Date:         Thu, 1 May 2003 17:35:00 -0400
Reply-To:     Nancy Crowell <nac2@georgetown.edu>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Nancy Crowell <nac2@georgetown.edu>
Organization: Georgetown University
Subject:      Re: Syntax help-working with lag & loops
Content-Type: text/plain; charset=us-ascii

Clarification:

I don't want to aggregate by job number.  The variable I want to create is
for each job the cumulative months employed -- so for job #1, the months
in the job equal the total months employed; for job #2, total months =
months in job 1 plus months in job 2.  and so forth.  I am going to be
running an event history analysis which requires that each job be a case
of its own, meaning that for any given id there will be multiple cases
unless they have only had one job in their entire career.

Hope that clarifies my question.

Thanks.

Richard Ristow wrote:

> At 11:16 AM 5/1/2003 -0400, Nancy Crowell wrote:
>
> >I have a data file that has multiple records for each id number--each
> >record corresponds to a job held by the individual. The first thing I
> >want to do is to create a variable 'total months employed'.
> >
> >Data looks like this
> >
> >ID    Jobnbr    Jstart    Jend    JMos
> >1    1     1/1982    6/1984    29
> >1    2    7/1984    6/1988    47
> >1    3    7/1988    6/1995    83
> >1    4    7/1995    6/1998    35
> >1    5    7/1998    2/2003    56
> >2    1    9/1977    5/1981    44
> >2    2    5/1981    8/1982    15
> >2    3    8/1982    7/1983    11
> >2    4    7/1983    1/2003    234
> >3    1    9/1974    8/1978    47
> >3    2    8/1980    1/2003    269
> >
> >For each successive job for each person I want to create a variable
> >that shows the number of total months employed (i.e. jstart for job=1
> >subtracted from jend for job in question).
>
> I'm not *quite* clear. It looks like your variable "JMos" has total
> months employed for that person for that job, as you were looking for.
> Are you having trouble calculating "JMos"?
>
> In the meantime, if you want to combine to total months employed per
> person in any job, while it can be done with LAG, AGGREGATE is much
> easier and more natural:
>
> AGGREGATE/OUTFILE=*/BREAK=ID/
>     EmpMos 'Total months employed, all jobs'= SUM(JMos).
>
> Does this get you near where you want to go?