Date: Mon, 26 Jun 2006 13:02:20 -0700 "Nordlund, Dan (DSHS)" "SAS(r) Discussion" "Nordlund, Dan (DSHS)" Re: Odds Ratios To: Brad Heins text/plain; charset=utf-8

> -----Original Message----- > From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Brad > Heins > Sent: Monday, June 26, 2006 11:57 AM > To: SAS-L@LISTSERV.UGA.EDU > Subject: Odds Ratios > > I am having a little trouble interpreting the odds ratios for a research > paper that I am currently writing. > > D30 is survival to 30 days. The data is coded as either a 1 (survived to > 30 > days) or 0 (culled before 30 days). > > I am trying to determine how to interpret the odds ratios below. If the > estimate for Scandinavian vs. Holstein in 3.061, would that say that the > odds of survival for Scandinavian is 3.061 times greater than Holstein? > or > are the odds ratios the other way around? > > Below is some of the output from the logistic regression. > > Any help would be appreciated? Thank you. > > > Brad Heins > > > > > > Response Variable D30 > Number of Response Levels 2 > Model binary logit > Optimization Technique Fisher's scoring > > Number of Observations Used 2246 > > Response Profile > > Ordered Total > Value D30 Frequency > 1 0 2191 > 2 1 55 > > Probability modeled is D30=0. > > > Odds Ratio Estimates > Point 95% Wald > Effect Estimate Confidence Limits > > > BREED SCANDINAVIAN vs HOLSTEIN 3.061 1.063 8.814 > BREED NORMANDE vs HOLSTEIN 2.649 1.026 6.843 > BREED MONTBELIARDE vs HOLSTEIN 3.175 1.124 8.973

When you say D30 was coded "0 (culled before 30 days)", what does that mean? Was it the case that all cows coded 0 failed to survive for 30 days, or were some cows censored before 30 days? This may (or may not) have a bearing won whether you should be doing something different from logistic regression.

As to the interpretation of your odds ratios, 1. you have apparently coded your data to compare each of 3 other breeds to Holstein, and 2. you are modeling the odds of being coded 0.

So, the odds of being coded 0 (i.e. failed to survive 30 days) is 3.061 times higher for Scandinavian than for Holstein (i.e. Scandinavian are less likely to survive).

Hope this helps,

Dan

Daniel J. Nordlund Research and Data Analysis Washington State Department of Social and Health Services Olympia, WA 98504-5204

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