Date:         Fri, 29 Dec 2000 12:37:39 -0500
Reply-To:     Raynald Levesque <rlevesque@VIDEOTRON.CA>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Raynald Levesque <rlevesque@VIDEOTRON.CA>
Subject:      Re: Aggregating data in SPSS
Comments: cc: Steven Delaronde <Delaronde@CONNECTICARE.COM>
In-Reply-To:  <850FE2CA73BAD01187FA0060976710C2051B1E52@CCI_EXCHANGE>
Content-Type: text/plain; charset="iso-8859-1"

Hi Steven

This seems to solve your data aggregation problem:

COMPUTE cost1=cost*(int=1).
COMPUTE cost2=cost*(int=2).

AGGREGATE
  /OUTFILE=*
  /BREAK=personid
  /precost = SUM (cost1) /postcost = SUM (cost2) .

Raynald Levesque rlevesque@videotron.ca


> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of
> Steven Delaronde
> Sent: Thursday, December 28, 2000 11:54 AM
> To: SPSSX-L@LISTSERV.UGA.EDU
> Subject: Aggregating data in SPSS
>
>
> I have an SPSS data set that is not arranged so that I can conduct a two
> related-samples test or dependent t-test.  Is there a way to
> aggregate data
> so that I can sum an aggregate variable only when a certain condition is
> met?  For example, I want to add an IF statement to the following syntax
> such that the new variables (precost and postcost) are equal to the sum of
> the cost for only those cases that meet a specific condition (i.e., before
> the intervention (int=1) and after the intervention (int=2),
> respectively).
>
> AGGREGATE
>   /OUTFILE='F:\Aggr.sav
>   /BREAK=personid
>   /precost = SUM (cost) /postcost = SUM (cost) .
>
> After I rearrange the data set so that I have a precost and postcost
> variable, I can compare the two costs using a dependent t-test.  After
> making this bivariate comparison, I'm not sure if SPSS has the capacity to
> perform multivariate analysis using related samples.  Any
> suggestions would
> be greatly appreciated.
>
> Steve Delaronde
> ConnectiCare, Inc.
> Farmington, Connecticut, USA
> E-mail: delaronde@connecticare.com