Re: st: Producing graph with predictions after IVPOIS

[Austin Nichols <austinnichols@gmail.com>]

Anat (Manes) Tchetchik <anatmanes@gmail.com>
In addition to the oft-repeated "nots" of the Statalist FAQ, one can
offer the positive advice:
*Do use publicly available data* to construct an example of what you
want or the error you run into.

clear
use http://fmwww.bc.edu/RePEc/bocode/i/ivp_bwt.dta
g lnbw=ln(bw)
loc x "parity white male"
loc z "edfwhite edmwhite incwhite cigtax88"
ivreg2 lnbw `x' (cigspreg=`z')
*gmm (bw*exp(-{xb:`x' cigspreg}-{b0})-1), inst(`x' `z') onestep
*above works only in Stata 11+
ivpois bw `x', endog(cigspreg) exog(`z')
ren cigspreg tmp
g double pbw=.
qui forv i=1/6 {
 g cigspreg=(`i'-1)*10
 predict double xb
 g double exb=exp(xb)
 su exb, mean
 replace pbw=r(mean) in `i'
 drop xb exb cigspreg
 }
g c=(_n-1)*10 in 1/6
la var pbw "Predicted mean birthweight"
la var c "Cigarettes during pregnancy"
la var tmp "Cigarettes during pregnancy"
la var bw "Birthweight"
ren tmp cigspreg
sc bw tmp, msize(vsmall)||sc pbw c, c(l)


On Thu, Mar 7, 2013 at 11:29 AM, Nick Cox <njcoxstata@gmail.com> wrote:
> Specifically
>
> not explaining where user-written programs you refer to come from
>
> not explaining what "doesn't work" means
>
> not using the -cmdname- convention (neither do most people, but it's
> still a good idea).
>
> In fact these points don't really hinge on what is or is not in the
> FAQ or even whether there is one.
>
> Just imagine you wrote a program and someone says to you "I tried your
> program but it doesn't work". This is essentially no information.
>
> Or just imagine that you are the local Stata expert and someone says
> to you "I have problems with the -foobar- command". Your first
> question has to be "What is that?"
>
> Unclear questions waste everyone's time, including the people who
> wrote them, because they won't get good answers.
>
> Nick
>
> On Thu, Mar 7, 2013 at 3:08 PM, Anat (Manes) Tchetchik
> <anatmanes@gmail.com> wrote:
>
>> Hi Nick,
>> Thanks for your answer.
>> I probably missed something, What was the thing I should have done
>> which I haven't? is there an archive with FAQs  I could have search
>> for the answer there?
>> Anyway, I have posted two questions in the past few months, I didn't
>> realize it is that much.
>> Anat
>>
>> On Thu, Mar 7, 2013 at 3:09 PM, Nick Cox <njcoxstata@gmail.com> wrote:
>>> I have rewritten this question using Statalist FAQ guidelines. Why is
>>> this (seemingly) so difficult to understand and to do? Being new to
>>> Statalist can't be a reason here as Anat has posted many questions
>>> over some time.
>>>
>>> Anat Tchetchik, PhD
>>>
>>> I ran an -ivpois- model and want to produce few figures.
>>>
>>> -ivpois- is a user-written command by Austin Nichols from SSC.
>>>
>>> In the Poisson model I used to employ the -prgen- command, for example:
>>>
>>> prgen income, x( spain=1) from(1) to(14) ncases(14) gen(meetingsbrypr)
>>>
>>> But it doesn't work for -ivpois-. -prgen- is another user-written
>>> command, which I downloaded from <I can't fill that in, as several
>>> places could qualify>.
>>>
>>> Any ideas for other commands?
>>>
>>> Nick
>>>
>>> On Thu, Mar 7, 2013 at 12:54 PM, Anat (Manes) Tchetchik
>>> <anatmanes@gmail.com> wrote:
>>>> Hi all.
>>>>
>>>> I ran an IVPOIS model and wan't to produce few figures.
>>>> In the poisson model I used to employ the prgen command
>>>> (for example:  prgen income, x( spain=1) from(1) to(14) ncases(14) gen
>>>> (meetingsbrypr)
>>>> But it doesn't work for the IVpois
>>>> Any ideas for other commands?
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/