https://communities.sas.com/t5/Statistical-Procedures/Odds-Ratios-do-not-match-the-exponentiated-log-odds/m-p/53321#M2466 Hopefully this is a silly question with an easy fix. However, my results from a binary logistic regression is providing me odds ratio estimates that do not equate to the exponentiated coefficient estimates. Does SAS calculate Odds Ratios in some other fashion? I have negative estimates, which I assumed would relate to an odds ratio under 1, but this is not always the case. That and the numbers just don't line up. <BR /> <BR /> Thanks for any help,<BR /> Peter Fri, 05 Feb 2010 18:24:27 GMT deleted_user 2010-02-05T18:24:27Z https://communities.sas.com/t5/Statistical-Procedures/Odds-Ratios-do-not-match-the-exponentiated-log-odds/m-p/53321#M2466 Hopefully this is a silly question with an easy fix. However, my results from a binary logistic regression is providing me odds ratio estimates that do not equate to the exponentiated coefficient estimates. Does SAS calculate Odds Ratios in some other fashion? I have negative estimates, which I assumed would relate to an odds ratio under 1, but this is not always the case. That and the numbers just don't line up. <BR /> <BR /> Thanks for any help,<BR /> Peter Fri, 05 Feb 2010 18:24:27 GMT https://communities.sas.com/t5/Statistical-Procedures/Odds-Ratios-do-not-match-the-exponentiated-log-odds/m-p/53321#M2466 deleted_user 2010-02-05T18:24:27Z https://communities.sas.com/t5/Statistical-Procedures/Odds-Ratios-do-not-match-the-exponentiated-log-odds/m-p/53322#M2467 See these usage notes:<BR /> <BR /> <A href="http://support.sas.com/kb/23087" target="_blank">http://support.sas.com/kb/23087</A><BR /> <A href="http://support.sas.com/kb/36478" target="_blank">http://support.sas.com/kb/36478</A><BR /> <BR /> The answers to many questions can be found in the Samples and SAS Notes in the searchable knowledgebase, <A href="http://support.sas.com/kb" target="_blank">http://support.sas.com/kb</A>. You can use the search engine there to find the answers you need. Fri, 05 Feb 2010 21:03:43 GMT https://communities.sas.com/t5/Statistical-Procedures/Odds-Ratios-do-not-match-the-exponentiated-log-odds/m-p/53322#M2467 StatDave 2010-02-05T21:03:43Z