Insights on Crash Injury Severity Control from Novice and Experienced Drivers: A Bivariate Random-Effects Probit Analysis

Table 1

Results for bivariate probit and bivariate random-effects probit models.

Variable	Bivariate probit				Bivariate random-effects probit
	Coef.	Std. Err.	z	> |z|	Coef.	Std. Err.	z	> |z|
Novice driver injury model
Crash type	−0.109	0.020	−5.49	0.000	−0.110	0.019	−5.01	0.000
Vehicle 2 type	−0.050	0.022	−2.25	0.024	−0.049	0.024	−2.22	0.012
Pedestrian	1.329	0.365	3.63	0.000	1.394	0.384	3.61	0.000
Motorcyclist	1.513	0.436	3.46	0.001	1.671	0.490	3.44	0.001
Constant	0.887	0.079	11.15	0.000	0.882	0.080	11.17	0.000
Experienced driver injury model
Crash type	−0.055	0.021	−2.61	0.009	−0.056	0.020	−2.61	0.005
Vehicle 2 driver condition	0.056	0.016	3.52	0.000	0.057	0.016	3.57	0.000
First harm	−0.106	0.013	−7.81	0.000	−0.106	0.014	−7.84	0.000
Highway factor	−0.616	0.182	−3.37	0.001	−0.629	0.174	−3.38	0.001
Constant	1.668	0.195	8.52	0.000	1.681	0.191	8.51	0.000
α1	1.052	0.189	6.82	0.000	1.112	0.155	7.18	0.000
α2	0.908	0.263	3.65	0.000	0.948	0.220	4.31	0.000
ρα	−0.256	0.087			−0.287	0.085
ρɛ					0.416	0.086
Goodness of fit
No. of observations	1999				1999
Log-likelihood at convergence	−925.747				−978.069
Log-likelihood at zero	−1809.635				−1894.337

Note. Coef. = coefficient; Std. Err. = standard error; ∗significant at the 5% level.