Re: st: Two-step Probit after merging

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Mon, 13/9/10, mmolina@uniroma3.it wrote:
> I would like to know, how realiable can be the results of a
> Probit model after merging two datasets.
> 
> I will try to explain it by using the example of the PDF
> information:
> 
> use http://www.stata-press.com/data/r11/sforce
> merge m:1 region using http://www.stata-press.com/data/r11/dollars
> 
> Suppose that the two data sets refer to two years (2 and
> 3). The model includes a dummy variable (sold=1, not sold=0) and the
> sales of year 3 depend on the costs of year 2. The model was run as:
> 
> ivprobit sold name i.region (sales3 = cost2), twostep
> 
> Looking at the ivprobit, timeseries varlists are admited.
> But in this case, these variables comes from two different datasets
> that have been merged.
> 
> Could you please tell me if there is another way to work
> with this merged dataset?

That really depends on the exact nature of your data, and on the kind of
problems your discipline thinks are important and which kind of 
problems are routinely ignored in your discipline. There are surprising
differences between disciplines on this account, which are to a large
extend responsible for the fact that there are many disciplines that
do not take each other seriously.

One thing that jumps out to me is that you seem to have a multi-level
structure. It would not do in my discipline to leave that as is, but
it may very well be that in your discipline this is considered 
unimportant. In the end you will need to please your peers, not mine.

-- 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/