A comparative evaluation of the safety performance of bus priority route configurations

European Transport Research Review

An Open Access Journal

Table 2 Poisson regression models fitted to total injury accidents on road sections with BPRs

(a) Model 1—no use of traffic volumes
Variables	Estimate	Standard error	DF	t Value	Pr >\|t\|
Intercept	− 1.53	1.08	87	− 1.41	0.162
BPR type: curbside	2.56	1.06	264	2.41	0.017
BPR type: open center-lane	1.00	1.08	264	0.93	0.355
BPR type: barrier-separated central	0
Area: urban, 2–3 lanes	− 1.55	0.73	264	− 2.11	0.036
Area: suburban	− 0.43	0.73	264	− 0.59	0.554
Area: urban, 1 lane	0

(b) Model 2—with continuous traffic volumes
Variables	Estimate	Standard error	DF	t Value	Pr >\|t\|
Intercept	− 1.76	1.11	85	− 1.58	0.117
BPR type: curbside	2.49	1.07	264	2.34	0.020
BPR type: open center-lane	1.03	1.08	264	0.96	0.340
BPR type: barrier-separated central	0
Area: urban, 2–3 lanes	− 1.74	0.77	264	− 2.26	0.025
Area: suburban	− 0.79	0.82	264	− 0.96	0.337
Area: urban, 1 lane	0
log (V_Bus)	− 0.23	0.38	264	− 0.60	0.547
log (V_Veh)	0.44	0.48	264	0.92	0.360

(c) Model 3—with categorical traffic volumes
Variables	Estimate	Standard error	DF	t Value	Pr >\|t\|
Intercept	− 1.11	1.53	81	− 0.73	0.470
BPR type: curbside	2.45	1.06	261	2.31	0.022
BPR type: open center-lane	0.93	1.08	261	0.87	0.387
BPR type: barrier-separated central	0
Area: Urban, 2–3 lanes	− 2.00	1.14	261	− 1.76	0.080
Area: Suburban	− 1.10	1.17	261	− 0.94	0.347
Area: Urban, 1 lane	0
Year: 1	0.66	0.29	261	2.31	0.022
Year: 2	0.51	0.29	261	1.72	0.087
Year: 3	− 0.24	0.35	261	− 0.69	0.494
Year: 4	0
Cat_veh: 1	− 0.96	1.11	261	− 0.86	0.391
Cat_veh: 2	− 0.41	0.57	261	− 0.72	0.471
Cat_veh: 3	− 0.33	0.45	261	− 0.73	0.465
Cat_veh: 4	0
Cat_bus: 1	− 0.18	0.44	261	− 0.40	0.687
Cat_bus: 2	0.45	0.40	261	1.14	0.254
Cat_bus: 3	− 0.10	0.37	261	− 0.28	0.776
Cat_bus: 4	0

(a) Model fit statistics: − 2 Log Likelihood = 361.95; AIC = 373.95; R² = 15.1%. Likelihood ratio test: p < 0.0001. Type III Tests of fixed effects: for BRT type p < 0.01; for Area p < 0.0001.
(b) Model fit statistics: − 2 Log Likelihood = 361.18; AIC = 375.18; R² = 15.2%. Likelihood ratio test: p < 0.0001. Type III Tests of fixed effects: for BRT type p < 0.05; for Area p < 0.001. log—logarithm.
(c) Model fit statistics: − 2 Log Likelihood = 342.92; AIC = 370.92; R² = 19.5%. Likelihood ratio test: p < 0.0005. Type III Tests of fixed effects: for BRT type, Area, Year p < 0.05; for Cat_veh p = 0.77, for Cat_bus p = 0.11.