Date: Mon, 7 Apr 1997 23:38:34 -0300
Reply-To: Carlos Tadeu <ctsdias@CARPA.CIAGRI.USP.BR>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Carlos Tadeu <ctsdias@CARPA.CIAGRI.USP.BR>
Subject: Re: test for proportions
Content-Type: text/plain; charset="iso-8859-1"
Thanks to Jim, Dale Jay for yours email.
I did overal chi-square test on the whole contingency table and it was significative.
I want to do test with only one marginal variable because the researcher wants to know if there are differences
among proportions of the levels of the second variable in each level of the first. This sound like pairwais comparations.
I tried to run separately for each row as suggested by Jay, but in the output we have just
Column Frequency Percent Cumulative Frequency Cumulative Percent
with the correctly calculated table, but no chi-square test.
In this mean time I found the Marascuillo procedure in the site: http://www.itl.nist.gov/div898/handbook/prc/section4/prc474.htm
and I would like to know if someone has worked with this test? References? It sound what I look for but I am not sure about this test.