There is a rather long FAQ on this by Ulrich
Kohler and friend at
How do I deal with multiple responses?
http://www.stata.com/support/faqs/data/multresp.html
and also a long discussion based on the
same material in
On structure & shape: the case of multiple responses.
Stata Journal 3(1): 81--99 (2003)
The first will be accessible to you even
if the second one isn't.
I suggest that you read either.
While you're it, you should check out
Ben Jann's substantial contributions to this field,
Tabulation of multiple responses.
Stata Journal 5(1): 92--122 (2005)
The software is downloadable without subscription
to the Stata Journal.
Finally, there are some utilities in -tab_chi-
on SSC. To some extent, those (and Ben's commands)
allow descriptive analyses of multiple response data
without any -reshape-, as the programs do it for you on the fly.
Note that in an -update-d Stata
. search multiple response
would have pointed to all but the last of
these (-tab_chi-).
Besides that, you clearly should consider
subscribing to the Stata Journal (or reading
it!).
Nick
[email protected]
Arnold H. Levinson, Ph.D.
> I have a long dataset with variables that allow multiple responses per
> subject. So it looks like:
>
> ID var1 var2
> 1 3 5
> 1 2 .
> 2 3 4
> 2 . 5
> etc.
> Reshape doesn't like multiple responses for a single
> variable. What's the
> most efficient way of reshaping to wide, with each variable becoming a
> response-specific dummy, i.e., var1_1, var1_2, etc.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
