Assessing the collective safety of automated vehicle groups: A duration modeling approach of accumulated distances between crashes.

Ideally, the evaluation of automated vehicles would involve the careful tracking of individual vehicles and recording of observed crash events. Unfortunately, due to the low frequency of crash events, such data would require many years to acquire, and potentially place the motorized public at risk if defective automated technologies were present. To acquire information on the safety effectiveness of automated vehicles more quickly, this paper uses the collective crash histories of a group of automated vehicles, and applies a duration modeling approach to the accumulated distances between crashes. To demonstrate the applicability of this approach as a method compare automated and conventional vehicles (human drivers), an empirical assessment was undertaken using two comparable sources of data. For conventional vehicles, police and non-police-reportable crashes were collected from the Second Strategic Highway Research Program's naturalistic driving study, and for automated vehicles, data from the California Department of Motor Vehicles Autonomous Vehicle Tester program were used (105 crashes from 59 permit holders driving ∼2.8 million miles were used for the analysis). The results of the empirical study showed that automated driving was safer at the 95% confidence level, with a higher number of miles between crashes, relative to their conventional vehicle counterparts. The findings indicate that the number of miles between crashes would be increased by roughly 27% when switching from conventional vehicles to automated vehicles. Despite limited data which mandated a group-vehicle approach, this study can be considered a reasonable initial approximation of automated vehicle safety.

Intended cycling frequency and the role of happiness and environmental friendliness after COVID-19.

Although the COVID-19 pandemic has contributed to an increase in cycling in many countries worldwide, it is not yet known whether this increase becomes a long-lasting change in mobility. The current study explores this increase by analyzing data collected in a U.S. nationwide longitudinal survey. Using a total of 7421 observations, a mixed logit model with heterogeneity in the means of random parameters was estimated. In the resulting sample, nearly 14 percent of the respondents stated that they were planning to cycle more while only 4 percent of the respondents stated that they were planning to cycle less post COVID-19 pandemic. The estimation results provide insights into socio-demographic and psychological factors that play a role in planned cycling behavior post COVID-19. The study also establishes that age, race, employment status, gender, and household size impact intended cycling frequency. The model estimation results further indicate that workers (full time and part time), individuals with a high degree of life satisfaction, and individuals who are environmentally friendly all have higher cycling-frequency probabilities relative to others. The findings can be used to support policies that target sustainable mobility and further our understanding of the transportation, psychology, and well-being relationships.

A statistical assessment of work-from-home participation during different stages of the COVID-19 pandemic.

Responses to the COVID-19 pandemic have dramatically transformed industry, healthcare, mobility, and education. Many workers have been forced to shift to work-from-home, adjust their commute patterns, and/or adopt new behaviors. Particularly important in the context of mitigating transportation-related emissions is the shift to work-from-home. This paper focuses on two major shifts along different stages of the pandemic. First, it investigates switching to work-from-home during the pandemic, followed by assessing the likelihood of continuing to work-from-home as opposed to returning to the workplace. This second assessment, being conditioned on workers having experienced work-from-home as the result of the pandemic, allows important insights into the factors affecting work-from-home probabilities. Using a survey collected in July and August of 2020, it is found that nearly 50 percent of the respondents who did not work-from-home before but started to work-from-home during the COVID-19 pandemic, indicated the willingness to continue work-from-home. A total of 1,275 observations collected using the survey questionnaire, that was administered through a U.S. nationwide panel (Prime Panels), were used in the model estimation. The methodological approach used to study work-from-home probabilities in this paper captures the complexities of human behavior by considering the effects of unobserved heterogeneity in a multivariate context, which allows for new insights into the effect of explanatory variables on the likelihood of working from home. Random parameters logit model estimations (with heterogeneity in the means and variances of random parameters) revealed additional insights into factors affecting work-from-home probabilities. It was found that gender, age, income, the presence of children, education, residential location, or job sectors including marketing, information technologies, business, or administration/administrative support all played significant roles in explaining these behavioral shifts and post-pandemic preferences.

Are forensic scientists too risk averse?

Fingerprint examiners maintain decision thresholds that represent the amount of evidence required for an identification or exclusion conclusion. As measured by error rate studies (Proc Natl Acad Sci USA. 2011;108(19):7733-8), these decision thresholds currently exhibit a preference for preventing erroneous identification errors at the expense of preventing erroneous exclusion errors. The goal of this study is to measure the decision thresholds for both fingerprint examiners and members of the general public, to determine whether examiners are more risk averse than potential jury members. To externally measure these decision thresholds, subjects manipulated decision criteria in a web-based visualization that reflects the trade-offs between erroneous identification decisions and erroneous exclusion decisions. Data from fingerprint examiners and the general public were compared to determine whether both groups have similar values as expressed by the placement of the decision criteria. The results of this study show that fingerprint examiners are more risk averse than members of the general public, although they align with error rate studies of fingerprint examiners. Demographic data demonstrate those factors that may contribute to differences in decision criterion placement, both between the two groups and between individuals within a group. The experimental methods provide a rich framework for measuring, interpreting, and responding to the values of society as applied to forensic decision-making.

The role of gender and temporal instability in driver-injury severities in crashes caused by speeds too fast for conditions.

The effect of inappropriate speed adjustment to adverse conditions on crash-injury severities, and how this effect might vary across male and female drivers, and over time, is not well understood. To study this, single-vehicle crashes occurring in rainy weather, where speed too fast for conditions is a driver action identified as a contributing factor to the crash, were considered. The differences between the resulting crash-injury severities of male and female drivers (and how these differences change over time) is then studied utilizing three years of Florida crash data and estimating random parameters multinomial logit models of driver injury severity while considering potential heterogeneity in the means and variances of parameter estimates. Model estimation results show that there were significant differences in the driver-injury severities of male and female drivers, and that the effect of factors that determine injury severities varied significantly over time (statistically significant temporal instability). This suggests that male and female drivers generally perceive and react to rainy weather conditions in fundamentally different ways, and that their responses, as reflected by the effect that explanatory variables have on injury severity probabilities, change over time. However, there were two explanatory variables that had relatively stable effects on injury-severity probabilities over time and across genders: an indicator variable for crashes involving non-collision factors (including overturn/rollover crashes) and an indicator variable for restraint usage. Policies that target these two variables could produce long-term reductions in crash-injury severities under adverse conditions.

Unobserved heterogeneity and the effects of driver nationality on crash injury severities in Saudi Arabia.

This paper investigates factors that significantly contribute to the injury severity of different drivers of different nationality backgrounds. Using the data from Riyadh, Saudi Arabia, a random parameters multinomial logit model of driver-injury severity was estimated to explore the effects of a wide range of variables on driver injury-severity outcomes. With three possible outcomes (no injury, injury, fatality), only single-vehicle crashes are considered and crashes involving domestic (Saudi) and international (non-Saudi) drivers were modeled separately. Model estimation results show that a wide range factors significantly affect the injury severity outcomes in single-vehicle crashes including driver attributes (such as nationality and age), vehicle characteristics (such as make, model and year of manufacture), driver actions (such as speeding and preoccupation on driving), and other factors (such as location and time of the accident); and that the influence that these variables have on injury-severity probabilities vary considerably between Saudi and non-Saudi drivers. While Saudi Arabia is rather unique because of the large numbers of non-national drivers, the results suggest that different nationalities, with their different cultural, educational and, behavioral backgrounds, may affect risk-taking behavior and resulting crash-injury severities.

The effects of drug and alcohol consumption on driver injury severities in single-vehicle crashes.

OBJECTIVE: It is well known that alcohol and drugs influence driving behavior by affecting the central nervous system, awareness, vision, and perception/reaction times, but the resulting effect on driver injuries in car crashes is not fully understood. The purpose of this study was to identify factors affecting the injury severities of unimpaired, alcohol-impaired, and drug-impaired drivers. METHOD: The current article applies a random parameters logit model to study the differences in injury severities among unimpaired, alcohol-impaired, and drug-impaired drivers. Using data from single-vehicle crashes in Cook County, Illinois, over a 9-year period from January 1, 2004, to December 31, 2012, separate models for unimpaired, alcohol-impaired, and drug-impaired drivers were estimated. A wide range of variables potentially affecting driver injury severity was considered, including roadway and environmental conditions, driver attributes, time and location of the crash, and crash-specific factors. RESULTS: The estimation results show significant differences in the determinants of driver injury severities across groups of unimpaired, alcohol-impaired, and drug-impaired drivers. The findings also show that unimpaired drivers are understandably more responsive to variations in lighting, adverse weather, and road conditions, but these drivers also tend to have much more heterogeneity in their behavioral responses to these conditions, relative to impaired drivers. In addition, age and gender were found to be important determinants of injury severity, but the effects varied significantly across all drivers, particularly among alcohol-impaired drivers. CONCLUSIONS: The model estimation results show that statistically significant differences exist in driver injury severities among the unimpaired, alcohol-impaired, and drug-impaired driver groups considered. Specifically, we find that unimpaired drivers tend to have more heterogeneity in their injury outcomes in the presence potentially adverse weather and road surface conditions. This makes sense because one would expect unimpaired drivers to apply their full knowledge/judgment range to deal with these conditions, and the variability of this range across the driver population (with different driving experiences, etc.) should be great. In contrast, we find, for the most part, that alcohol-impaired and drug-impaired drivers have far less heterogeneity in the factors that affect injury severity, suggesting an equalizing effect resulting from the decision-impairing substance.

An empirical analysis of risk-taking in car driving and other aspects of life.

The link between risk-taking behavior in various aspects of life has long been an area of debate among economists and psychologists. Using an extensive data set from Denmark, this study provides an empirical investigation of the link between risky driving and risk taking in other aspects of life, including risk-taking behavior in financial and labor-market decisions. Specifically, we establish significant positive correlations between individuals' risk-taking behavior in car driving and their risk-taking behavior in financial and labor-market decisions. However, we find that the strength of these correlations vary significantly between genders, and across risk decisions. These correlations and their differences across genders get stronger when we construct more "homogenous" groups by restricting our sample to those individuals with at least some stock-market participation. Overall, the empirical results in this study suggest that risk-taking behavior in various aspects of life can be associated, and our results corroborate previous evidence on the link between individuals' risk preferences across various aspects of life. This implies that individuals' driving behavior, which is commonly unobservable, can be more fully understood using observable labor market and financial decisions of individuals.

Safety impacts of signal-warning flashers and speed control at high-speed signalized intersections.

For many years, to reduce the crash frequency and severity at high-speed signalized intersections, warning flashers have been used to alert drivers of potential traffic-signal changes. Recently, more aggressive countermeasures at such intersections include a speed-limit reduction in addition to warning flashers. While such speed-control strategies have the potential to further improve the crash-mitigation effectiveness of warning flashers, a rigorous statistical analysis of crash data from such intersections has not been undertaken to date. This paper uses 10-year crash data from 28 intersections in Nebraska (all with intersection approaches having signal-warning flashers; some with no speed-limit reduction, and the others with either 5 mi/h or 10 mi/h reduction in speed limit) to estimate a random parameters negative binomial model of crash frequency and a nested logit model of crash-injury severity. The estimation findings show that, while a wide variety of factors significantly influence the frequency and severity of crashes, the effect of the 5 mi/h speed-limit reduction is ambiguous-decreasing the frequency of crashes on some intersection approaches and increasing it on others, and decreasing some crash-injury severities and increasing others. In contrast, the 10 mi/h reduction in speed limit unambiguously decreased both the frequency and injury-severity of crashes. It is speculated that, in the presence of potentially heterogeneous driver responses to decreased speed limits, the smaller distances covered during reaction time at lower speeds (allowing a higher likelihood of crash avoidance) and the reduced energy of crashes associated with lower speed limits are not necessarily sufficient to unambiguously decrease the frequency and severity of crashes when the speed-limit reduction is just 5 mi/h. However, they are sufficient to unambiguously decrease the frequency and severity of crashes when the speed-limit reduction is 10 mi/h. Based on this research, speed-limit reductions in conjunction with signal-warning flashers appear to be an effective safety countermeasure, but only clearly so if the speed-limit reduction is at least 10 mi/h.

A study of factors affecting highway accident rates using the random-parameters tobit model.

A large body of previous literature has used a variety of count-data modeling techniques to study factors that affect the frequency of highway accidents over some time period on roadway segments of a specified length. An alternative approach to this problem views vehicle accident rates (accidents per mile driven) directly instead of their frequencies. Viewing the problem as continuous data instead of count data creates a problem in that roadway segments that do not have any observed accidents over the identified time period create continuous data that are left-censored at zero. Past research has appropriately applied a tobit regression model to address this censoring problem, but this research has been limited in accounting for unobserved heterogeneity because it has been assumed that the parameter estimates are fixed over roadway-segment observations. Using 9-year data from urban interstates in Indiana, this paper employs a random-parameters tobit regression to account for unobserved heterogeneity in the study of motor-vehicle accident rates. The empirical results show that the random-parameters tobit model outperforms its fixed-parameters counterpart and has the potential to provide a fuller understanding of the factors determining accident rates on specific roadway segments.

A multivariate tobit analysis of highway accident-injury-severity rates.

Relatively recent research has illustrated the potential that tobit regression has in studying factors that affect vehicle accident rates (accidents per distance traveled) on specific roadway segments. Tobit regression has been used because accident rates on specific roadway segments are continuous data that are left-censored at zero (they are censored because accidents may not be observed on all roadway segments during the period over which data are collected). This censoring may arise from a number of sources, one of which being the possibility that less severe crashes may be under-reported and thus may be less likely to appear in crash databases. Traditional tobit-regression analyses have dealt with the overall accident rate (all crashes regardless of injury severity), so the issue of censoring by the severity of crashes has not been addressed. However, a tobit-regression approach that considers accident rates by injury-severity level, such as the rate of no-injury, possible injury and injury accidents per distance traveled (as opposed to all accidents regardless of injury-severity), can potentially provide new insights, and address the possibility that censoring may vary by crash-injury severity. Using five-year data from highways in Washington State, this paper estimates a multivariate tobit model of accident-injury-severity rates that addresses the possibility of differential censoring across injury-severity levels, while also accounting for the possible contemporaneous error correlation resulting from commonly shared unobserved characteristics across roadway segments. The empirical results show that the multivariate tobit model outperforms its univariate counterpart, is practically equivalent to the multivariate negative binomial model, and has the potential to provide a fuller understanding of the factors determining accident-injury-severity rates on specific roadway segments.

The statistical analysis of highway crash-injury severities: a review and assessment of methodological alternatives.

Reducing the severity of injuries resulting from motor-vehicle crashes has long been a primary emphasis of highway agencies and motor-vehicle manufacturers. While progress can be simply measured by the reduction in injury levels over time, insights into the effectiveness of injury-reduction technologies, policies, and regulations require a more detailed empirical assessment of the complex interactions that vehicle, roadway, and human factors have on resulting crash-injury severities. Over the years, researchers have used a wide range of methodological tools to assess the impact of such factors on disaggregate-level injury-severity data, and recent methodological advances have enabled the development of sophisticated models capable of more precisely determining the influence of these factors. This paper summarizes the evolution of research and current thinking as it relates to the statistical analysis of motor-vehicle injury severities, and provides a discussion of future methodological directions.

The effects of road-surface conditions, age, and gender on driver-injury severities.

Drivers' adaptation to weather-induced changes in roadway-surface conditions is a complex process that can potentially be influenced by many factors including age and gender. Using a mixed logit analysis, this research assesses the effects that age, gender, and other factors have on crash severities by considering single-vehicle crashes that occurred on dry, wet, and snow/ice-covered roadway surfaces. With an extensive database of single-vehicle crashes from Indiana in 2007 and 2008, estimation results showed that there were substantial differences across age/gender groups under different roadway-surface conditions. For example, for all females and older males, the likelihood of severe injuries increased when crashes occurred on wet or snow/ice surfaces-but for male drivers under 45 years of age, the probability of severe injuries decreased on wet and snow/ice surfaces - relative to dry-surface crashes. This and many other significant differences among age and gender groups suggest that drivers perceive and react to pavement-surface conditions in very different ways, and this has important safety implications. Furthermore, the empirical findings of this study highlight the value of considering subsets of data to unravel the complex relationships within crash-injury severity analysis.

An empirical assessment of fixed and random parameter logit models using crash- and non-crash-specific injury data.

Traditional crash-severity modeling uses detailed data gathered after a crash has occurred (number of vehicles involved, age of occupants, weather conditions at the time of the crash, types of vehicles involved, crash type, occupant restraint use, airbag deployment, etc.) to predict the level of occupant injury. However, for prediction purposes, the use of such detailed data makes assessing the impact of alternate safety countermeasures exceedingly difficult due to the large number of variables that need to be known. Using 5-year data from interstate highways in Indiana, this study explores fixed and random parameter statistical models using detailed crash-specific data and data that include the injury outcome of the crash but not other detailed crash-specific data (only more general data are used such as roadway geometrics, pavement condition and general weather and traffic characteristics). The analysis shows that, while models that do not use detailed crash-specific data do not perform as well as those that do, random parameter models using less detailed data still can provide a reasonable level of accuracy.

A note on modeling pedestrian-injury severity in motor-vehicle crashes with the mixed logit model.

Pedestrian-injury severity has been traditionally modeled with approaches that have assumed that the effect of each variable is fixed across injury observations. This assumption ignores possible unobserved heterogeneity which is likely to be particularly important in pedestrian injuries because unobserved physical health, strength, and behavior may significantly affect the pedestrians' ability to absorb collision forces. To address such unobserved heterogeneity, this research applies a mixed logit model to analyze pedestrian-injury severity in pedestrian-vehicle crashes. Using police-reported collision data from 1997 through 2000 from North Carolina, several factors were found to more than double the average probability of fatal injury for pedestrians in motor-vehicle crashes including: darkness without streetlights (400% increase in fatality probability), vehicle is a truck (370% increase), freeway (330% increase), speeding involved (360% increase), and collisions involving a motorist who had been drinking (250% increase). It was also found that the effect of pedestrian age was normally distributed across observations, and that as pedestrians became older the probability of fatal injury increased substantially. Heterogeneity in the mean of the random parameters for the freeway and pedestrian-solely-at-fault collision indicators was related to pedestrian gender, and heterogeneity in the mean of the random parameters for the traffic-sign and motorist-back-up indicators was related to pedestrian age.

Zero-state Markov switching count-data models: an empirical assessment.

In this study, a two-state Markov switching count-data model is proposed as an alternative to zero-inflated models to account for the preponderance of zeros sometimes observed in transportation count data, such as the number of accidents occurring on a roadway segment over some period of time. For this accident-frequency case, zero-inflated models assume the existence of two states: one of the states is a zero-accident count state, which has accident probabilities that are so low that they cannot be statistically distinguished from zero, and the other state is a normal-count state, in which counts can be non-negative integers that are generated by some counting process, for example, a Poisson or negative binomial. While zero-inflated models have come under some criticism with regard to accident-frequency applications - one fact is undeniable - in many applications they provide a statistically superior fit to the data. The Markov switching approach we propose seeks to overcome some of the criticism associated with the zero-accident state of the zero-inflated model by allowing individual roadway segments to switch between zero and normal-count states over time. An important advantage of this Markov switching approach is that it allows for the direct statistical estimation of the specific roadway-segment state (i.e., zero-accident or normal-count state) whereas traditional zero-inflated models do not. To demonstrate the applicability of this approach, a two-state Markov switching negative binomial model (estimated with Bayesian inference) and standard zero-inflated negative binomial models are estimated using five-year accident frequencies on Indiana interstate highway segments. It is shown that the Markov switching model is a viable alternative and results in a superior statistical fit relative to the zero-inflated models.

Empirical assessment of the impact of highway design exceptions on the frequency and severity of vehicle accidents.

Compliance to standardized highway design criteria is considered essential to ensure roadway safety. However, for a variety of reasons, situations arise where exceptions to standard-design criteria are requested and accepted after review. This research explores the impact that such design exceptions have on the frequency and severity of highway accidents in Indiana. Data on accidents at carefully selected roadway sites with and without design exceptions are used to estimate appropriate statistical models of the frequency and severity of accidents at these sites using recent statistical advances with mixing distributions. The results of the modeling process show that presence of approved design exceptions has not had a statistically significant effect on the average frequency or severity of accidents - suggesting that current procedures for granting design exceptions have been sufficiently rigorous to avoid adverse safety impacts. However, the findings do suggest that the process that determines the frequency of accidents does vary between roadway sites with design exceptions and those without.

Markov switching multinomial logit model: An application to accident-injury severities.

In this study, two-state Markov switching multinomial logit models are proposed for statistical modeling of accident-injury severities. These models assume Markov switching over time between two unobserved states of roadway safety as a means of accounting for potential unobserved heterogeneity. The states are distinct in the sense that in different states accident-severity outcomes are generated by separate multinomial logit processes. To demonstrate the applicability of the approach, two-state Markov switching multinomial logit models are estimated for severity outcomes of accidents occurring on Indiana roads over a four-year time period. Bayesian inference methods and Markov Chain Monte Carlo (MCMC) simulations are used for model estimation. The estimated Markov switching models result in a superior statistical fit relative to the standard (single-state) multinomial logit models for a number of roadway classes and accident types. It is found that the more frequent state of roadway safety is correlated with better weather conditions and that the less frequent state is correlated with adverse weather conditions.

Markov switching negative binomial models: an application to vehicle accident frequencies.

In this paper, two-state Markov switching models are proposed to study accident frequencies. These models assume that there are two unobserved states of roadway safety, and that roadway entities (roadway segments) can switch between these states over time. The states are distinct, in the sense that in the different states accident frequencies are generated by separate counting processes (by separate Poisson or negative binomial processes). To demonstrate the applicability of the approach presented herein, two-state Markov switching negative binomial models are estimated using five-year accident frequencies on Indiana interstate highway segments. Bayesian inference methods and Markov Chain Monte Carlo (MCMC) simulations are used for model estimation. The estimated Markov switching models result in a superior statistical fit relative to the standard (single-state) negative binomial model. It is found that the more frequent state is safer and it is correlated with better weather conditions. The less frequent state is found to be less safe and to be correlated with adverse weather conditions.

A note on modeling vehicle accident frequencies with random-parameters count models.

In recent years there have been numerous studies that have sought to understand the factors that determine the frequency of accidents on roadway segments over some period of time, using count data models and their variants (negative binomial and zero-inflated models). This study seeks to explore the use of random-parameters count models as another methodological alternative in analyzing accident frequencies. The empirical results show that random-parameters count models have the potential to provide a fuller understanding of the factors determining accident frequencies.