[n2m]*`meanm'+_b[t1m]*`meanm'+_b[t2m]*`meanm')+ 
normden(`xb')*( _b[n1m]) in 1, se(sen1m)

But STATA report that the expression of last command"predictnl" is too long. 
So, I can't get results.
I drop one interaction term in my model, and I can run the command above 
very well and get results. So, I am rather confused why the expression is 
too long.

Thanks very much!


Cheers,
Zhang Shuang




> From: "Maarten Buis" <M.Buis@fsw.vu.nl>
> Reply-To: statalist@hsphsun2.harvard.edu
> To: <statalist@hsphsun2.harvard.edu>
> Subject: st: RE: predictnl for effects and se of interaction terms in 
> probit model
> Date: Fri, 2 Dec 2005 17:15:03 +0100
>
> Dear Zhang Shuang:
>
> I cannot correct your command since you have not given it to us, or the 
> exact error message. But calculating marginal effects with interaction 
> terms is a FAQ: see: 
> http://www.stata.com/support/faqs/stat/mfx_interact.html and threads 
> started with 
> http://newwww.stata.com/statalist/archive/2004-11/msg00254.html and 
> http://www.stata.com/statalist/archive/2004-11/msg00346.html . These 
> contain lots of example code that might help.
>
> HTH,
> Maarten
>
>
> -----------------------------------------
> Maarten L. Buis
> Department of Social Research Methodology
> Vrije Universiteit Amsterdam
> Boelelaan 1081
> 1081 HV Amsterdam
> The Netherlands
>
> visiting adress:
> Buitenveldertselaan 3 (Metropolitan), room Z214
>
> +31 20 5986715
>
> http://home.fsw.vu.nl/m.buis/
> -----------------------------------------
>
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu]On Behalf Of shuang zhang
> Sent: vrijdag 2 december 2005 17:04
> To: statalist@hsphsun2.harvard.edu
> Subject: st: predictnl for effects and se of interaction terms in probit 
> model
>
> Hi,
>
> In my probit model, I have independent variables X1,X2,X3,X4,M, and the
> interaction terms X1*M, X2*M, X3*M, X4*M. I want to estimate the marginal
> effects of these variables, especially that of interaction terms. I know 
> the
> command "predictnl" may help me to get right marginal effect and se.
> However, when I do the command with all these variables above, I meet
> problems.
> 1, I can get marginal effect and se of X1,X2,X3,X4,M by "predictnl".
> 2, When I want to estimate the interaction terms and write the command for
> them, the STATA report that my expression of "predictnl" is too long!
> 3, For a check, I use just X1,X2,X3,M and X1*M, X2*M, X3*M, which including
> three interaction terms. I can get the marginal effect and se of 
> interaction
> terms by "predictnl". But I can't get results when I add X4 and have four
> interaction terms, where STATA told me my expression was too long.
>
> I am rather confused by "predictnl". I am sure my command writtened
> correctly, but it can't work when I have four interaction terms!
> Any advice on this problem is welcome!
>
> Thanks very much!
>
>
> Cheers,
> Zhang Shuang
>
> _________________________________________________________________
> On the road to retirement? Check out MSN Life Events for advice on how to
> get there! http://lifeevents.msn.com/category.aspx?cid=Retirement
>
> *
> *   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/
>
> *
> *   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/
_________________________________________________________________
Don�t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/

*
*   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/