For logistic models, the odds ratio is another way to represent a model’s coefficients. The odds ratio is calculated as exp(β). In terms of the odds ratio, a value of 1 corresponds to no effect. Variants of the odds ratio are known as the incidental-risk ratio (for count models), the relative risk ratio (for multinomial logit models), and the hazard ratio (for proportional hazards models).
To view a model’s odds ratios rather than the original coefficients:
The standard errors and confidence intervals in the explanatory variables table will then apply to the odds ratios, rather than to the unexponentiated coefficients. The Predictor view will also display its calculation in terms of odds ratios.
See also: