An example of the full logistic regression model used to compare the proportion of crashes where the response variable was one of the driving restrictions (night-time curfew in this example) and the main explanatory variable was type of licence (in this example pre-GDL drivers were compared with restricted licence drivers)
| Estimates | df | Estimate | SE | χ2 | p Value>χ2 | Odds ratio | 95% confidence interval |
|---|---|---|---|---|---|---|---|
| Intercept | 1 | 1.809 | 1.504 | 1.45 | 0.229 | ||
| Licence type | |||||||
| Pre-GDL: restricted | 1 | −0.42 | 0.14 | 9.13 | 0.003 | 0.66 | 0.50 to 0.86 |
| Driver age (years) | |||||||
| 19 : 15 | 1 | 0.786 | 0.120 | 15.5 | <0.0001 | 2.20 | 1.50 to 3.29 |
| 19 : 16 | 1 | 0.606 | 0.124 | 23.7 | <0.0001 | 1.83 | 1.44 to 2.34 |
| 19 : 17 | 1 | 0.221 | 0.106 | 4.34 | 0.037 | 1.25 | 1.01 to 1.54 |
| 19 : 18 | 1 | 0.041 | 0.098 | 0.18 | 0.676 | 1.04 | 0.86 to 1.26 |
| Gender | |||||||
| Female: male | 1 | 0.463 | 0.095 | 23.55 | <0.0001 | 1.59 | 1.32 to 1.92 |
| Year of crash | 1 | −0.013 | 0.016 | 0.65 | 0.422 | 0.99 | 0.96 to 1.02 |