Date:         Fri, 9 Sep 2005 20:49:13 -0400
Reply-To:     Richard Ristow <wrristow@mindspring.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Richard Ristow <wrristow@mindspring.com>
Subject:      Re: Combining binary columns into a single column
Comments: To: Margaret MacDougall <margmacd@talk21.com>
In-Reply-To:  <20050909085313.66251.qmail@web86607.mail.ukl.yahoo.com>
Content-Type: text/plain; charset="us-ascii"; format=flowed

(I've returned to the original subject line so my E-mail program will
'thread' the exchange correctly.)

At 04:53 AM 9/9/2005, Margaret MacDougall wrote:

>Yesterday, I posted the query, [how to] code the pairs forthcoming
>from two
>'binary columns' into a single column after the following fashion:
>
>0,0-->1
>0,1-->2
>1,0-->3
>1,1-->4
>1,blank-->blank
>blank,1-->blank
>blank,blank-->blank

It looks like everybody understood what you wanted to do. Nevertheless,
it's better not to speak of a 'blank' value for a numeric variable.
Numeric quantities cannot be 'blank'. The value most analogous is
"system-missing", SYSMIS for short, which is probably what your
"blanks" are.

>I noticed that I could define a function in SPSS using the Compute
>option from the transorm menu which solves my problem.  Perhaps other
>users will find this useful and so I offer the solution below:
>
>DEFINE THE FUNCTION 'COMPOSITE' WITH VALUES GIVEN BY:
>COMPOSITE(c1CELL,c2CELL)=2^c1CELL*3^c2CELL,
>
>where c1CELL and c2CELL are corresponding entries in the first and
>second column, respectively. [THIS MAPS THE ORIGINAL PAIRS TO 1,3,2,6,
>RATHER THAN 1,2,3,4 BUT THIS GIVES A 4-VALUED COLUMN WHICH SUFFICES
>FOR MY NEEDS.]

Good. But if you wanted the exact coding you specified, then try
(though this is untested),

COMPOSITE(c1CELL,c2CELL)=2*c1CELL + c2CELL + 1.