Re: st: Selecting a sample to compromise between significant size and geographical dispersion

[Maarten Buis <maartenlbuis@gmail.com>]

I do not think there is a tradeoff, I think you have not decided yet
what your population of interest is. Either it is markets all over the
world or it is markets of a given size. This means you have to choose
one criteria and live with the consequences. Looking at data cannot
and certainly should not help you with that decision, nor can Stata or
any other software. This is a matter of theory alone. So ask yourself
why do you want to look at only large markets and why do you want to
look at markets all over the world. Than be very critical of your own
answers, because you need to eliminate one of these criteria.

Hope this helps,
Maarten

On Thu, Sep 15, 2011 at 10:31 AM, Partho Sarkar
<partho.ss+lists@gmail.com> wrote:
> Thanks for the prompt response Marten!  So there IS a trade-off it
> seems, right?  If I set a certain cut-ff level of market size, I get a
> certain geographical pattern.  What I want is a way to visualise this
> relationship- e.g., a table or graph that would show the pairs of
> cut-off size versus a dispersion "index" , so that I could then choose
> the "optimal" (subjectively) cut-off.  Or doesn't this make sense?  By
> the way, I wonder if this could be handled through Stata's Survey
> commands, which I have never gone into.
>
> Thanks again.
>
> Partho
>
> On Thu, Sep 15, 2011 at 1:36 PM, Maarten Buis <maartenlbuis@gmail.com> wrote:
>> What is the population that you want to generalize to? My suspicion is
>> that as soon as you have defined that, you'll see that there is no
>> trade-off. For example: if the population of interest is markets with
>> more than median trading volume than the geographic dispersion is
>> given by that constraint. Having two conflicting criteria suggests to
>> me that your definition of the population is still too fuzzy.
>>
>>
>> Hope this helps,
>> Maarten
>>
> *
> *   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/
>



-- 
--------------------------
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/