OK. Try this:
input a b
5 5
6 11
end
gen long id = _n
expand 2 if a != b
bysort id : replace id = cond(_n == 1, a, b)
replace a = a == id
replace b = b == id
l
Nick
[email protected]
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Michael
McCulloch
Sent: 21 October 2009 18:21
To: [email protected]
Subject: Re: st: RE: creating expanded records with indicator variables,
where different ID's reside in the same rows
Sorry Nick, yes you're correct, it should be:
id A B
5 1 1
6 1 0
11 0 1
Michael
On Oct 21, 2009, at 10:18 AM, Nick Cox wrote:
> Sorry, but I can't even follow your first two steps. Why isn't the
> second observation 6 1 0?
>
> Nick
> [email protected]
>
> Michael McCulloch
>
> I have a data structure that has been gathered as follows, where each
> cell contains the subject ID if A is true, and similarly for B.
> Therefore, for subject 5, I have A=yes and B=yes, whereas for subjects
> 6 and 11, only B=yes.
>
> A B
> 5 5
> 6 11
>
> I'd like to transform the data structure such that each row would have
> ID, and indicator variables for A and B.
> id A B
> 5 1 1
> 6 0 1
> 11 0 1
>
> I've tried an approach like this, where I create indicator variables
> but A and B, but am not sure how to "expand" from the data structure
> where different ID's reside in the same rows.
> . gen id1=a
> . gen id2=b
> . gen aa=1 if a!=.
> . gen bb=1 if b!=.
>
> It also looks like -reshape- is not suitable because different ID's
> reside in the same rows.
> Any advice would be 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/
