Date:         Tue, 11 Apr 2006 15:17:15 +1000
Reply-To:     Jason Burke <burke.jasonm@gmail.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Jason Burke <burke.jasonm@gmail.com>
Subject:      Re: Order of labels in Custom Table Columns
Comments: To: Mike Trotman <mike.trotman@datalucid.com>
In-Reply-To:  <443B05D4.1060308@datalucid.com>
Content-Type: text/plain; charset=ISO-8859-1

Yes there is.

Move the cluster membership variable to the column dimension in your
table, then select "Categories and Totals" from the lower right hand
side of the Custom Tables UI. Here you will see (assuming that each
unique value in you variable has a label) and list box which will
enable you to specify the order of the values for that dimension.

Or, with the CATEGORIES VARIABLES subcommand you can specify it using
syntax (no need for value labels). Say, for a variable clust_mem

   /CATEGORIES VARIABLES=clust_mem [5,2,3,4,1]

Regards,


Jason

On 4/11/06, Mike Trotman <mike.trotman@datalucid.com> wrote:
> I have nominal variables (cluster membership) whose categories I want to
> use as columns in a table.
>
> I want these columns to be in a specific order (based on summary values
> calculated elsewhere - e.g. Average of a variable for the cluster).
> I don't want to change the values for the cluster membership variable.
>
> Is there any way to accomplish this without
> 1 - defining new variables with new values
> or
> 2 - adding a sort order to the beginning of the values labels (and
> sorting on value label)?
>
> I would prefer to use Custom Tables - but would consider other methods.
>
> I have (optimistically) tried declaring the value label pairs in the
> order in which I want them displayed - to no effect.
>
>
> Any suggestions appreciated.
>
> Thanks
>
> Mike
>
> --
> Datalucid Limited
> email: mike.trotman@datalucid.com
> web: http://www.datalucid.com
>
>
> Message Scanned by ClamAV on datalucid.com
>