Exploring the Application of the Linear Poisson Autoregressive Model for Analyzing the Dynamic Impact of Traffic Laws on Fatal Traffic Accident Frequency

Table 2

Modelling results for different orders of PAR models.

Model	PAR (1)	PAR (2)	PAR (3)	PAR (4)	PAR (5)
Intercept	−6.216 (0.375)	−6.821 (0.328)	−6.811 (0.322)	−7.028 (0.341)	−6.787 (NA)
Gasoline price	−0.152 (0.038)	−0.074 (0.028)	−0.074 (0.029)	−0.101 (0.037)	−0.124 (0.048)
Beer consumption	2.12 (0.244)	2.37 (0.204)	2.363 (0.197)	2.476 (0.209)	2.491 (0.225)
Belt law	−0.196 (0.046)	−0.157 (0.037)	−0.159 (0.038)	−0.118 (0.042)	−0.093 (0.049)
DUI toughened penalties	−0.379 (0.05)	−0.279 (0.043)	−0.278 (0.05)	−0.206 (0.076)	−0.451 (NA)
Red running	0.393 (0.083)	0.301 (0.059)	0.296 (0.065)	0.383 (0.089)	0.434 (0.113)
Alcohol law	−0.237 (0.067)	−0.219 (0.049)	−0.216 (0.052)	−0.247 (0.062)	−0.276 (0.075)
	0.628 (0.057)	0.706 (0.093)	0.704 (0.095)	0.754 (0.095)	0.718 (0.098)
	—	−0.194 (0.079)	−0.179 (0.119)	−0.218 (0.12)	−0.157 (0.127)
	—	—	−0.015 (0.086)	−0.265 (0.118)	−0.262 (0.12)
	—	—	—	0.287 (0.082)	0.144 (0.12)
	—	—	—	—	0.147 (0.09)
N^1	42	42	42	42	42
Log-likelihood	−234.125	−222.116	−217.543	−200.139	−197.746
AIC	482.249	460.232	453.069	420.277	417.492
BIC	494.414	474.133	468.725	437.655	436.606

¹N is the number of samples.