Table 2.

Binary logit and heteroscedastic logit model results for single motorcycle crashes.

Variable	Binary Logit Model		Mixed Logit Model
	Coefficient Estimate	p-Value	Coefficient Estimate (Standard Deviation)	p-Value
Intercept	−7.453	0.0000	−7.620	<0.001
Crash and environment characteristics
Midnight (00–06)	0.742	<0.001	0.774	<0.001
Rural	0.139	0.0191	0.130	0.0395
Speed limit > 50 km/h	0.479	<0.001	0.514	<0.001
Horizontal or vertical curve	0.554	<0.001	0.591	<0.001
Lane separation-markings	0.251	<0.001	0.266	<0.001
Lane separation-traffic Island	0.473	<0.001	0.489	<0.001
Pavement surface-dry	0.188	0.0249	0.183	0.0368
Crash type
Run off	1.630	<0.001	1.698	<0.001
Hit tree/pole	2.389	<0.001	2.508	<0.001
Hit traffic device	2.156	<0.001	1.897 (1.106)	<0.001 (0.0392)
Hit island/barrier	1.684	<0.001	1.759	<0.001
Motorcyclist characteristics
Male	0.607	<0.001	0.624	<0.001
Young (<25 years old)	0.310	<0.001	0.321	<0.001
Aged (55 years old+)	0.877	<0.001	0.623 (0.902)	<0.001 (<0.001)
Without wearing a helmet	1.534	<0.001	1.719	<0.001
Without a valid driving license	0.445	<0.001	0.487	<0.001
Distracted	0.214	<0.001	0.220	<0.001
Speed violation	0.868	<0.001	0.900	<0.001
LHM (250 cc+)	0.732	<0.001	0.019 (1.554)	0.9695 (n.s.) (0.0021)
Motorcyclist’s BAC (+<0.3%)	1.827	<0.001	1.105 (1.696)	0.0148 (<0.001)
Motorcyclist’s BAC (≥0.3%)	0.828	<0.001	0.288 (1.228)	0.2640 (n.s.) (<0.001)
Motorcyclist’s BAC unknown	3.152	<0.001	3.252	0.0392
Model statistics
LL0	−8472.6		−8472.6
LL	−6052.7		−6041.1
R2	0.286		0.287
AIC	12,151		12,138

Note: n.s.: not significant at the 95% confidence interval.