```Date: Wed, 9 Apr 2003 09:56:10 -0700 Reply-To: Dale McLerran Sender: "SAS(r) Discussion" From: Dale McLerran Subject: Re: test for proportions In-Reply-To: <8kd89vsbncq9ihkb8cnblqtcov1qqeknu3@4ax.com> Content-Type: text/plain; charset=us-ascii In addition to testing each row separately, you can sum the chi-square values for each row and test this against the chi-square distribution with df=5*9. I would look at this pooled test first, and if it is significant then examine the row-specific chi-square tests. If the pooled test is not significant, do not interpret the row-specific tests. You are introducing a multiple comparisons problem with the multiple row-specific tests. If you examine row-specific tests conditional on the the pooled test, you effectively control the experimentwise error rate. Dale --- Jay Weedon wrote: > On 9 Apr 03 00:13:56 GMT, ctsdias@CARPA.CIAGRI.USP.BR (Carlos Tadeu) > wrote: > > >I am working in an assessorship in a problem of contingency table > 5x10 and the researcher wants to know if there are difference among > proportions inside a level of the first factor. I would like to know > if there is a test for this in the SAS and where I can find > references about. This sound like multiple comparison of proportions > in a contingency table that has one row and 10 columns. > > > > A B C D E F G H I J K total > >M p1 p2 p3 p4 p5 p6 p7 p8 p9 p10 N > > > >Where pi=Ni/N, AND sum(Ni)=N > > Generate one-way frequency tables and request the chi-square test, > e.g., > > proc freq; > tables column /chisq; > by row; > run; > > This will generate (separately for each row) a chi-square > goodness-of-fit test of the null hypothesis that the row proportions > are all equal. > > JW ===== --------------------------------------- Dale McLerran Fred Hutchinson Cancer Research Center mailto: dmclerra@fhcrc.org Ph: (206) 667-2926 Fax: (206) 667-5977 --------------------------------------- __________________________________________________ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com ```

Back to: Top of message | Previous page | Main SAS-L page