1 Introduction
1.1 Research questions
-
systematically explore the crash performances of both urban segments and intersections, with the related influential variables thus searching for the best subsets of segments and intersections with homogeneous characteristics for modelling purposes, among different possibilities.
-
Compare the significant predictors highlighted in the modelling stages with the significant crash predictors retrieved in previous research, to reveal specific local differences which may be of interest for further studies.
-
Explore the influence of several other potential crash predictors, which are usually not considered in safety prediction studies, besides the traditional geometric and traffic control variables used in previous research.
2 Methods
2.1 Main dataset
2.2 Samples of sites
2.3 Crash predictors
Variables [for segments] | Modalities | Homogeneous Segments (n = 320a) | One-way homogeneous segments (n = 190) | Two-way homogeneous segments (n = 130) | |||
Mean (st. dev.)/ Counts | Min-max/ Percentages | Mean (st. dev.)/ Counts | Min-max/ Percentages | Mean (st. dev.)/ Counts | Min-max/ Percentages | ||
Fatal+Injury (F + I) crashes | Numerical | 689 | – | 283 | – | 406 | – |
F + I crashes/segment/year | Numerical | 0.43 (0.86) | 0–6 | 0.30 (0.68) | 0–5 | 0.63 (1.03) | 0–6 |
One-way | 0 – No (two-way) | 0–130 | 0–0.41 | 0–0 | 0–0.00 | 0–130 | 0–1.00 |
1 – Yes | 1–190 | 1–0.59 | 1–190 | 1–1.00 | 1–0 | 1–0.00 | |
One-lane | 0 –No (multilane) | 0–226 | 0–0.71 | 0–96 | 0–0.51 | 0–130 | 0–1.00 |
1 – Yes | 1–94 | 1–0.29 | 1–94 | 1–0.49 | 1–0 | 1–0.00 | |
AADT | Numerical | 8637.2 (5889.3) | 300.0–29,960.0 | 6228.8 (4569.5) | 300.0–22,600.0 | 12,157.1 (5843.5) | 1160.0–29,960.0 |
Length (m) | Numerical | 138.3 (183.3) | 30.0–1510.0 | 124.1 (175.4) | 30.0–1510.0 | 159.1 (193.0) | 30.0–1415.0 |
Speed Limit = 50 km/h | 0 – Yes | 0–309 | 0–0.97 | 0–190 | 0–1.00 | 0–119 | 0–0.92 |
1 – No (> 50 km/h) | 1–11 | 1–0.03 | 1–0 | 1–0.00 | 1–11 | 1–0.08 | |
Road width (m) | Numerical | 10.5 (3.0) | 4.1–28.0 | 9.9 (2.6) | 5.5–20.0 | 11.3 (3.5) | 4.1–28.0 |
Type of lanes | 0–1 lane | 0–116 | 0–0.36 | 0–116 | 0–0.61 | 0–0 | 0–0.00 |
1 – > 1 lane (up to 3) | 1–74 | 1–0.23 | 1–74 | 1–0.39 | 1–0 | 1–0.00 | |
2–1 + 1 lane | 2–89 | 2–0.28 | 2–0 | 2–0.00 | 2–89 | 2–0.68 | |
3 – > 1 + 1 lanes (up to 3 + 3) | 3–41 | 3–0.13 | 3–0 | 3–0.00 | 3–41 | 3–0.32 | |
Density of minor roads/ driveways per km | Numerical | 26.9 (26.0) | 0.0–143.0 | 22.4 (22.8) | 0.0–142.9 | 33.3 (28.9) | 0.0–143.0 |
Vertical signs on minor roads/ driveways | 0 – No | 0–256 | 0–0.80 | 0–158 | 0–0.84 | 0–98 | 0–0.75 |
1 – Yes | 1–64 | 1–0.20 | 1–32 | 1–0.16 | 1–32 | 1–0.25 | |
Poor maintenance of pavements | 0 – No | 0–71 | 0–0.22 | 0–49 | 0–0.26 | 0–22 | 0–0.17 |
1 – Yes | 1–249 | 1–0.78 | 1–141 | 1–0.74 | 1–108 | 1–0.83 | |
Road markings | 0 – No/Partially absent or illegible | 0–60 | 0–0.19 | 0–24 | 0–0.13 | 0–36 | 0–0.28 |
1 – Illegible | 1–124 | 1–0.39 | 1–82 | 1–0.43 | 1–42 | 1–0.32 | |
2 – Yes | 2–136 | 2–0.42 | 2–84 | 2–0.44 | 2–52 | 2–0.40 | |
Parking type | 0 – Prohibited | 0–85 | 0–0.27 | 0–26 | 0–0.14 | 0–59 | 0–0.45 |
1 – One side | 1–109 | 1–0.34 | 1–74 | 1–0.39 | 1–35 | 1–0.27 | |
2 – Both sides | 2–119 | 2–0.37 | 2–87 | 2–0.46 | 2–32 | 2–0.25 | |
3 – Mixed | 3–7 | 3–0.02 | 3–3 | 3–0.02 | 3–4 | 3–0.03 | |
Cycle paths | 0 – No | 0–291 | 0–0.91 | 0–163 | 0–0.86 | 0–128 | 0–0.98 |
1 – Yes | 1–29 | 1–0.09 | 1–27 | 1–0.14 | 1–2 | 1–0.02 | |
Sidewalks | 0 – No | 0–31 | 0–0.10 | 0–4 | 0–0.02 | 0–27 | 0–0.21 |
1 – Yes | 1–289 | 1–0.90 | 1–186 | 1–0.98 | 1–103 | 1–0.79 | |
Median | 0 – No | 0–244 | 0–0.76 | 0–114 | 0–0.60 | 0–130 | 0–1.00 |
1 – Yes | 1–76 | 1–0.24 | 1–76 | 1–0.40 | 1–0 | 1–0.00 | |
Bus stop | 0 – No | 0–227 | 0–0.71 | 0–143 | 0–0.75 | 0–84 | 0–0.65 |
1 – Yes | 1–93 | 1–0.29 | 1–47 | 1–0.25 | 1–46 | 1–0.35 | |
Bus/taxi lane | 0 – No | 0–284 | 0–0.89 | 0–154 | 0–0.81 | 0–130 | 0–1.00 |
1 – Yes | 1–36 | 1–0.11 | 1–36 | 1–0.19 | 1–0 | 1–0.00 | |
Land use | 0 – Residential | 0–115 | 0–0.36 | 0–71 | 0–0.37 | 0–44 | 0–0.34 |
1 – Mainly commercial | 1–150 | 1–0.47 | 1–110 | 1–0.58 | 1–40 | 1–0.31 | |
2 – Other | 2–55 | 2–0.17 | 2–9 | 2–0.05 | 2–46 | 2–0.35 | |
Variables [for intersections] | Modalities | Intersections (n = 120a) | Three-legged intersections (n = 48) | Four-legged intersections (n = 72b) | |||
Mean (st. dev.)/ Counts | Min-max/ Percentages | Mean (st. dev.)/ Counts | Min-max/ Percentages | Mean (st. dev.)/ Counts | Min-max/ Percentages | ||
Fatal+Injury (F + I) crashes | Numerical | 906 | – | 232 | – | 674 | – |
F + I crashes/intersection/year | Numerical | 1.51 (1.84) | 0–15 | 0.97 (1.41) | 0–15 | 1.87 (1.99) | 0–12 |
Traffic lights | 0 – No | 0–59 | 0–0.49 | 0–38 | 0–0.79 | 0–21 | 0–0.29 |
1 – Yes | 1–61 | 1–0.51 | 1–10 | 1–0.21 | 1–51 | 1–0.71 | |
Type = 4 legs | 0 – No | 0–48 | 0–0.40 | 0–48 | 0–1.00 | 0–0 | 0–0.00 |
1 – Yes | 1–72 | 1–0.60 | 1–0 | 1–0.00 | 1–72 | 1–1.00 | |
Main AADT | Numerical | 12,192.4 (5313.2) | 2475.0–23,680.0 | 12,588.4 (5350.8) | 3550.0–22,340.0 | 11,928.4 (5309.1) | 2475.0–23,680.0 |
Secondary AADT | Numerical | 5084.1 (4271.5) | 300.0–17,880.0 | 4008.3 (3897.4) | 300.0–18,520.0 | 5801.4 (4384.2) | 345.0–17,880.0 |
Total AADT | Numerical | 17,276.5 (7981.8) | 4175.0–41,560.0 | 16,596.6 (7018.3) | 5420.0–32,300.0 | 17,729.7 (8582.2) | 4175.0–41,560.0 |
Main AADT/ Total AADT | Numerical | 0.53 (0.03) | 0.50–0.62 | 0.55 (0.03) | 0.50–0.62 | 0.53 (0.02) | 0.50–0.61 |
Lane balance | 0 – No | 0–81 | 0–0.67 | 0–38 | 0–0.79 | 0–43 | 0–0.60 |
1 – Yes | 1–39 | 1–0.33 | 1–10 | 1–0.21 | 1–29 | 1–0.40 | |
Median on the main road | 0 – No | 0–100 | 0–0.83 | 0–48 | 0–1.00 | 0–52 | 0–0.72 |
1 – Yes | 1–20 | 1–0.17 | 1–0 | 1–0.00 | 1–20 | 1–0.28 | |
Median on the secondary road | 0 – No | 0–107 | 0–0.89 | 0–43 | 0–0.90 | 0–64 | 0–0.89 |
1 – Yes | 1–13 | 1–0.11 | 1–5 | 1–0.10 | 1–8 | 1–0.11 | |
Entering lanes on the main road | Numerical | 1.5 (1.1) | 0.0–5.0 | 1.1 (1.0) | 0.0–4.0 | 1.8 (1.1) | 0.0–5.0 |
Mean lane width at intersections (m) | Numerical | 4.5 (1.1) | 2.8 (8.7) | 4.3 (1.0) | 2.8–6.6 | 4.6 (1.2) | 2.8–8.7 |
Critical sight distance (m) | Numerical | 31.2 (26.0) | 0.0–139.5 | 30.0 (28.1) | 0.0–139.5 | 32.0 (24.6) | 1.6–100.0 |
One-way legs | Numerical | 1.7 (1.5) | 0.0–4.0 | 1.4 (1.1) | 0.0–3.0 | 2.0 (1.6) | 0.0–4.0 |
Specialized turning lane | 0 – No | 0–88 | 0–0.73 | 0–43 | 0–0.90 | 0–45 | 0–0.63 |
1 – Yes | 1–32 | 1–0.27 | 1–5 | 1–0.10 | 1–27 | 1–0.37 | |
Traffic control | 0 – No | 0–22 | 0–0.18 | 0–16 | 0–0.33 | 0–6 | 0–0.08 |
1 – Give-way/Stop | 1–37 | 1–0.31 | 1–22 | 1–0.46 | 1–15 | 1–0.21 | |
2 – Traffic lights | 2–35 | 2–0.29 | 2–7 | 2–0.15 | 2–28 | 2–0.39 | |
3 – Traffic lights with dedicated turning | 3–26 | 3–0.22 | 3–3 | 3–0.06 | 3–23 | 3–0.32 | |
Poor maintenance of pavements | 0 – No | 0–18 | 0–0.15 | 0–9 | 0–0.19 | 0–9 | 0–0.13 |
1 – Yes | 1–102 | 1–0.85 | 1–39 | 1–0.81 | 1–63 | 1–0.87 | |
Cycle path crossing | 0 – No | 0–103 | 0–0.86 | 0–36 | 0–0.75 | 0–67 | 0–0.93 |
1 – Yes | 1–17 | 1–0.14 | 1–12 | 1–0.25 | 1–5 | 1–0.07 | |
Sidewalks | 0 – No | 0–8 | 0–0.07 | 0–3 | 0–0.06 | 0–5 | 0–0.07 |
1 – Yes (both sides) | 1–112 | 1–0.93 | 1–45 | 1–0.94 | 1–67 | 1–0.93 | |
Bus stops | 0 – No | 0–111 | 0–0.93 | 0–46 | 0–0.96 | 0–65 | 0–0.90 |
1 – Yes | 1–9 | 1–0.07 | 1–2 | 1–0.04 | 1–7 | 1–0.10 | |
Bus/taxi lanes on intersecting roads | 0 – No | 0–92 | 0–0.77 | 0–43 | 0–0.90 | 0–49 | 0–0.68 |
1 – Yes | 1–28 | 1–0.23 | 1–5 | 1–0.10 | 1–23 | 1–0.32 |
2.4 Data analysis techniques
-
preliminary models for the whole datasets of segments and intersections were run;
-
two sub-categories for each family of sites (segments and intersections) were selected as based on results from preliminary models;
-
models for each sub-category were run.
3 Results
3.1 Predictive models for segments
Predictor | Estimate | Std. Error | z- value | p-value |
---|---|---|---|---|
All segments | ||||
(Intercept) | −6.054 | 0.728 | −8.318 | < 0.001 |
Ln (AADT) | 0.375 | 0.084 | 4.449 | < 0.001 |
Ln(L) | 0.292 | 0.060 | 4.907 | < 0.001 |
Type of lanes 1: > 1 (reference: 1 lane) | 0.079 | 0.167 | 0.470 | 0.638 |
Type of lanes 2: 1 + 1 (ref.: 1 lane) | 0.524 | 0.151 | 3.471 | < 0.001 |
Type of lanes 3: > 1 + 1 (ref.: 1 lane) | 0.707 | 0.181 | 3.906 | < 0.001 |
Vertical signs on driveways/minor roads (Yes - ref.: No) | 0.652 | 0.110 | 5.943 | < 0.001 |
Poor maintenance of pavements (Yes - ref.: No) | −0.268 | 0.114 | −2.348 | 0.019 |
Parking type 1: At one side (ref.: No parking) | 0.218 | 0.131 | 1.668 | 0.095 |
Parking type 2: At both sides (ref.: No parking) | 0.258 | 0.126 | 2.053 | 0.040 |
Parking type 3: Mixed (ref.: No parking) | −1.092 | 0.556 | −1.964 | 0.050 |
Cycle paths (Yes – ref.: No) | 0.308 | 0.182 | 1.694 | 0.090 |
Goodness-of-fit measures | ||||
LRT (reference: null model) = 200.96 (df = 11), p < .001; LRT (reference: full model) = 6.64 (df = 11), p = .828 | ||||
Nagelkerke R2 = 0.158, AIC = 2609.9 | ||||
Over-dispersion parameter (theta) = 1.08 (std. error: 0.17) | ||||
Interval of continuous predictors | ||||
AADT: 300–29,960 vehicles/day, L: 30–1510 m | ||||
One-way segments | ||||
(Intercept) | −8.794 | 0.961 | −9.151 | < 0.001 |
Ln (AADT) | 0.695 | 0.108 | 6.434 | < 0.001 |
Ln(L) | 0.333 | 0.095 | 3.505 | < 0.001 |
Driveways/minor roads per km | 0.010 | 0.003 | 3.163 | 0.002 |
Vertical signs on accesses/minor roads (Yes - ref.: No) | 0.477 | 0.169 | 2.825 | 0.005 |
Poor maintenance of pavements (Yes - ref.: No) | −0.497 | 0.159 | −3.137 | 0.002 |
Goodness-of-fit measures | ||||
LRT (reference: null model) = 101.68 (df = 5), p < .001; LRT (reference: full model) = 9.77 (df = 10), p = .460 | ||||
Nagelkerke R2 = 0.153, AIC = 1226.8 | ||||
Over-dispersion parameter (theta) = 1.01 (std. error: 0.26) | ||||
Interval of continuous predictors | ||||
AADT: 300–22,600 vehicles/day, L: 30–1510 m, Density of driveways/minor roads: 0.0–142.9 number/km | ||||
Two-way segments | ||||
(Intercept) | −3.764 | 1.122 | −3.356 | < 0.001 |
Ln (AADT) | 0.134 | 0.117 | 1.140 | 0.254 |
Ln(L) | 0.310 | 0.079 | 3.944 | < 0.001 |
Vertical signs on accesses/minor roads (Yes - ref.: No) | 0.432 | 0.141 | 3.074 | 0.002 |
Markings 1: Illegible (ref.: No/partially absent/illegible) | 0.317 | 0.177 | 1.791 | 0.073 |
Markings 2: Yes (ref.: No/partially absent/illegible) | 0.309 | 0.175 | 1.767 | 0.077 |
Parking type 1: At one side (ref.: No parking) | 0.287 | 0.156 | 1.836 | 0.066 |
Parking type 2: At both sides (ref.: No parking) | 0.441 | 0.152 | 2.903 | 0.004 |
Parking type 3: Mixed (ref.: No parking) | −0.973 | 0.640 | −1.521 | 0.128 |
Goodness-of-fit measures | ||||
LRT (reference: null model) = 57.69 (df = 7), p < 0.001; LRT (reference: full model) = 5.63 (df = 7), p = 0.584 | ||||
Nagelkerke R2 = 0.102, AIC = 1369.2 | ||||
Over-dispersion parameter (theta) = 1.32 (std. error: 0.29) | ||||
Interval of continuous predictors | ||||
AADT: 1160–29,960 vehicles/day, L: 30–1415 m |
3.2 Predictive models for intersections
Predictor | Estimate | Std. Error | z- value | p-value |
---|---|---|---|---|
All intersections | ||||
(Intercept) | 0.930 | 1.524 | 0.610 | 0.542 |
Ln (Total AADT) | 0.285 | 0.106 | 2.675 | 0.007 |
Main AADT/Total AADT | −6.870 | 2.021 | −3.399 | 0.001 |
Type: Four-legged (reference: Three-legged) | 0.443 | 0.105 | 4.210 | < 0.001 |
Critical sight distance | 0.005 | 0.002 | 2.762 | 0.006 |
Specialized turning lane (Yes – ref.: no) | 0.496 | 0.114 | 4.362 | < 0.001 |
Poor maintenance of pavements (Yes - ref.: No) | −0.397 | 0.115 | −3.444 | < 0.001 |
Cycle path crossing (Yes – ref.: No) | 0.327 | 0.122 | 2.677 | 0.007 |
Goodness-of-fit measures | ||||
LRT (reference: null model) = 5.63 (df = 12), p = .934; LRT (reference: full model) = 161.11 (df = 12), p < .001 | ||||
Nagelkerke R2 = 0.283, AIC = 1877.6 | ||||
Over-dispersion parameter (theta) = 2.84 (std. error: 0.52) | ||||
Interval of continuous predictors | ||||
Total AADT: 4175–41,560 vehicles/day, Main/Total AADT: 0.50–0.62, Critical sight distance: 0.0–139.5 m | ||||
Three-legged intersections | ||||
(Intercept) | 4.696 | 2.895 | 1.609 | 0.108 |
Ln (Total AADT) | 0.366 | 0.220 | 1.661 | 0.097 |
Main AADT/Total AADT | −16.376 | 3.711 | −4.413 | < 0.001 |
Number of entering lanes on the main road | −0.211 | 0.119 | −1.768 | 0.077 |
Critical sight distance | 0.006 | 0.003 | 2.009 | < 0.001 |
Specialized turning lane (Yes – ref.: no) | 0.927 | 0.228 | 4.061 | < 0.001 |
Traffic control 1: give-way/stop (ref.: no control)a | 0.359 | 0.214 | 1.675 | 0.094 |
Traffic control 2: traffic lights (ref.: no control)a | 0.377 | 0.289 | 1.306 | 0.192 |
Traffic control 3: main + turning lights (ref.: no control)a | 0.556 | 0.313 | 1.777 | 0.076 |
Cycle path crossing (Yes – ref.: No) | 0.900 | 0.235 | 3.830 | < 0.001 |
Goodness-of-fit measures | ||||
LRT (reference: null model) = 63.39 (df = 9), p < .001; LRT (reference: full model) = 5.79 (df = 9), p = .761 | ||||
Nagelkerke R2 = 0.292, AIC = 610.4 | ||||
Over-dispersion parameter (theta) = 5.26 (std. error: 2.75) | ||||
Interval of continuous predictors | ||||
Total AADT: 5420–32,300 vehicles/day, Main /Total AADT: 0.50–0.62, Critical sight distance: 0.0–139.5 m, Number of entering lanes on the main road: 0–4 | ||||
Four-legged intersections | ||||
(Intercept) | −0.127 | 1.623 | −0.078 | 0.938 |
Ln (Total AADT) | 0.372 | 0.108 | 3.450 | 0.006 |
Main AADT/Total AADT | −5.427 | 2.624 | −2.068 | 0.039 |
Critical sight distance | 0.004 | 0.002 | 2.102 | 0.036 |
Specialized turning lane (Yes – ref.: no) | 0.410 | 0.134 | 3.060 | 0.002 |
Poor maintenance of pavements (Yes - ref.: No) | −0.446 | 0.151 | −2.961 | 0.003 |
LRT (reference: null model) = 77.98 (df = 5), p < .001; LRT (reference: full model) = 7.55 (df = 12), p = .819 | ||||
Nagelkerke R2 = 0.225, AIC = 1263.6 | ||||
Over-dispersion parameter (theta) = 2.79 (std. error: 0.59) | ||||
Interval of continuous predictors | ||||
Total AADT: 4175–41,560 vehicles/day, Main /Total AADT: 0.50–0.61, Critical sight distance: 1.6–100.0 m | ||||
All intersections (alternative) | ||||
(Intercept) | 1.713 | 1.513 | 1.132 | 0.257 |
Ln (Total AADT) | 0.371 | 0.103 | 3.613 | < 0.001 |
Main AADT/Total AADT | −9.502 | 2.161 | −4.397 | < 0.001 |
Median on the main road (Yes – reference: No) | 0.226 | 0.117 | 1.930 | 0.054 |
Critical sight distance | 0.004 | 0.002 | 2.359 | 0.018 |
Type of intersection (Signalized – ref.: Unsignalized) | 0.268 | 0.104 | 2.580 | 0.010 |
Poor maintenance of pavements (Yes - ref.: No) | −0.321 | 0.118 | −2.730 | 0.006 |
Cycle path crossing (Yes - ref.: No) | 0.253 | 0.123 | 2.058 | 0.040 |
Goodness-of-fit measures | ||||
LRT (reference: null model) = 124.59 (df = 7), p < .001; LRT (reference: full model) = 8.29 (df = 8), p = .406 | ||||
Nagelkerke R2 = 0.222, AIC = 1914.1 | ||||
Over-dispersion parameter (theta) = 2.40 (std. error: 0.41) | ||||
Interval of continuous predictors | ||||
Total AADT: 4175–41,560 vehicles/day, Main/Total AADT: 0.50–0.62, Critical sight distance: 0.0–139.5 m | ||||
Signalized intersections | ||||
(Intercept) | 2.390 | 2.293 | 1.043 | 0.297 |
Ln (Total AADT) | 0.407 | 0.142 | 2.863 | 0.004 |
Main AADT/Total AADT | −11.302 | 3.586 | −3.152 | 0.002 |
Type: Four-legged (reference: Three-legged) | 0.452 | 0.182 | 2.479 | 0.013 |
Critical sight distance | 0.004 | 0.002 | 1.986 | 0.047 |
Specialized turning lane (Yes – ref.: no) | 0.307 | 0.148 | 2.076 | 0.038 |
Poor maintenance of pavements (Yes - ref.: No) | −0.630 | 0.148 | −4.243 | < 0.001 |
Presence of bus stops (Yes – ref.: No) | −0.914 | 0.357 | −2.559 | 0.010 |
Goodness-of-fit measures | ||||
LRT (reference: null model) = 95.60 (df = 7), p < 0.001; LRT (reference: full model) = 4.39 (df = 9), p = 0.884 | ||||
Nagelkerke R2 = 0.320, AIC = 1078.2 | ||||
Over-dispersion parameter (theta) = 3.08 (std. error: 0.71) | ||||
Interval of continuous predictors | ||||
Total AADT: 5800–41,560 vehicles/day, Main/Total AADT: 0.50–0.61, Critical sight distance: 1.6–139.5 m | ||||
Unsignalized intersections | ||||
(Intercept) | −2.604 | 2.185 | −1.192 | 0.233 |
Ln (Total AADT) | 0.759 | 0.163 | 6.343 | < 0.001 |
Main AADT/Total AADT | −7.868 | 0.173 | 4.389 | 0.001 |
Type: Four-legged (reference: Three-legged) | 1.036 | 2.406 | −3.270 | < 0.001 |
Median on the main road (Yes – reference: No) | −0.998 | 0.329 | −3.031 | 0.002 |
Specialized turning lane (Yes – ref.: No) | 0.516 | 0.204 | 2.523 | 0.012 |
Cycle path crossing (Yes - ref.: No) | 0.997 | 0.193 | 5.156 | < 0.001 |
Presence of sidewalks (Yes - ref.: No) | −1.008 | 0.331 | −3.042 | 0.002 |
LRT (reference: null model) = 65.07 (df = 7), p < .001; LRT (reference: full model) = 11.96 (df = 11), p = .367 | ||||
Nagelkerke R2 = 0.238, AIC = 769.8 | ||||
Over-dispersion parameter (theta) = 8.11 (std. error: 6.14) | ||||
Interval of continuous predictors | ||||
Total AADT: 4175–31,500 vehicles/day, Main/Total AADT: 0.50–0.62, Critical sight distance: 0.0–100.0 m |