Bicyclist injury severity in traffic crashes: A spatial approach for geo-referenced crash data to uncover non-stationary correlates

Introduction: Bicyclists are among vulnerable road users with their safety a key concern. This study generates new knowledge about their safety by applying a spatial modeling approach to uncover non-stationary correlates of bicyclist injury severity in traffic crashes. Method: The approach is Geographically Weighted Ordinal Logistic Regression (GWOLR), extended from the regular Ordered Logistic Regression (OLR) by incorporating the spatial perspective of traffic crashes. The GWOLR modeling approach allows the relationships between injury severity and its contributing factors to vary across the spatial domain, to account for the spatial heterogeneity. This approach makes use of geo-referenced data. This study explored more than 7,000 geo-referenced bicycle--motor-vehicle crashes in North Carolina. Results: This study performed a series of non-stationarity tests to identify local relationships that vary substantially across the spatial domain. These local relationships are related to the bicyclist (bicyclist age, bicyclist behavior, bicyclist intoxication, bicycle direction, bicycle position), motorist (driver age, driver intoxication, driver behavior, vehicle speed, vehicle type) and traffic (traffic volume). Conclusions: Results from the regular OLR are in general consistent with previous findings. For example, an increased bicyclist injury severity is associated with older bicyclists, bicyclist being intoxicated, and higher motor-vehicle speeds. Results from the GWOLR show local (rather than global) relationships between contributing factors and bicyclist injury severity. Practical Applications: Researchers and practitioners may use GWOLR to prioritize cycling safety countermeasures for specific regions. For example, GWOLR modeling estimates in the study highlighted the west part (from Charlotte to Asheville) of North Carolina for increased bicyclist injury severity due to the intoxication of road users including both bicyclists and drivers. Therefore, if a countermeasure is concerned with the road user intoxication, there may be a priority for the region from Charlotte to Asheville (relative to other areas in North Carolina).     

Behavioral pathways in bicycle-motor vehicle crashes: From contributing factors,pre-crash actions,to injury severities

Introduction: This study performed a path analysis to uncover the behavioral pathways (from contributing factors, pre-crash actions to injury severities) in bicycle-motor vehicle crashes. Method: The analysis investigated more than 7,000 bicycle-motor vehicle crashes in North Carolina between 2007 and 2014. Pre-crash actions discussed in this study are actions of cyclists and motorists prior to the event of a crash, including “bicyclist failed to yield,” “motorist failed to yield,” “bicyclist overtaking motorist,” and “motorist overtaking bicyclist.” Results: Model results show significant correlates of pre-crash actions and bicyclist injury severity. For example, young bicyclists (18 years old or younger) are 23.5% more likely to fail to yield to motor traffic prior to the event of a crash than elder bicyclists. The “bicyclist failed to yield” action is associated with increased bicyclist injury severity than other actions, as this behavior is associated with an increase of 5.88 percentage points in probability of a bicyclist being at least evidently injured. The path analysis can highlight contributing factors related to risky pre-crash actions that lead to severe injuries. For example, bicyclists traveling on regular vehicle travel lanes are found to be more likely to involve the “bicyclist failed to yield” action, which resulted in a total 44.38% (7.04% direct effect + 37.34% indirect effect) higher likelihood of evident or severe injuries. The path analysis can also identify factors (e.g., intersection) that are not directly but indirectly correlated with injury severity through pre-crash actions. Practical Applications: This study offers a methodological framework to quantify the behavioral pathways in bicycle-motor vehicle crashes. The findings are useful for cycling safety improvements from the perspective of bicyclist behavior, such as the educational program for cyclists.     

Exploring bicyclist injury severity in bicycle-vehicle crashes using latent class clustering analysis and partial proportional odds models

Introduction: Bicyclists are more vulnerable compared to other road users. Therefore, it is critical to investigate the contributing factors to bicyclist injury severity to help provide better biking environment and improve biking safety. According to the data provided by National Highway Traffic Safety Administration (NHTSA), a total of 8,028 bicyclists were killed in bicycle-vehicle crashes from 2007 to 2017. The number of fatal bicyclists had increased rapidly by approximately 11.70% during the past 10 years (NHTSA, 2019). Methods: This paper conducts a latent class clustering analysis based on the police reported bicycle-vehicle crash data collected from 2007 to 2014 in North Carolina to identify the heterogeneity inherent in the crash data. First, the most appropriate number of clusters is determined in which each cluster has been characterized by the distribution of the featured variables. Then, partial proportional odds models are developed for each cluster to further analyze the impacts on bicyclist injury severity for specific crash patterns. Results: Marginal effects are calculated and used to evaluate and interpret the effect of each significant explanatory variable. The model results reveal that variables could have different influence on the bicyclist injury severity between clusters, and that some variables only have significant impacts on particular clusters. Conclusions: The results clearly indicate that it is essential to conduct latent class clustering analysis to investigate the impact of explanatory variables on bicyclist injury severity considering unobserved or latent features. In addition, the latent class clustering is found to be able to provide more accurate and insightful information on the bicyclist injury severity analysis. Practical Applications: In order to improve biking safety, regulations need to be established to prevent drinking and lights need to be provided since alcohol and lighting condition are significant factors in severe injuries according to the modeling results.     

The unintended impact of helmet use on bicyclists’ risk-taking behaviors

Introduction: Safety is a critical factor in promoting sustainable urban non-motorized travel modes like bicycles. Helmets have shown to be effective in reducing injury severity in bicycle crashes, however, their effects on bicyclists’ behaviors still requires deeper understanding, especially amid the emerging trend of using shared bicycles. Risk compensation effects suggest that bicyclists may offset perceived gains in safety from wearing a helmet by increasing risk-taking behaviors. A better understanding of these compensation effects can be useful in assessing various bicycle safety related programs. Method: Using a sample of 131 bicyclists from the San Francisco Bay area, this research studies how bicyclists respond with respect to risk-taking behaviors under various urban-street conditions, as a function of helmet use. Study participants were each shown 12 videos, shot in Berkeley, California, from the perspective of a bicyclist riding behind another bicyclist. A fractional factorial experiment design was used to systematically vary contextual attributes (e.g., speed, bike lane facilities, on-street parking, passing vehicles) across the videos. After each video, participants were asked to indicate if they would overtake the bicyclist in the video. With the help of data adaptive estimation techniques, targeted maximum likelihood estimation (TMLE) was applied to estimate the average risk difference between helmeted users and non-users, controlling for self-selection effects. Individual-based nonparametric bootstrap was performed to assess the uncertainty associated with the estimator. Results: Our findings suggest, on average, individuals more likely to wear a helmet are 15.6% more likely to undertake a risky overtaking maneuver. Practical Applications: This study doesn’t try to oppose mandatory helmet laws, but rather serves as a cautionary warning that road safety programs may need to consider strategies in which unintended impact of bicycle helmet use can be mitigated. Moreover, our findings also provide additional evaluation component when it comes to the cost-benefit assessment of helmet-related laws.     

Built environment effects on bike crash frequency and risk in Beijing

IntroductionBuilding a safe biking environment is crucial to encouraging bicycle use. In developed areas with higher density and more mixed land use, the built environment factors that pose a crash risk may vary. This study investigates the connection between biking risk factors and the compact built environment, using data for Beijing.MethodIn the context of China, this paper seeks to answer two research questions. First, what types of built environment factors are correlated with bike-automobile crash frequency and risk? Second, how do risk factors vary across different types of bikes? Poisson lognormal random effects models are employed to examine how land use and roadway design factors are associated with the bike-automobile crashes.ResultsThe main findings are: (1) bike-automobile crashes are more likely to occur in densely developed areas, which is characterized by higher population density, more mixed land use, denser roads and junctions, and more parking lots; (2) areas with greater ground transit are correlated with more bike-automobile crashes and higher risks of involving in collisions; (3) the percentages of wider streets show negative associations with bike crash frequency; (4) built environment factors cannot help explain factors contributing to motorcycle-automobile crashes.Practical ApplicationsIn China's dense urban context, important policy implications for bicycle safety improvement drawn from this study include: prioritizing safety programs in urban centers, applying safety improvements to areas with more ground transit, placing bike-automobile crash countermeasures at road junctions, and improving bicycle safety on narrower streets.     

Towards safer bicyclist responses to the presence of a truck near an urban loading zone: Analysis of bicyclist perceived level of comfort

IntroductionWhile sophisticated plans have been adopted nationally and globally to increase bicycling’s share of daily commutes, safety concerns have negatively impacted targeted bicycling growth. To investigate people’s preferences for bicycling in dense urban areas, it is important to recognize how bicycling perceived level of comfort (PLOC) is constructed and how it could relate to safe versus risky behavior while interacting with motorized modes of transportation. Method: To examine these issues, we analyzed results from an online survey with 342 participants. Structural Equation Modeling (SEM) was employed to systematically investigate the construct of bicycling PLOC and simultaneously analyze bicyclists’ responses to the presence of a truck in the adjacent lane near an urban loading zone. Results: SEM estimation results indicated that participants who said that they engaged in more frequent distracted bicycling reported lower PLOC. On the other hand, those who felt that road users were more lawful and predictable, and who had more bicycling experience, reported higher levels of PLOC. Participants who bicycled for commuting purposes, who made shorter trips, who bicycled more frequently, and who had more exposure to downtown bicycling also reported higher levels of PLOC. Finally, findings showed that higher PLOC was significantly associated with the choice of a safe, rather than a risky response to the presence of a truck, suggesting that a way to improve bicyclist safety would be to build an environment that could increase bicyclists’ PLOC.     

Crash histories,safety perceptions,and attitudes among Virginia bicyclists

IntroductionCycling injury and fatality rates are on the rise, yet there exists no comprehensive database for bicycle crash injury data.MethodWidely used for safety analysis, police crash report datasets are automobile-oriented and widely known to under-report bicycle crashes. This research is one attempt to address gaps in bicycle data in sources like police crash reports. A survey was developed and deployed to enhance the quality and quantity of available bicycle safety data in Virginia. The survey captures bicyclist attitudes and perceptions of safety as well as bicycle crash histories of respondents.ResultsThe results of this survey most notably show very high levels of under-reporting of bicycle crashes, with only 12% of the crashes recorded in this survey reported to police. Additionally, the results of this work show that lack of knowledge concerning bicycle laws is associated with lower levels of cycling confidence. Count model results predict that bicyclists who stop completely at traffic signals are 40% less likely to be involved in crashes compared to counterparts who sometimes stop at signals. In this dataset, suburban and urban roads with designated bike lanes had more favorable injury severity profiles, with lower percentages of severe and minor injury crashes compared to similar roads with a shared bike/automobile lane or no designated bike infrastructure.     

An in-depth analysis of self-reported cycling injuries in single and multiparty bicycle crashes in Denmark

Introduction: Cycling is one of the main forms of transportation in Denmark. However, while the number of traffic crash fatalities in the country has decreased over the past decade, the frequency of cyclists killed or seriously injured has increased. The high rate of serious injuries and fatalities associated with cycling emphasizes the increasing need for mitigating the severity of such crashes. Method: This study conducted an in-depth analysis of cyclist injury severity resulting from single and multiparty bicycle-involved crashes. Detailed information was collected using self-reporting data undertaken in Denmark for a 12-month period between 1 November 2012 and 31 October 2013. Separate multilevel logistic (MLL) regression models were applied to estimate cyclist injury severity for single and multiparty crashes. The goodness-of-fit measures favored the MLL models over the standard logistic models, capturing the intercorrelation among bicycle crashes that occurred in the same geographical area. Results: The results also showed that single bicycle-involved crashes resulted in more serious outcomes when compared to multiparty crashes. For both single and multiparty bicycle crash categories, non-urban areas were associated with more serious injury outcomes. For the single crashes, wet surface condition, autumn and summer seasons, evening and night periods, non-adverse weather conditions, cyclists aged between 45 and 64 years, male sex, riding for the purpose of work or educational activities, and bicycles with light turned-off were associated with severe injuries. For the multiparty crashes, intersections, bicycle paths, non-winter season, not being employed or retired, lower personal car ownership, and race bicycles were directly related to severe injury consequences. Practical Applications: The findings of this study demonstrated that the best way to promote cycling safety is the combination of improving the design and maintenance of cycling facilities, encouraging safe cycling behavior, and intensifying enforcement efforts.     

Using latent class clustering and binary logistic regression to model Australian cyclist injury severity in motor vehicle–bicycle crashes

Introduction: In recent years, Australia is seeing an increase in the total number of cyclists. However, the rise of serious injuries and fatalities to cyclists has been a major concern. Understanding the factors affecting the fatalities and injuries of bicyclists in crashes with motor vehicles is important to develop effective policy measures aimed at improving the safety of bicyclists. This study aims to identify the factors affecting motor vehicle-bicycle (MVB) crashes in Victoria, Australia and introducing effective countermeasures for the identified risk factors. Method: A data set of 14,759 MVB crash records from Victoria, Australia between 2006 and 2019 was analyzed using the binary logit model and latent class clustering. Results: It was observed that the factors that increase the risk of fatalities and serious injuries of bicyclists (FSI) in all clusters are: elderly bicyclist, not using a helmet, and darkness condition. Likewise, in areas with no traffic control, clear weather, and dry surface condition (cluster 1), high speed limits increase the risk of FSI, but the occurrence of MVB crashes in cross intersection and T-intersection has been significantly associated with a reduction in the risk of FSI. In areas with traffic control and unfavorable weather conditions (cluster 2), wet road surface increases the risk of FSI, but the areas with give way sign and pedestrian crossing signs reduce the risk of FSI. Practical Applications: Recommendations to reduce the risk of fatalities or serious injury to bicyclists are: improvement of road lighting and more exposure of bicyclists using reflective clothing and reflectors, separation of the bicycle and vehicle path in mid blocks especially in high-speed areas, using a more stable bicycle for the older people, monitoring helmet use, improving autonomous emergency braking, and using bicyclist detection technology for vehicles.     

Factors related to youth self-efficacy for injury prevention bicycle skills

Introduction: Bicycle riding is a common activity for children, but they are prone to bicycle-related injuries. It is well-established that injury prevention measures such as wearing a helmet and correctly riding a bicycle can reduce the severity of an injury and the likelihood of having an accident. However, how to increase bicycle injury prevention behaviors among children, who collectively fail to engage in injury prevention behaviors, is less well understood. Self-efficacy is consistently predictive of injury prevention behavior, making it an important approach to understanding injury prevention skills among this key population. The objective of this study was to explore and identify factors internal to the child as well as factors about his or her environment that predict a child’s self-efficacy for injury prevention skills. Method: Two generalized linear mixed effects models were created from survey data collected from elementary school students (n = 2,255) as part of a school-based bicycle education program. Models focused on self-efficacy for riding a bicycle and self-efficacy for wearing a helmet correctly. Results: In both models, road safety knowledge, opportunity for skill building through owning appropriate equipment (a bicycle or helmet), and situation through perception of neighborhood safety were predictive. The analyses reveal these variables as key factors for greater confidence, with feeling safe riding in the neighborhood, in particular, emerging as highly predictive of self-efficacy for injury prevention skills. Conclusions: These findings highlight the interplay of individual and environmental factors within confidence for injury prevention behavior. Given self-efficacy’s strong relationship to prevention behavior, these findings indicate actionable strategies. Practical Applications: The key factors highlighted in this study can be used by policymakers to target specific areas (e.g., neighborhood safety) to promote self-efficacy and thus improve injury prevention. These factors can also inform strategies for establishing safety skills in bicycle-safety education programs.     

Highway safety assessment and improvement through crash prediction by injury severity and vehicle damage using Multivariate Poisson-Lognormal model and Joint Negative Binomial-Generalized Ordered Probit Fractional Split model

Introduction: Predicting crash counts by severity plays a dominant role in identifying roadway sites that experience overrepresented crashes, or an increase in the potential for crashes with higher severity levels. Valid and reliable methodologies for predicting highway accidents by severity are necessary in assessing contributing factors to severe highway crashes, and assisting the practitioners in allocating safety improvement resources. Methods: This paper uses urban and suburban intersection data in Connecticut, along with two sophisticated modeling approaches, i.e. a Multivariate Poisson-Lognormal (MVPLN) model and a Joint Negative Binomial-Generalized Ordered Probit Fractional Split (NB-GOPFS) model to assess the methodological rationality and accuracy by accommodating for the unobserved factors in predicting crash counts by severity level. Furthermore, crash prediction models based on vehicle damage level are estimated using the same two methodologies to supplement the injury severity in estimating crashes by severity when the sample mean of severe injury crashes (e.g., fatal crashes) is very low. Results: The model estimation results highlight the presence of correlations of crash counts among severity levels, as well as the crash counts in total and crash proportions by different severity levels. A comparison of results indicates that injury severity and vehicle damage are highly consistent. Conclusions: Crash severity counts are significantly correlated and should be accommodated in crash prediction models. Practical application: The findings of this research could help select sound and reliable methodologies for predicting highway accidents by injury severity. When crash data samples have challenges associated with the low observed sampling rates for severe injury crashes, this research also confirmed that vehicle damage can be appropriate as an alternative to injury severity in crash prediction by severity.     

Analysis of the severity of vehicle-bicycle crashes with data mining techniques

Introduction: Although cycling is increasingly being promoted for transportation, the safety concern of bicyclists is one of the major impediments to their adoption. A thorough investigation on the contributing factors to fatalities and injuries involving bicyclist. Method: This paper designs an integrated data mining framework to determine the significant factors that contribute to the severity of vehicle-bicycle crashes based on the crash dataset of Victorian, Australia (2013–2018). The framework integrates imbalanced data resampling, learning-based feature extraction with gradient boosting algorithm and marginal effect analysis. The top 10 significant predictors of the severity of vehicle-bicycle crashes are extracted, which gives an area under ROC curve (AUC) value of 0.8236 and computing time as 37.8 s. Results: The findings provide insights for understanding and developing countermeasures or policy initiatives to reduce severe vehicle-bicycle crashes.     

Investigating hazardous factors affecting freeway crash injury severity incorporating real-time weather data: Using a Bayesian multinomial logit model with conditional autoregressive priors

Introduction: It has been demonstrated that weather conditions have significant impacts on freeway safety. However, when employing an econometric model to examine freeway crash injury severity, most of the existing studies tend to categorize several different adverse weather conditions such as rainy, snowy, and windy conditions into one category, “adverse weather,” which might lead to a large amount of information loss and estimation bias. Hence, to overcome this issue, real-time weather data, the value of meteorological elements when crashes occurred, are incorporated into the dataset for freeway crash injury analysis in this study. Methods: Due to the possible existence of spatial correlations in freeway crash injury data, this study presents a new method, the spatial multinomial logit (SMNL) model, to consider the spatial effects in the framework of the multinomial logit (MNL) model. In the SMNL model, the Gaussian conditional autoregressive (CAR) prior is adopted to capture the spatial correlation. In this study, the model results of the SMNL model are compared with the model results of the traditional multinomial logit (MNL) model. In addition, Bayesian inference is adopted to estimate the parameters of these two models. Result: The result of the SMNL model shows the significance of the spatial terms, which demonstrates the existence of spatial correlation. In addition, the SMNL model has a better model fitting ability than the MNL model. Through the parameter estimate results, risk factors such as vertical grade, visibility, emergency medical services (EMS) response time, and vehicle type have significant effects on freeway injury severity. Practical Application: According to the results, corresponding countermeasures for freeway roadway design, traffic management, and vehicle design are proposed to improve freeway safety. For example, steep slopes should be avoided if possible, and in-lane rumble strips should be recommended for steep down-slope segments. Besides, traffic volume proportion of large vehicles should be limited when the wind speed exceeds a certain grade.     

Considering built environment and spatial correlation in modeling pedestrian injury severity

Objective: This study looks at mitigating and aggravating factors that are associated with the injury severity of pedestrians when they have crashes with another road user and overcomes existing limitations in the literature by focusing attention on the built environment and considering spatial correlation across crashes.

Method: Reports for 6,539 pedestrian crashes occurred in Denmark between 2006 and 2015 were merged with geographic information system resources containing detailed information about the built environment and exposure at the crash locations. A linearized spatial logit model estimated the probability of pedestrians sustaining a severe or fatal injury conditional on the occurrence of a crash with another road user.

Results: This study confirms previous findings about older pedestrians and intoxicated pedestrians being the most vulnerable road users and crashes with heavy vehicles and in roads with higher speed limits being related to the most severe outcomes. This study provides novel perspectives by showing positive spatial correlations of crashes with the same severity outcomes and emphasizing the role of the built environment in the proximity of the crash.

Conclusions: This study emphasizes the need for thinking about traffic calming measures, illumination solutions, road maintenance programs, and speed limit reductions. Moreover, this study emphasizes the role of the built environment, because shopping areas, residential areas, and walking traffic density are positively related to a reduction in pedestrian injury severity. Often, these areas have in common a larger pedestrian mass that is more likely to make other road users more aware and attentive, whereas the same does not seem to apply to areas with lower pedestrian density.     

Experimental evaluation of alternative conspicuity-enhancement techniques for pedestrians and bicyclists

A field experiment was conducted to determine the extent of conspicuity enhancement provided pedestrians and bicyclists at night by various commercially available retroreflective materials and lights. The conspicuous materials were designed to be worn or carried by the pedestrians and bicyclists. Detection and recognition distances for the various experimental and baseline conditions were determined using subjects driving instrumented vehicles over a predetermined route on a realistic closed-course roadway system. Field experimenters were used to model the conspicuity-enhancing materials employing natural motion associated with walking and bicycling. Comparisons of the detection and recognition distances suggested that pedestrians and bicyclists can greatly enhance their conspicuity to drivers at night by wearing certain types of apparel and by using devices that are currently available in the marketplace. Nevertheless, it was concluded that nighttime pedestrian and bicyclist activity is inherently dangerous, even with these devices, and should be avoided.     

Assessment of temporal stability in risk factors of crashes at horizontal curves on rural two-lane undivided highways

Introduction: Safety of horizontal curves on rural two-lane, two-way undivided roadways is not fully explored. This study investigates factors that impact injury severity of such crashes. Method: To achieve the aim of this paper, issues associated with police-reported crash data such as unobserved heterogeneity and temporal stability need to be accounted for. Hence, a mixed logit model was estimated, while heterogeneity in means and variances is investigated by considering four injury severity outcomes for drivers: severe injury, moderate injury, possible injury, and no injury. Crash data for the period between 2011 and 2016 for crashes that occurred in the state of Oregon was analyzed. Temporal stability in factors determining the injury severity was investigated by identifying three time periods through splitting crash data into 2011–2012, 2013–2014, and 2015–2016. Results: Despite some factors affecting injuries in all specified time periods, the values of the marginal effects showed relative differences. The estimation results revealed that some factors increased the risk of being involved in severe injury crashes, including head-on collisions, drunk drivers, failure to negotiate curves, older drivers, and exceeding the speed limits. Conclusions: The hypothesis that attributes of injury severity are temporally stable is rejected. For example, young drivers (30 years old and younger) and middle-aged drivers were found to be temporally instable over time. Practical applications: The findings could help transportation authorities and safety professionals to enhance the safety of horizontal curves through appropriate and effective countermeasures.     

Exploring the severity of bicycle–vehicle crashes using latent class clustering approach in India

IntroductionBicyclists are vulnerable users in the shared asset like roadways. However, people still prefer to use bicycles for environmental, societal, and health benefits. In India, the bicycle plays a role in supporting the mobility to more people at lower cost and are often associated with the urban poor. Bicyclists represents one of the road user categories with highest risk of injuries and fatalities. According to the report by the Ministry of Road Transport and Highways (Accidents, 2017) in India, there is a sharp increase in the number of fatal victims for bicyclists in 2017 over 2016. The number of cyclists killed jumped from 2,585 in 2016 to 3,559 in 2017, a 37.7% increase. Method: Few studies have only investigated the crash risk perceived by the bicyclists while interacting with other road users. The present paper investigates the injury severity of bicyclists in bicycle-vehicle crashes that occurred in the state of Tamilnadu, India during the nine year period (2009–2017). The analyses demonstrate that dividing bicycle-vehicle collision data into five clusters helps in reducing the systematic heterogeneity present in the data and identify the hidden relationship between the injury severity levels of bicyclists and cyclists demographics, vehicle, environmental, temporal cause for the crashes. Results: Latent Class Clustering (LCC) approach was used in the present study as a preliminary tool for the segmentation of 9,978 crashes. Later, logistic regression analysis was used to identify the factors that influence bicycle crash severity for the whole dataset as well as for the clusters that were obtained from the LCC model. Results of this study show that combined use of both techniques reveals further information that wouldn’t be obtained without prior segmentation of the data. Few variables such as season, weather conditions, and light conditions were significant for certain clusters that were hidden in the whole dataset. This study can help domain experts or traffic safety researchers to segment traffic crashes and develop targeted countermeasures to mitigate injury severity.     

Hierarchical ordered model for injury severity of pedestrian crashes in South Korea

Introduction: The high percentage of fatalities in pedestrian-involved crashes is a critical social problem. The purpose of this study is to investigate factors influencing injury severity in pedestrian crashes by examining the demographic and socioeconomic characteristics of the regions where crashes occurred. Method: To understand the correlation between the unobserved characteristics of pedestrian crashes in a defined region, we apply a hierarchical ordered model, in which we set crash characteristics as lower-level variables and municipality characteristics as upper-level. Pedestrian crash data were collected and analyzed for a three-year period from 2011 to 2013. The estimation results show the statistically significant factors that increase injury severity of pedestrian crashes. Results: At the crash level, the factors associated with increased severity of pedestrian injury include intoxicated drivers, road-crossing pedestrians, elderly pedestrians, heavy vehicles, wide roads, darkness, and fog. At the municipality level, municipalities with low population density, lower level of financial independence, fewer doctors, and a higher percentage of elderly residents experience more severe pedestrian crashes. Municipalities ranked as having the top 10% pedestrian fatality rate (fatalities per 100,000 residents) have rates 7.4 times higher than municipalities with the lowest 10% rate of fatalities. Their demographic and socioeconomic characteristics also have significant differences. The proposed model accounts for a 7% unexplained variation in injury severity outcomes between the municipalities where crashes occurred. Conclusion: To enhance the safety of vulnerable pedestrians, considerable investments of time and effort in pedestrian safety facilities and zones should be made. More certain and severe punishments should be also given for the traffic violations that increase injury severity of pedestrian crashes. Furthermore, central and local governments should play a cooperative role to reduce pedestrian fatalities. Practical applications: Based on our study results, we suggest policy directions to enhance pedestrian safety.