https://communities.sas.com/t5/SAS-Programming/CMH-P-Value/m-p/557597#M155458 <P>It depends. Read the documentation for these statistics. If your variables TRT01PN and ACCGR0FL are both ordinal, you should go for the "Stratified Spearman"&nbsp; correlation labelled "Nonzero Correlation". If either variable is nominal,, go for the "General Association" measure.</P> Thu, 09 May 2019 20:00:26 GMT PGStats 2019-05-09T20:00:26Z https://communities.sas.com/t5/SAS-Programming/CMH-P-Value/m-p/557575#M155455 <P>Hi all,&nbsp;</P><P>&nbsp;</P><P>I want to take "Stratified Cochran-Mantel-Haenszel Test P-value" from proc freq. So I am using below procedure for that.</P><P>&nbsp;</P><PRE><CODE class=" language-sas">ods trace on; ods output CMH=cmhpval; proc freq data=xxx; tables ACCRES*TRT01PN*ACCGR0FL / norow nocum cmh; run; ods output close; ods trace off;</CODE></PRE><P>&nbsp;</P><P>So I am getting below output.</P><P>&nbsp;</P><P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="cmhpval.PNG" style="width: 600px;"><img src="https://communities.sas.com/t5/image/serverpage/image-id/29384iB4B20B47660DCC0C/image-size/large?v=v2&amp;px=999" role="button" title="cmhpval.PNG" alt="cmhpval.PNG" /></span></P><P>So which value I need to choose from these three?</P><P>&nbsp;</P><P>Thanks,</P><P>Adithya</P> Thu, 09 May 2019 19:07:46 GMT https://communities.sas.com/t5/SAS-Programming/CMH-P-Value/m-p/557575#M155455 chinna0369 2019-05-09T19:07:46Z https://communities.sas.com/t5/SAS-Programming/CMH-P-Value/m-p/557597#M155458 <P>It depends. Read the documentation for these statistics. If your variables TRT01PN and ACCGR0FL are both ordinal, you should go for the "Stratified Spearman"&nbsp; correlation labelled "Nonzero Correlation". If either variable is nominal,, go for the "General Association" measure.</P> Thu, 09 May 2019 20:00:26 GMT https://communities.sas.com/t5/SAS-Programming/CMH-P-Value/m-p/557597#M155458 PGStats 2019-05-09T20:00:26Z