Re: st: Actor-partner Interdependence Model data structure (redo)

[Scott Merryman <scott.merryman@gmail.com>]

clear
 input egoid egodelinq egoattach altid altdelinq altattach
 111         3           4          222     8           9
 444         5           1          333     2           7
 end

gen dyad = _n
rename (egoid altid) (person1 person2)
reshape long person,i(dy)
order dy person egod altd egoa alta

bys dy: replace egodel = altdel[_n-1] if _j ==2
bys dy: replace altdel = egodel[_n-1] if _j ==2
gen attachment = egoa if _j ==1
replace attach = alta if _j ==2
drop egoa alta _j
l

Scott


On Sat, Apr 28, 2012 at 1:30 PM, Carter Rees <carterrees@gmail.com> wrote:
>
> Sorry for the messy format of the previous post.
>
> Hi Statalisters,
>
> I am trying to structure my data as a pairwise data set (see Kenny,
> Kashy, Cook (2006) Dyadic Data Analysis) and then run a multilevel
> APIM model.
>
> My data set is currently at the individual level as follows:
>
> egoid egodelinq egoattach altid altdelinq altattach
>  111         3           4          222     8           9
>  444         5           1          333     2           7
>
> What I would like it to look like (also individual level):
>
> dyad    person egodelinq altdelinq attachment
> 1            111        3            8           4
> 1            222        8            3           9
> 2            444        5            2           1
> 2            333        2            5           7
>
> The outcome variable is friendship attachment level. The predictor
> variables are ego delinquency and alter (friend) delinquency. Each
> individual's outcome score is associated with both his or her own
> predictor scores and his or her partner's scores.
>
> Much appreciated,
>
> Carter

*
*   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/