Development of Accident Frequency Models with Random Parameters on Interstate Roadway Segments with and without Lighting Systems
Table 4
Model results for segments without lighting systems.
Variable
Constant parameter
Random parameter
Mean
t-statistic
Distribution
Mean
t-statistic
Standard deviation
t-statistic
Constant
−7.055
−36.444
n.a.
Exposure
Logarithm of averaged daily traffic volume
n.a.
Normal
0.076
3.577
0.074
10.204
Logarithm of segment length
n.a.
Normal
0.042
2.320
0.179
16.938
Lanes
Proportion of segments with standard number of lanes
4.162
5.006
n.a.
Proportion of segments with low number of lanes
n.a.
Normal
0.011
5.393
0.185
11.805
Left shoulder width
Proportion of segments with low left shoulder width
n.a.
Normal
0.035
3.239
0.399
46.452
Proportion of segments with standard left shoulder width
n.a.
Normal
−0.085
−15.047
0.021
6.559
Right shoulder width
Proportion of segments with standard left shoulder width
n.a.
Normal
−0.059
−8.694
0.033
22.017
Proportion of segments with high right shoulder width
−0.057
−8.004
n.a.
Proportion of segments with low right shoulder width
n.a.
Normal
0.043
3.358
0.398
102.567
Curves
Number of horizontal curves in segment
0.031
7.776
n.a.
Number of vertical curves in segment
0.019
1.833
n.a.
Scale parameter for overdispersion
5.228
23.788
n.a.
Number of observations: 385. McFadden’s pseudo : 0.60. Log likelihood at convergence of RPNB model: −9,277.67. Log likelihood at the constant only: −23,444.89. n.a.: not applicable.
