Date:         Tue, 14 Jan 1997 09:48:35 +0000
Reply-To:     n.w.nelson@education.leeds.ac.uk
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         nick nelson <n.w.nelson@EDUCATION.LEEDS.AC.UK>
Organization: Leeds University
Subject:      Re: Aggregating
Content-Type: text/plain; charset=us-ascii

phewett@popcouncil.org wrote:
>
> Hello..
>
> I am attempting to create a variable that is a count (or mean, or max,
> etc.) of the number of observations of another variable (across
> cases), by subgroubs (splitfile). However, I want to assign this
> variable across all the cases, by subgroup. In Stata this is simply
> done using egen, but I don't know if there is an SPSS equivalent. The
> aggregate function does not seem to be too useful here. Any help ??
>
> ex:
>
> smoke pot ?     sex     newvar (count)   newvar (mean)
> 1               1               3                   .75
> 0               1               3                   .75
> 1               1               3                   .75
> 1               1               3                   .75
> 1               0               2                   .66
> 1               0               2                   .66
> 0               0               2                   .66
 
SORT CASES BY
  sex (A) .
AGGREGATE
  /OUTFILE='C:\XFILES\DATA\AGGR.SAV'
  /BREAK=sex
  /smoke_1 = MEAN(smoke)
  /N_BREAK=N.
MATCH FILES /FILE=*
 /TABLE='C:\XFILES\DATA\AGGR.SAV'
 /BY sex.
EXECUTE.
 
Regards, Nick.
 
BTW, this is pretty straightforward using dialogues rather than
syntax.