Modeling crash outcome probabilities at rural intersections: application of hierarchical binomial logistic models

It is important to examine the nature of the relationships between roadway, environmental, and traffic factors and motor vehicle crashes, with the aim to improve the collective understanding of causal mechanisms involved in crashes and to better predict their occurrence. Statistical models of motor vehicle crashes are one path of inquiry often used to gain these initial insights. Recent efforts have focused on the estimation of negative binomial and Poisson regression models (and related deviants) due to their relatively good fit to crash data. Of course analysts constantly seek methods that offer greater consistency with the data generating mechanism (motor vehicle crashes in this case), provide better statistical fit, and provide insight into data structure that was previously unavailable. One such opportunity exists with some types of crash data, in particular crash-level data that are collected across roadway segments, intersections, etc. It is argued in this paper that some crash data possess hierarchical structure that has not routinely been exploited. This paper describes the application of binomial multilevel models of crash types using 548 motor vehicle crashes collected from 91 two-lane rural intersections in the state of Georgia. Crash prediction models are estimated for angle, rear-end, and sideswipe (both same direction and opposite direction) crashes. The contributions of the paper are the realization of hierarchical data structure and the application of a theoretically appealing and suitable analysis approach for multilevel data, yielding insights into intersection-related crashes by crash type.     

Temporal and spatial analyses of rear-end crashes at signalized intersections

In this study, the generalized estimating equations with the negative binomial link function were used to model rear-end crash frequencies at signalized intersections to account for the temporal or spatial correlation among the data. The longitudinal data for 208 signalized intersections over 3 years and the spatially correlated data for 476 signalized intersections which are located along different corridors were collected in the state of Florida. The modeling results showed that there are high correlations between the longitudinal or spatially correlated rear-end crashes. Some intersection related variables are identified as significantly influencing rear-end crash occurrences at signalized intersections. Intersections with heavy traffic on the major and minor roadways, having more right and left-turn lanes on the major roadway, having a large number of phases per cycle (indicated by the left-turn protection on the minor roadway), with high speed limits on the major roadway, and in high population areas are correlated with high rear-end crash frequencies. On the other hand, intersections with three legs, having channelized or exclusive right-turn lanes on the minor roadway, with protected left-turning on the major roadway, with medians on the minor roadway, and having longer signal spacing have a lower frequency of rear-end crashes.     

Gender differences in injury severity risks in crashes at signalized intersections

This paper analyzes gender differences in crash risk severities using data for signalized intersections. It estimates gender models for injury severity risks and finds that driver condition, type of crash, type of vehicle driven and vehicle safety features have different effects on females’ and males’ injury severity risks. Also, it finds some variables which are significantly related to females’ injury severity risks but not males’ and others which affect males’ injury severity risks but not females’. It concludes that better and more in-depth information about gender differences in injury severity risks is gained by estimating separate models for females and males.     

Applying the random effect negative binomial model to examine traffic accident occurrence at signalized intersections

Poisson and negative binomial (NB) models have been used to analyze traffic accident occurrence at intersections for several years. There are however, limitations in the use of such models. The Poisson model requires the variance-to-mean ratio of the accident data to be about 1. Both the Poisson and the NB models require the accident data to be uncorrelated in time. Due to unobserved heterogeneity and serial correlation in the accident data, both models seem to be inappropriate. A more suitable alternative is the random effect negative binomial (RENB) model, which by treating the data in a time-series cross-section panel, will be able to deal with the spatial and temporal effects in the data. This paper describes the use of RENB model to identify the elements that affect intersection safety. To establish the suitability of the model, several goodness-of-fit statistics are used. The model is then applied to investigate the relationship between accident occurrence and the geometric, traffic and control characteristics of signalized intersections in Singapore. The results showed that 11 variables significantly affected the safety at the intersections. The total approach volumes, the numbers of phases per cycle, the uncontrolled left-turn lane and the presence of a surveillance camera are among the variables that are the highly significant.     

Contributory factors to traffic crashes at signalized intersections in Hong Kong

Efficient geometric design and signal timing not only improve operational performance at signalized intersections by expanding capacity and reducing traffic delays, but also result in an appreciable reduction in traffic conflicts, and thus better road safety. Information on the incidence of crashes, traffic flow, geometric design, road environment, and traffic control at 262 signalized intersections in Hong Kong during 2002 and 2003 are incorporated into a crash prediction model. Poisson regression and negative binomial regression are used to quantify the influence of possible contributory factors on the incidence of killed and severe injury (KSI) crashes and slight injury crashes, respectively, while possible interventions by traffic flow are controlled. The results for the incidence of slight injury crashes reveal that the road environment, degree of curvature, and presence of tram stops are significant factors, and that traffic volume has a diminishing effect on the crash risk. The presence of tram stops, number of pedestrian streams, road environment, proportion of commercial vehicles, average lane width, and degree of curvature increase the risk of KSI crashes, but the effect of traffic volume is negligible.     

Severity of driver injury and vehicle damage in traffic crashes at intersections: a Bayesian hierarchical analysis

Most crash severity studies ignored severity correlations between driver-vehicle units involved in the same crashes. Models without accounting for these within-crash correlations will result in biased estimates in the factor effects. This study developed a Bayesian hierarchical binomial logistic model to identify the significant factors affecting the severity level of driver injury and vehicle damage in traffic crashes at signalized intersections. Crash data in Singapore were employed to calibrate the model. Model fitness assessment and comparison using intra-class correlation coefficient (ICC) and deviance information criterion (DIC) ensured the suitability of introducing the crash-level random effects. Crashes occurring in peak time and in good street-lighting condition as well as those involving pedestrian injuries tend to be less severe. But crashes that occur in night time, at T/Y type intersections, and on right-most lane, as well as those that occur in intersections where red light cameras are installed tend to be more severe. Moreover, heavy vehicles have a better resistance on severe crash and thus induce less severe injuries, while crashes involving two-wheel vehicles, young or aged drivers, and the involvement of offending party are more likely to result in severe injuries.     

Cyclist activity and injury risk analysis at signalized intersections: A Bayesian modelling approach

This study proposes a two-equation Bayesian modelling approach to simultaneously study cyclist injury occurrence and bicycle activity at signalized intersections as joint outcomes. This approach deals with the potential presence of endogeneity and unobserved heterogeneities and is used to identify factors associated with both cyclist injuries and volumes. Its application to identify high-risk corridors is also illustrated. Montreal, Quebec, Canada is the application environment, using an extensive inventory of a large sample of signalized intersections containing disaggregate motor-vehicle traffic volumes and bicycle flows, geometric design, traffic control and built environment characteristics in the vicinity of the intersections. Cyclist injury data for the period of 2003–2008 is used in this study. Also, manual bicycle counts were standardized using temporal and weather adjustment factors to obtain average annual daily volumes. Results confirm and quantify the effects of both bicycle and motor-vehicle flows on cyclist injury occurrence. Accordingly, more cyclists at an intersection translate into more cyclist injuries but lower injury rates due to the non-linear association between bicycle volume and injury occurrence. Furthermore, the results emphasize the importance of turning motor-vehicle movements. The presence of bus stops and total crosswalk length increase cyclist injury occurrence whereas the presence of a raised median has the opposite effect. Bicycle activity through intersections was found to increase as employment, number of metro stations, land use mix, area of commercial land use type, length of bicycle facilities and the presence of schools within 50–800 m of the intersection increase. Intersections with three approaches are expected to have fewer cyclists than those with four. Using Bayesian analysis, expected injury frequency and injury rates were estimated for each intersection and used to rank corridors. Corridors with high bicycle volumes, located mainly in the central neighbourhoods of Montreal, have lower risk of injury. These results may reflect the “safety in numbers” hypothesis or cyclist preference towards safer intersections and corridors. Despite these corridors having a lower individual risk, they are nevertheless associated with a greater number of injuries.     

Exploring the overall and specific crash severity levels at signalized intersections

Many studies have shown that intersections are among the most dangerous locations of a roadway network. Therefore, there is a need to understand the factors that contribute to injuries at such locations. This paper addresses the different factors that affect crash injury severity at signalized intersections. It also looks into the quality and completeness of the crash data and the effect that incomplete data has on the final results. Data from multiple sources have been cross-checked to ensure the completeness of all crashes including minor crashes that are usually unreported or not coded into crash databases. The ordered probit modeling technique has been adopted in this study to account for the fact that injury levels are naturally ordered variables. The tree-based regression methodology has also been adopted in this study to explore the factors that affect each severity level. The probit model results showed that a combination of crash-specific information and intersection characteristics result in the highest prediction rate of injury level. More specifically, having a divided minor roadway or a higher speed limit on the minor roadway decreased the level of injury while crashes involving a pedestrian/bicyclist and left turn crashes had the highest probability of a more severe crash. Several regression tree models showed a difference in the significant factors that affect the different severity types. Completing the data with minor non injury crashes improved the modeling results and depicted differences when modeling the no injury crashes.     

Multimodal injury risk analysis of road users at signalized and non-signalized intersections

This paper proposes a multimodal approach to study safety at intersections by simultaneously analysing the safety and flow outcomes for both motorized and non-motorized traffic. This study uses an extensive inventory of signalized and non-signalized intersections on the island of Montreal, Quebec, Canada, containing disaggregate motor-vehicle, cyclist and pedestrian flows, injury data, geometric design, traffic control and built environment characteristics in the vicinity of each intersection. Bayesian multivariate Poisson models are used to analyze the injury and traffic flow outcomes and to develop safety performance functions for each mode at both facilities. After model calibration, contributing injury frequency factors are identified. Injury frequency and injury risk measures are then generated to carry out a comparative study to identify which mode is at greatest risk at intersections in Montreal. Among other results, this study identified the significant effect that motor-vehicle traffic imposes on cyclist and pedestrian injury occurrence. Motor-vehicle traffic is the main risk determinant for all injury and intersection types. This highlights the need for safety improvements for cyclists and pedestrians who are, on average, at 14 and12 times greater risk than motorists, respectively, at signalized intersections. Aside from exposure measures, this work also identifies some geometric design and built environment characteristics affecting injury occurrence for cyclists, pedestrians and motor-vehicle occupants.     

Road factors and bicycle-motor vehicle crashes at unsignalized priority intersections

In this study, the safety of cyclists at unsignalized priority intersections within built-up areas is investigated. The study focuses on the link between the characteristics of priority intersection design and bicycle–motor vehicle (BMV) crashes. Across 540 intersections that are involved in the study, the police recorded 339 failure-to-yield crashes with cyclists in four years. These BMV crashes are classified into two types based on the movements of the involved motorists and cyclists:

• • 

  type I: through bicycle related collisions where the cyclist has right of way (i.e. bicycle on the priority road);

• • 

  type II: through motor vehicle related collisions where the motorist has right of way (i.e. motorist on the priority road).

The probability of each crash type was related to its relative flows and to independent variables using negative binomial regression. The results show that more type I crashes occur at intersections with two-way bicycle tracks, well marked, and reddish coloured bicycle crossings. Type I crashes are negatively related to the presence of raised bicycle crossings (e.g. on a speed hump) and other speed reducing measures. The accident probability is also decreased at intersections where the cycle track approaches are deflected between 2 and 5 m away from the main carriageway. No significant relationships are found between type II crashes and road factors such as the presence of a raised median.     

Using hierarchical tree-based regression model to predict train-vehicle crashes at passive highway-rail grade crossings

This paper applies a nonparametric statistical method, hierarchical tree-based regression (HTBR), to explore train-vehicle crash prediction and analysis at passive highway-rail grade crossings. Using the Federal Railroad Administration (FRA) database, the research focuses on 27 years of train-vehicle accident history in the United States from 1980 through 2006. A cross-sectional statistical analysis based on HTBR is conducted for public highway-rail grade crossings that were upgraded from crossbuck-only to stop signs without involvement of other traffic-control devices or automatic countermeasures. In this study, HTBR models are developed to predict train-vehicle crash frequencies for passive grade crossings controlled by crossbucks only and crossbucks combined with stop signs respectively, and assess how the crash frequencies change after the stop-sign treatment is applied at the crossbuck-only-controlled crossings. The study results indicate that stop-sign treatment is an effective engineering countermeasure to improve safety at the passive grade crossings. Decision makers and traffic engineers can use the HTBR models to examine train-vehicle crash frequency at passive crossings and assess the potential effectiveness of stop-sign treatment based on specific attributes of the given crossings.     

Investigating different approaches to develop informative priors in hierarchical Bayesian safety performance functions

The Bayesian inference method has been frequently adopted to develop safety performance functions. One advantage of the Bayesian inference is that prior information for the independent variables can be included in the inference procedures. However, there are few studies that discussed how to formulate informative priors for the independent variables and evaluated the effects of incorporating informative priors in developing safety performance functions. This paper addresses this deficiency by introducing four approaches of developing informative priors for the independent variables based on historical data and expert experience. Merits of these informative priors have been tested along with two types of Bayesian hierarchical models (Poisson-gamma and Poisson-lognormal models). Deviance information criterion (DIC), R-square values, and coefficients of variance for the estimations were utilized as evaluation measures to select the best model(s). Comparison across the models indicated that the Poisson-gamma model is superior with a better model fit and it is much more robust with the informative priors. Moreover, the two-stage Bayesian updating informative priors provided the best goodness-of-fit and coefficient estimation accuracies. Furthermore, informative priors for the inverse dispersion parameter have also been introduced and tested. Different types of informative priors’ effects on the model estimations and goodness-of-fit have been compared and concluded. Finally, based on the results, recommendations for future research topics and study applications have been made.     

Using hierarchical Bayesian binary probit models to analyze crash injury severity on high speed facilities with real-time traffic data

Severe crashes are causing serious social and economic loss, and because of this, reducing crash injury severity has become one of the key objectives of the high speed facilities’ (freeway and expressway) management. Traditional crash injury severity analysis utilized data mainly from crash reports concerning the crash occurrence information, drivers’ characteristics and roadway geometric related variables. In this study, real-time traffic and weather data were introduced to analyze the crash injury severity. The space mean speeds captured by the Automatic Vehicle Identification (AVI) system on the two roadways were used as explanatory variables in this study; and data from a mountainous freeway (I-70 in Colorado) and an urban expressway (State Road 408 in Orlando) have been used to identify the analysis result's consistence. Binary probit (BP) models were estimated to classify the non-severe (property damage only) crashes and severe (injury and fatality) crashes. Firstly, Bayesian BP models’ results were compared to the results from Maximum Likelihood Estimation BP models and it was concluded that Bayesian inference was superior with more significant variables. Then different levels of hierarchical Bayesian BP models were developed with random effects accounting for the unobserved heterogeneity at segment level and crash individual level, respectively. Modeling results from both studied locations demonstrate that large variations of speed prior to the crash occurrence would increase the likelihood of severe crash occurrence. Moreover, with considering unobserved heterogeneity in the Bayesian BP models, the model goodness-of-fit has improved substantially. Finally, possible future applications of the model results and the hierarchical Bayesian probit models were discussed.     

Corridor-level signalized intersection safety analysis in Shanghai,China using Bayesian hierarchical models

Most traffic crashes in Chinese cities occur at signalized intersections. Research on the intersection safety problem in China is still in its early stage. The recent development of an advanced traffic information system in Shanghai enables in-depth intersection safety analyses using road design, traffic operation, and crash data. In Shanghai, the road network density is relatively high and the distance between signalized intersections is small, averaging about 200 m. Adjacent signalized intersections located along the same corridor share similar traffic flows, and signals are usually coordinated. Therefore, when studying intersection safety in Shanghai, it is essential to account for intersection correlations within corridors. In this study, data for 195 signalized intersections along 22 corridors in the urban areas of Shanghai were collected. Mean speeds and speed variances of corridors were acquired from taxis equipped with Global Positioning Systems (GPS). Bayesian hierarchical models were applied to identify crash risk factors at both the intersection and the corridor levels. Results showed that intersections along corridors with lower mean speeds were associated with fewer crashes than those with higher speeds, and those intersections along two-way roads, under elevated roads, and in close proximity to each other, tended to have higher crash frequencies.     

Driver injury severity outcome analysis in rural interstate highway crashes: a two-level Bayesian logistic regression interpretation

There is a high potential of severe injury outcomes in traffic crashes on rural interstate highways due to the significant amount of high speed traffic on these corridors. Hierarchical Bayesian models are capable of incorporating between-crash variance and within-crash correlations into traffic crash data analysis and are increasingly utilized in traffic crash severity analysis. This paper applies a hierarchical Bayesian logistic model to examine the significant factors at crash and vehicle/driver levels and their heterogeneous impacts on driver injury severity in rural interstate highway crashes. Analysis results indicate that the majority of the total variance is induced by the between-crash variance, showing the appropriateness of the utilized hierarchical modeling approach. Three crash-level variables and six vehicle/driver-level variables are found significant in predicting driver injury severities: road curve, maximum vehicle damage in a crash, number of vehicles in a crash, wet road surface, vehicle type, driver age, driver gender, driver seatbelt use and driver alcohol or drug involvement. Among these variables, road curve, functional and disabled vehicle damage in crash, single-vehicle crashes, female drivers, senior drivers, motorcycles and driver alcohol or drug involvement tend to increase the odds of drivers being incapably injured or killed in rural interstate crashes, while wet road surface, male drivers and driver seatbelt use are more likely to decrease the probability of severe driver injuries. The developed methodology and estimation results provide insightful understanding of the internal mechanism of rural interstate crashes and beneficial references for developing effective countermeasures for rural interstate crash prevention.     

The hierarchical Tobit model: A case study in Bayesian computing

In this paper we discuss the potentials of a new Bayesian inference tool, called the Gibbs sampler, for the analysis of the censored regression or Tobit model. Tobit models have a wide range of applications in empirical sciences, like econometrics and biometrics. The estimation results of the simple Tobit model will be compared to a hierarchical Tobit model, and the Gibbs sampling approach to the related classical algorithm of expectation-maximisation (EM). The underlying botanical example of this paper is concerned with the censoring mechanism in plant reproduction and proposes the Bayesian Tobit model for the growth relationship between the reproductive part and the rest of the plant.     

Examining driver injury severity outcomes in rural non-interstate roadway crashes using a hierarchical ordered logit model

Rural non-interstate crashes induce a significant amount of severe injuries and fatalities. Examination of such injury patterns and the associated contributing factors is of practical importance. Taking into account the ordinal nature of injury severity levels and the hierarchical feature of crash data, this study employs a hierarchical ordered logit model to examine the significant factors in predicting driver injury severities in rural non-interstate crashes based on two-year New Mexico crash records. Bayesian inference is utilized in model estimation procedure and 95% Bayesian Credible Interval (BCI) is applied to testing variable significance. An ordinary ordered logit model omitting the between-crash variance effect is evaluated as well for model performance comparison. Results indicate that the model employed in this study outperforms ordinary ordered logit model in model fit and parameter estimation. Variables regarding crash features, environment conditions, and driver and vehicle characteristics are found to have significant influence on the predictions of driver injury severities in rural non-interstate crashes. Factors such as road segments far from intersection, wet road surface condition, collision with animals, heavy vehicle drivers, male drivers and driver seatbelt used tend to induce less severe driver injury outcomes than the factors such as multiple-vehicle crashes, severe vehicle damage in a crash, motorcyclists, females, senior drivers, driver with alcohol or drug impairment, and other major collision types. Research limitations regarding crash data and model assumptions are also discussed. Overall, this research provides reasonable results and insight in developing effective road safety measures for crash injury severity reduction and prevention.     

Data-driven uncertainty quantification and propagation in structural dynamics through a hierarchical Bayesian framework

In the presence of modeling errors, the mainstream Bayesian methods seldom give a realistic account of uncertainties as they commonly underestimate the inherent variability of parameters. This problem is not due to any misconceptions in the Bayesian framework since it is robust with respect to the modeling assumptions and the observed data. Rather, this issue has deep roots in users’ inability to develop an appropriate class of probabilistic models. This paper bridges this significant gap, introducing a novel Bayesian hierarchical setting, which breaks time-history vibration responses into several segments so as to capture and identify the variability of inferred parameters over the segments. Since the computation of the posterior distributions in hierarchical models is expensive and cumbersome, novel marginalization strategies, asymptotic approximations, and maximum a posteriori estimations are proposed and outlined in a computational algorithm aiming to handle both uncertainty quantification and propagation. For the first time, the connection between the ensemble covariance matrix and hyper distribution parameters is characterized through approximate estimations. Experimental and numerical examples are employed to illustrate the efficacy and efficiency of the proposed method. It is observed that, when the segments correspond to various system operating conditions and input characteristics, the proposed method delivers robust parametric uncertainties with respect to unknown phenomena such as ambient conditions, input characteristics, and environmental factors.     

Bayesian transformed models for small area estimation

Sample surveys are usually designed and analyzed to produce estimates for larger areas. However, sample sizes are often not large enough to give adequate precision for small area estimates of interest. To overcome such difficulties,borrowing strength from related small areas via modeling becomes an appropriate approach. In line with this, we propose hierarchical models with power transformations for improving the precision of small area predictions. The proposed methods are applied to satellite data in conjunction with survey data to estimate mean acreage under a specified crop for counties in Iowa.     

Using a reliability process to reduce uncertainty in predicting crashes at unsignalized intersections

The negative binomial (NB) model has been used extensively by traffic safety analysts as a crash prediction model, because it can accommodate the over-dispersion criterion usually exhibited in crash count data. However, the NB model is still a probabilistic model that may benefit from updating the parameters of the covariates to better predict crash frequencies at intersections. The objective of this paper is to examine the effect of updating the parameters of the covariates in the fitted NB model using a Bayesian updating reliability method to more accurately predict crash frequencies at 3-legged and 4-legged unsignalized intersections. For this purpose, data from 433 unsignalized intersections in Orange County, Florida were collected and used in the analysis. Four Bayesian-structure models were examined: (1) a non-informative prior with a log-gamma likelihood function, (2) a non-informative prior with an NB likelihood function, (3) an informative prior with an NB likelihood function, and (4) an informative prior with a log-gamma likelihood function. Standard measures of model effectiveness, such as the Akaike information criterion (AIC), mean absolute deviance (MAD), mean square prediction error (MSPE) and overall prediction accuracy, were used to compare the NB and Bayesian model predictions. Considering only the best estimates of the model parameters (ignoring uncertainty), both the NB and Bayesian models yielded favorable results. However, when considering the standard errors for the fitted parameters as a surrogate measure for measuring uncertainty, the Bayesian methods yielded more promising results. The full Bayesian updating framework using the log-gamma likelihood function for updating parameter estimates of the NB probabilistic models resulted in the least standard error values.