Date:         Thu, 9 Jan 1997 17:26:46 -0800
Reply-To:     John Taenzler <DATATACTICS-RESEARCH@WORLDNET.ATT.NET>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         John Taenzler <DATATACTICS-RESEARCH@WORLDNET.ATT.NET>
Organization: DataTactics, Inc
Subject:      Matching Cases with LOOPs
Content-Type: text/plain; charset=us-ascii

I have a problem . . .
 
I have a data set with two sets of subjects (A & B).  Each set has
identical numbers and locations of variables. (i.e., set a is in records
1-50 and set B is in records 51-100).
 
The problem is this, how can I set up a series of conditional loop
commands so that I can detemine which cases from set B best match a
given case from set A on 4 different variables?
 
But, wait, there's more.
 
The variables on which matches are determined are weighted, so that if
case A1 id being compared with case B1, if they have identical values
for  variables W & X but different values for varaibles Y & Z, they are
a better match than if they have differnt values for W & X, but have
identical values for Y & Z!
 
Essentually, I'm attempting to creat a post-hoc blocking design, and I
need to pair up cases (A-B) across a series of weighted variables.  Is
there anyway, with SPSS Command language to write a series of commands
so that for every case A, an X number of B's will be coded as being
possible matches?
 
See how much fun I'm having!
 
Thanks in advance.