Date:         Wed, 24 Dec 2003 11:37:14 -0500
Reply-To:     mwhitcomb@WESLEYAN.EDU
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
Comments:     To: Jonathan Goldberg <jlgoldberg@BRICK.NET>
From:         Michael Whitcomb <mwhitcomb@WESLEYAN.EDU>
Subject:      Re: Proc Tabulate and chisq test
Comments: To: SAS-L@LISTSERV.VT.EDU
In-Reply-To:  <5e61e30a.0312231234.2c7097cb@posting.google.com>
Content-Type: text/plain

.. but could you do two steps:

1) do a PROC FREQ with CHISQ to an ODS output,
2) merge this dat file back with your original dataset and incorporate the
chisq stats into PROC
TABULATE??


Quoting Jonathan Goldberg <jlgoldberg@BRICK.NET>:

> Ta:
>
> You haven't found it because it's not there.   Tabulate is basically a
> counter/formatter, as its name implies.  It has no statistical
> analysis capabilities (unless you count computing proportions, using
> the ever-popular denomonator definition).
>
> Cheers,
>
> Jonathan
>
>
> tliu@U.WASHINGTON.EDU (Ta Liu) wrote in message
> news:<Pine.A41.4.58.0312221123060.35128@dante61.u.washington.edu>...
> > Dear SAS-L,
> >
> > Does anybody know how to do a chisq test in Proc Tabulate? It really
> > surprised me since I could not find any reference to it and doing it
> in
> > Proc Freq is straightforward.
> >
> > Thanks in advance.
> >
> > -Ta
> >
> >
> >
> >
> >
> > ----------------------------------
> > Ta Liu
> > Postdoctoral Research Associate
> > Center for Innovation and Research
> > in  Graduate Education (CIRGE)
> > UNIVERSITY OF WASHINGTON
> >
> > M319 Miller
> > Box 353600
> > Seattle, Washington, 98195-3600
> > Tel: (206) 616-6364
> > Fax (206) 616-6762
>