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: goodness of fit measure fir ivtobit


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: goodness of fit measure fir ivtobit
Date   Tue, 25 Sep 2012 09:32:13 -0400

Anat (Manes) Tchetchik <anatmanes@gmail.com>:
1. You can specify that max as exposure, but it sounds a bit odd to
me--why not just include ln(number_of_years_rel_abroad) and ln(age-17)
as predictors and let the regression estimate the coefs?  If one coef
turns out to be close to one, then it looks like an exposure variable:
an exposure X just means that ln(X) enters with a coef of one
(elasticity of one). An offset X just means that X enters with a coef
of one (semi-elasticity of one).
3. Yes, predict yhat and calculate squared corr of y and yhat as a
pseudo-R-squared.

On Tue, Sep 25, 2012 at 7:49 AM, Anat (Manes) Tchetchik
<anatmanes@gmail.com> wrote:
> Austin hi,
> I have read the material regarding the ivpois model you have referred
> me to and ran it on my data set. The results look pretty plausible .
> Just before sticking to this model I have few last questions:
>
> 1. Can I use the exposure variable (in my case the maximum of the two
> values: respondent's adulthood years and and the no. of years his/her
> relative is staying abroad) also as an independent var.? (or given
> that I do it- I should not use the exposure option)
>
> 2. I'm not sure i understand when I should use the exposure option and
> when the offset one
>
> 3. Do I calculate goodness of fit measure using the same procedure you
> recommended earlier
> (http://www.stata.com/support/faqs/statistics/r-squared/) ?
>
> Thank you very much!
> Anat
>
> On Mon, Sep 24, 2012 at 6:38 PM, Austin Nichols <austinnichols@gmail.com> wrote:
>>
>> Anat (Manes) Tchetchik <anatmanes@gmail.com> :
>> Indeed, no censoring in your model. You have exactly the case I
>> referred to when I wrote "I suspect you have a lower limit at zero
>> which is actually a very low conditional mean rounded down to zero."
>> Unless you believe that the Tobit model somehow correctly captures
>> bunching at zero due to utility functions which would imply a negative
>> demand for travel abroad, if such a thing were possible, which is
>> implausible at best, you are much better off assuming that people with
>> zero travel abroad simply have very low demand, and a group that has a
>> conditional mean of 1/100000 trips will indeed have a lot of zeros
>> observed.  The -ivpois- package on SSC, and the -gmm- specifications
>> that supersede it, are designed to allow instrumental variables in a
>> count model, or a regression with nonnegative outcomes more generally:
>> http://www.stata.com/meeting/boston10/boston10_nichols.pdf
>> http://fmwww.bc.edu/repec/bocode/i/ivpois.html
>>
>> On Mon, Sep 24, 2012 at 10:54 AM, Anat (Manes) Tchetchik
>> <anatmanes@gmail.com> wrote:
>> > I haven't thought about the count model, I will definitely try to run
>> > it! thanks much!
>> >
>> > On Mon, Sep 24, 2012 at 5:38 PM, Maarten Buis <maartenlbuis@gmail.com> wrote:
>> >> That does not sound like censoring at all. I would think of this as a
>> >> regular count model. There are examples on how to deal with such an
>> >> iv-model in -help gmm-.
>> >>
>> >> Hope this helps,
>> >> Maarten
>> >>
>> >> On Mon, Sep 24, 2012 at 4:11 PM, Anat (Manes) Tchetchik
>> >> <anatmanes@gmail.com> wrote:
>> >>> Austin Hi,
>> >>> Thank you very much for your reply!
>> >>> What I have as a dependent var. are 500 respondents' reports of the
>> >>> number of times they travelled abroad to visit their friends and
>> >>> relatives over the course of their adult lives.  Some respondents yet,
>> >>> who have relatives abroad, did not travel at all.
>> >>> So the observations are censored at zero, with mean =2.2, max =50 and
>> >>> stdev= 3.8.
>> >>> Do you think in that case that the general methods of moments will be better?
>> >>> Thanks much!!!
>> >>> Anat
>> >>>
>> >>> On Sun, Sep 23, 2012 at 5:49 AM, Austin Nichols <austinnichols@gmail.com> wrote:
>> >>>>
>> >>>> Anat (Manes) Tchetchik <anatmanes@gmail.com>:
>> >>>> You can always -predict- and compute the squared correlation of
>> >>>> predictions with observed values:
>> >>>> http://www.stata.com/support/faqs/statistics/r-squared/
>> >>>> but are you sure your -ivtobit- model is justified?  What is the
>> >>>> process that results in observations being censored?  I suspect you
>> >>>> have a lower limit at zero which is actually a very low conditional
>> >>>> mean rounded down to zero--am I right?  You may be better off with a
>> >>>> -gmm- model.
>> >>>>
>> >>>> On Sat, Sep 22, 2012 at 5:33 PM, Anat (Manes) Tchetchik
>> >>>> <anatmanes@gmail.com> wrote:
>> >>>> > Dear statalisters,
>> >>>> >
>> >>>> > I wonder if anyone knows any goodness of fit that is appropriate for
>> >>>> > tobit with endogenous
>> >>>> > variables (ivtobit). Not as in "regular" tobit, stata does not report any
>> >>>> > goodness of fit measure, any idea how to estimate such a measure?
>> >>>> > Any response will be greatly appreciated..
*
*   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