Re: st: referencing values of others on a variable

[Nick Cox <njcoxstata@gmail.com>]

Let's look at

samerace= race==race[idstudent=idalter]

and fill in the implicit -generate- and correct the illegal = to ==

gen samerace= race==race[idstudent==idalter]

This is now legal but a long way from what you want.

The expression

idstudent==idalter

will be true (1) for observations in which idstudent == idalter and
false (0) otherwise. That means  false always, presumably. So Stata
sees

gen samerace = race==race[0]

but -race[0]- is outside the range of the data and therefore evaluated
as missing. So you end with  an indicator for missing race.

A strategy for your kind of data was spelled out at
http://www.stata-journal.com/sjpdf.html?articlenum=dm0043 As it is
quite detailed, I will not repeat it here.

Nick
njcoxstata@gmail.com


On 9 July 2013 09:58, Viktor Emonds <Viktor.Emonds@soc.kuleuven.be> wrote:

> I have a file where for each student, there are as many observations as there are possible ties to other classmates (directed, based on nomination), with the following basic structure:
>
> id student  id alter     tie
>
> 001               002      1
> 001               003      0
> 001               004      0
> 001               005      1
> 002              001        0
> 002             003         0
> 002              004       0
> 002              005       1
>
> I would like to compute indices of friendship segregation along ethnic lines and need to evaluate whether the ethnicities of ego and alter match or not. How to do so in Stata syntax?
>
> I know you can refer to values on a variable for another observation, e.g. samevalue= var==var[_n-1], but I cannot seem to extend this to something like samerace= race==race[idstudent=idalter]. Any ideas on how I can do this?
*
*   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/