An odds ratio in an ordinal response model is interpreted the same as in a binary model — it gives the change in odds for a unit increase in a continuous predictor or when changing levels of a categorical (CLASS) predictor. For the proportional odds model that PROC LOGISTIC fits to an ordinal response, this interpretation applies regardless of how the ordinal response is dichotomized into two levels. For example, suppose the response has levels 0, 1, or 2 and that the odds ratio estimate for a continuous predictor is 2. Then, by default^Note, this means that the odds of a lower response increase by a factor of two when increasing the continuous predictor by one unit. This interpretation applies whether you divide the response into low and high levels between the 0 and 1 levels (that is, contrasting 0 vs. 1 and 2) or between the 1 and 2 levels (that is, contrasting 0 and 1 vs. 2).

See the example titled "Ordinal Logistic Regression" in the LOGISTIC documentation which fits a logistic model to a taste rating response measured on an ordinal scale. The example interprets the parameters and odds ratios of a categorical predictor in the fitted model.

Note that the proportional odds model that LOGISTIC fits to ordinal data may not be supported by the data. A test of the proportional odds assumption is provided in the LOGISTIC results.

_____

NOTE: By default, LOGISTIC models the lower response levels so the odds are for a lower response. See the Response Profile table in the LOGISTIC results to see the ordering of the response levels. You can change the level ordering by using one or both of the response variable options ORDER= and DESCENDING in the MODEL statement.

Operating System and Release Information

Product Family	Product	System	SAS Release
			Reported	Fixed*
SAS System	SAS/STAT	All	n/a

* For software releases that are not yet generally available, the Fixed Release is the software release in which the problem is planned to be fixed.

---

Type:	Usage Note
Priority:	low
Topic:	SAS Reference ==> Procedures ==> LOGISTIC Analytics ==> Categorical Data Analysis Analytics ==> Regression

Date Modified:	2008-07-01 10:28:10
Date Created:	2005-07-22 15:48:31