We still want to use the **mpg** data file and there is no binary variable in that data set. So we need to combine the variable about the number of cylinders (**cyl**) into two groups: <6 or >= 6. The new variable **cyl_two** will be coded as 1 is cyl >= 6 or 0 if cyl < 6. **Cyl_two** will be used as outcome variable and use highway mpg (**hwy**) as predictor variable. The **expb** option in the **model** statement will give you the odds ratios.

data mpg; set mpg; if cyl >= 6 then cyl_two = 1; else cyl_two = 0; run; proc logistic data = mpg desc; model cyl_two = hwy / expb; run;

The results indicate that highway mpg (**hwy**) is a statistically significant predictor of the number of cylinders (i.e., being >= 6), Wald chi-square =46.9416, p = <.0001.

*See Also:*