Date: Fri, 18 Mar 2011 15:56:12 -0500
Reply-To: Anthony Babinec <tbabinec@sbcglobal.net>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Anthony Babinec <tbabinec@sbcglobal.net>
Subject: Re: Overriding Automatic Assignment of Measurement Level to New
Variables?
In-Reply-To: <F1897D1F26C0DA47B9489E2EA692DA7D6C715C0541@MBXC.exchange.cornell.edu>
Content-Type: multipart/alternative;
Look up the VARIABLE LEVEL command in help.
Tony Babinec
<mailto:tbabinec@sbcglobal.net> tbabinec@sbcglobal.net
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
Daniel J. Robertson
Sent: Friday, March 18, 2011 2:13 PM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Overriding Automatic Assignment of Measurement Level to New
Variables?
When SPSS computes a new variable, it assigns a measurement level to that
variable based on a set of conditions which are described in Help (I'm on
19, but I believe this automatic assignment has been in place for several
releases). However, we've noticed that SPSS often classifies new variables
as Nominal when, based on the documented conditions, they seem like they
should be Continuous.
For example, with Employee Data.sav open, if I run the command "compute
varNew = sum(salary, salbegin)", the resulting variable is assigned a
measurement level of Nominal, even though both of the source variables are
Continuous and the new variable appears to meet the documented conditions
for being automatically assigned a Continuous measurement level. One of my
colleagues thinks this behavior changed with v19. However, I can't find any
documentation of a change in v19, and I can't find anything in the
Knowledgebase on the support site to suggest a bug (resolution 59054 is
similar but relates to importing Excel files, and this is occurring with SAV
files).
So, assuming this is behaving as designed, is there a way to override this
automatic assignment of measurement level, and tell SPSS to set all new
variables to Continuous by default? I've tried in Options setting the unique
value variable under Reading External Data to '1', but it doesn't appear to
be making a difference. I suppose I could get in the habit of explicitly
setting the measurement level of all imported and computed variables, which
in some ways might be good practice, but it would be nice if there were a
simpler solution.
Thanks,
Dan
--
Daniel Robertson
Senior Research and Planning Associate
Institutional Research and Planning
Cornell University / irp.cornell.edu
[text/html]
|