A Risk-Based Method for Modeling Traffic Fatalities

We describe a risk-based analytical framework for estimating traffic fatalities that combines the probability of a crash and the probability of fatality in the event of a crash. As an illustrative application, we use the methodology to explore the role of vehicle mix and vehicle prevalence on long-run fatality trends for a range of transportation growth scenarios that may be relevant to developing societies. We assume crash rates between different road users are proportional to their roadway use and estimate case fatality ratios (CFRs) for the different vehicle-vehicle and vehicle-pedestrian combinations. We find that in the absence of road safety interventions, the historical trend of initially rising and then falling fatalities observed in industrialized nations occurred only if motorization was through car ownership. In all other cases studied (scenarios dominated by scooter use, bus use, and mixed use), traffic fatalities rose monotonically. Fatalities per vehicle had a falling trend similar to that observed in historical data from industrialized nations. Regional adaptations of the model validated with local data can be used to evaluate the impacts of transportation planning and safety interventions, such as helmets, seat belts, and enforcement of traffic laws, on traffic fatalities.     

The Dynamic Nature of Risk Perceptions After a Fatal Transit Accident

In 2009, two trains of Washington, DC's Metrorail system collided, resulting in nine deaths and 50 serious injuries. Based on a multiwave survey of Metrorail users in the months after the crash, this article reports how the accident appears to have (1) changed over time the tradeoffs among safety, speed, frequency of service, cost, and reliability that the transit users stated they were willing to make in the postaccident period and (2) altered transit users’ concerns about safety as a function of time and distance from the accident site. We employ conditional logit models to examine tradeoffs among stated preferences for system performance measures after the accident, as well as the influence that respondent characteristics of transit use, location, income, age, and gender have on these preference tradeoffs. As expected, respondents appear averse to longer headways between trains, longer travel durations, higher travel costs, a higher number of late trains, and a higher number of fatalities. The models also show evidence of higher aversion to fatalities from transit system operation among females compared to males. In addition, respondents less experienced with Metrorail travel and those with lower household incomes show higher aversion to fatalities, and this aversion increases as a subject's psychological distance from the accident site decreases. Contrary to expectations shaped by previous studies, aversion to fatalities appears to have increased between the early months after the accident and the end of the survey period, and the expected relationship between age and aversion to fatalities is not statistically significant.     

Real-World Personal Conversations Using a Hands-Free Embedded Wireless Device While Driving: Effect on Airbag-Deployment Crash Rates

A wireless device embedded in the vehicle allowed the user to engage in a personal hands-free conversation (HFC), and automatically placed an emergency notification call to an OnStar call center if the vehicle was involved in a crash in which its airbag deployed. A database stored the exact counts, start timestamps, and billed durations of all HFC and airbag notification calls. In 30 months of naturalistic driving, there were 91 million HFC calls from an average of 323,994 drivers per month who made calls. There were 14 airbag deployments in 276 million driver-minutes of HFC conversation for an exposed incidence rate of 5.08 airbag crashes per 100 million driver-minutes. There were 2,023 airbag deployments in an estimated 24.7 billion driver-minutes of no HFC conversation for a not-exposed incidence rate of 8.18 airbag crashes per 100 million driver-minutes. The crash incidence rate ratio (IRR) is the ratio of these two rates or 0.62 (95% C.I. 0.37 to 1.05). Sensitivity analyses controlled for the impact on the crash IRR of estimated time spent driving per day and calls by passengers. Counting all crashes as much as 20 minutes later than a call as related to that call gave similar results. We conclude that for personal conversations using a hands-free embedded device the risk of an airbag crash is somewhere in a range from a moderately lower risk to a risk near that of driving without a recent personal conversation. These results are not consistent with the large increase in crash risk reported in epidemiological studies using the case-crossover method.     

General Deterrence of Drunk Driving: Evaluation of Recent American Policies

A testable hypothesis of deterrence theory is that efforts to increase the expected cost of criminal activity by increasing the threat of punishment should, other things being equal, reduce the crime rate. In this paper, we examine whether the incidence of drinking and driving is responsive to escalation of the punitive threat. The recent national campaign against drunk driving provides a natural experiment in which to test the predictions of deterrence theory. Using state level data over the 1975-1986 period, we report no conclusive evidence that any specific form of punitive legislation is having a measurable effect on motor vehicle fatalities. We report suggestive evidence that multiple laws designed to increase the certainty of punishment (e.g., sobriety checkpoints and preliminary breath tests) have a synergistic deterrent effect. The most striking finding is that mandatory seat belt use laws and beer taxes may be more effective at reducing drunk driving fatalities than policies aimed at general deterrence.     

Trait Predictors of Aggression and Crash‐Related Behaviors Across Drivers from the United Kingdom and the Irish Republic

Aggressive driving is acknowledged as a contributor to motor vehicle crashes. This study explored a theoretical model of aggressive expression and crash‐related outcomes using self‐report data collected, using an online questionnaire, from drivers in the United Kingdom and the Republic of Ireland. The proposed model tested whether the personality traits of boredom proneness, sensation seeking, and impulsivity, coupled with trait driving anger, predicted aggressive driving; and whether aggressive driving predicted crash‐related outcomes (loss of concentration and control, near misses, and moving violations). The structural model was confirmed, with aggressive expressions of anger being found to mediate the relationships driving anger and impulsivity had with the crash‐related outcomes. Multigroup invariance analysis showed that the model remained invariant across drivers from the United Kingdom and Ireland, suggesting that the contributing factors for aggressive expression and crash involvement are similar across both countries. When self‐reported crash‐related conditions were compared between drivers in the United Kingdom and Ireland, drivers in the United Kingdom reported more aggressive driving, more minor crashes, more incidents of road rage, and more frequent losses of concentration and vehicle control.     

Exposure to Traffic and Risk of Hospitalization Due to Injuries

Research on the risk of motor vehicle injuries and their relationship with the amount of travel has been only partially analyzed. The few individual exposure assessments are related to very specific subsets of the driving and traveling populations. This study analyzes the relationship between kilometers traveled and hospitalization due to motor vehicle injuries. Twelve thousand three hundred and sixty nine Spanish university graduates from the Seguimiento Universidad de Navarra multipurpose cohort study were evaluated. They had not been hospitalized due to motor vehicle injuries at baseline and were followed up to eight years. Biannual questionnaires allowed for self‐reporting of kilometers traveled in motor vehicles, together with incidence of hospitalization. Covariates in the Cox regression models included age and gender and baseline use of safety belt while driving, driving a vehicle with driver‐side airbag, driving a motorcycle, and drinking and driving. There were 49,766 participant‐years with an average yearly travel of 7,828 km per person‐year. Thirty‐six subjects reported a first hospitalization event during this time. The adjusted hazard ratio per additional kilometer traveled was 1.00005 (95% confidence interval 1.000013 to 1.000086). Even the smallest of reductions in the amount of kilometers traveled (from an average of 3,250 km per year to 1,000) has a statistically significant protective effect on the likelihood of sustaining hospitalization due to motor vehicle injury (aHR 0.9, 95% CI 0.78 to 0.98). In light of current policies aimed to reduce motorized traffic due to environmental concerns, it may be appropriate to consider the additional health benefit related to reductions in injuries.     

Should Persons with Diabetes Be Licensed to Drive Trucks?—Risk Management

How should a regulatory agency interpret a risk analysis that concludes there is a small increase in risk? The agency must decide on behalf of society whether the increased risk is large enough to justify banning the risky activity or taking some other step to lessen the risks. In a companion paper (Songer et al.), we conclude that licensing insulin using persons to drive commercial motor vehicles in interstate commerce would result in 42 additional crashes each year. Here we address risk management issues by interpreting the number of additional crashes and the relative risks of the prospective handicapped drivers. Are the number of additional crashes (42) significant? Is the increase in the annual crash risk (from 0.00785 to 0.032 for non-insulin dependent and 0.048 for insulin dependent persons) significant? Are the relative risks significant for all insulin using drivers (4.7)? For drivers with a history of severe hypoglycemic reactions (19.8)? How should society tradeoff risk increases for increases in opportunity for these handicapped persons? We review other social decisions concerning highway safety: Accepting the increasing risks of letting 16 year olds drive, allowing extremely light cars, allowing some unsafe highways, and allowing extremely unsafe driving conditions at some times of day. We conclude that the additional risks from insulin using persons are well within the current accepted range of risks. Currently, 70% of states permit insulin using persons to drive trucks within their state. Nonetheless, the social cost, due to fatalities, injuries, and property damage from allowing a person with a history of severe hypoglycemic reactions to drive is more than $19,700 per year.     

Should I Text or Call Here? A Situation‐Based Analysis of Drivers’ Perceived Likelihood of Engaging in Mobile Phone Multitasking

This study investigated how situational characteristics typically encountered in the transport system influence drivers’ perceived likelihood of engaging in mobile phone multitasking. The impacts of mobile phone tasks, perceived environmental complexity/risk, and drivers' individual differences were evaluated as relevant individual predictors within the behavioral adaptation framework. An innovative questionnaire, which includes randomized textual and visual scenarios, was administered to collect data from a sample of 447 drivers in South East Queensland‐Australia (66% females; n = 296). The likelihood of engaging in a mobile phone task across various scenarios was modeled by a random parameters ordered probit model. Results indicated that drivers who are female, are frequent users of phones for texting/answering calls, have less favorable attitudes towards safety, and are highly disinhibited were more likely to report stronger intentions of engaging in mobile phone multitasking. However, more years with a valid driving license, self‐efficacy toward self‐regulation in demanding traffic conditions and police enforcement, texting tasks, and demanding traffic conditions were negatively related to self‐reported likelihood of mobile phone multitasking. The unobserved heterogeneity warned of riskier groups among female drivers and participants who need a lot of convincing to believe that multitasking while driving is dangerous. This research concludes that behavioral adaptation theory is a robust framework explaining self‐regulation of distracted drivers.     

The Risks of Licensing Persons with Diabetes to Drive Trucks

The 1990 Americans with Disabilities Act forbids employers to bar disabled persons from jobs unless employers can show the disabled person cannot perform the tasks. The Federal Highway Administration will not license persons with diabetes mellitus to drive commercial motor vehicles in interstate commerce. These individuals may experience severe hypoglycemia, greatly increasing their risk of losing control of the truck. This prohibition is currently being reexamined. We describe the disease process leading to severe hypoglycemia and its physical manifestations. To quantify the risks of licensing persons with diabetes to use insulin, we first estimate the number of potential insulin-using drivers. We estimate that 1420 insulin-using persons would seek licenses in the United States if they were permitted to do so (920 noninuslin dependent and 500 insulin dependent). Next, we estimate the annual incidence of mild and severe hypoglycemia in these populations. The third step is to estimate the number of hypoglycemic episodes while driving. Estimating the likelihood of a crash due to a mild or severe hypoglycemic episode is the fourth step. We estimate that an additional 42 crashes each year would occur if insulin using persons were licensed to drive commercial motor vehicles in interstate commerce (20 from insulin dependent and 22 from non-insulin dependent drivers).     

Estimating Fatality Reductions from Increased Safety Belt Use

Fatality reductions from increases in safety belt use are estimated taking into account that drivers who change from being nonusers to being users have lower accident involvement rates than the remaining nonusers, a process referred to as "selective recruitment." Analytical functions are derived which express expected fatality reductions in terms of changes in safety belt use rates from an initial rate. The function parameters are determined by requiring that computed average crash rates for nonusers be 53% higher than the rates for users, a recently determined empirical value. These functions show that, depending on the initial use rate and use rate increase, selective recruitment may increase or decrease expected fatality reductions. However, effects are relatively small, in no case exceeding +/- 5.3%.     

Cost-Benefit Analysis of All-Terrain Vehicles at the CPSC1

The present value of the costs of injuries and deaths associated with each unit of a product is a useful number in conducting cost-benefit analysis at the CPSC. Before a remedy is known, this value indicates the approximate amount that it is worth spending on a fix; similarly, if a fix is proposed, this number indicates whether it is cost-justified. The method of analysis is applied to all-terrain vehicles. Injury costs per vehicle are estimated. Determinants of accidents are estimated using a logit regression model applied to surveys of users and of accident victims; this indicates factors associated with accidents, and is used to estimate marginal probabilities. Costs of various vehicle components are estimated using a hedonic price index. This information (injury costs, accident probabilities, and component cost) is combined in a benefit–cost analysis of proposed remedies.     

Safety Impacts of SUVs, Vans, and Pickup Trucks in Two-Vehicle Crashes

Policy makers, vehicle manufacturers, and consumers have shown growing concern about the relative safety of sport utility vehicles (SUVs), vans, pickups, and cars. Empirical analysis of real-world crashes is complicated by the possibility that apparent relationships between vehicle type and safety may be confounded by other factors, such as driver behavior and crash circumstances. This study compares different vehicle types with respect to their crashworthiness (self-protection) and aggressivity (risk to others) in crashes between two passenger vehicles. The U.S. Crashworthiness Data System is used to analyze detailed information on 6,481 drivers involved in crashes during 1993-1999. Logistic regression analysis is used to model the risk of serious injury or death to a driver, conditional on a crash occurring. Covariates include the body type of each vehicle in the crash; the driver's age, gender, and restraint use; and the configuration of the crash. A unique feature of this study is the use of "delta-v" to represent the joint effects of vehicle mass and crash severity. While estimated effects are somewhat sensitive to the injury severity level used as the outcome variable, SUVs, vans, and pickups appear to be more aggressive and may be more crashworthy than cars. Effects of pickups are most pronounced. Drivers in pickups face less risk of serious injury than car drivers (odds ratio [OR], 0.35; 95% confidence interval [CI], 0.20-0.60), and drivers who collide with pickups experience more than twice the risk than those who collide with a car (OR, 2.18; 95% CI, 1.03-4.62). While vehicle mass and crash severity contribute to the apparent crashworthiness and aggressivity of passenger vehicles, other vehicle characteristics associated with body type (e.g., the stiffness and height of the underlying structure of the vehicle) also influence safety risks.     

Chlorine Truck Attack Consequences and Mitigation

We develop and apply an integrated modeling system to estimate fatalities from intentional release of 17 tons of chlorine from a tank truck in a generic urban area. A public response model specifies locations and actions of the populace. A chemical source term model predicts initial characteristics of the chlorine vapor and aerosol cloud. An atmospheric dispersion model predicts cloud spreading and movement. A building air exchange model simulates movement of chlorine from outdoors into buildings at each location. A dose‐response model translates chlorine exposures into predicted fatalities. Important parameters outside defender control include wind speed, atmospheric stability class, amount of chlorine released, and dose‐response model parameters. Without fast and effective defense response, with 2.5 m/sec wind and stability class F, we estimate approximately 4,000 (half within ～10 minutes) to 30,000 fatalities (half within ～20 minutes), depending on dose‐response model. Although we assume 7% of the population was outdoors, they represent 60–90% of fatalities. Changing weather conditions result in approximately 50–90% lower total fatalities. Measures such as sheltering in place, evacuation, and use of security barriers and cryogenic storage can reduce fatalities, sometimes by 50% or more, depending on response speed and other factors.     

Cellular Telephones and Driving Performance: The Effects of Attentional Demands on Motor Vehicle Crash Risk

This study examines the effects of conversation mode and split-attention communication training on driving performance. The study is based on an experiment where drivers with and without communication training (pilots vs. nonpilots) completed a simulated driving course while involved in one of three conversation modes: no conversation, conversation with passenger, or conversation on a hands-free cellular telephone. Results indicate that cellular telephone conversations consume more attention and interfere more with driving than passenger conversations. Cell phone conversations lack the nonverbal cues available during close-contact conversations and conversation participants expend significant cognitive resources to compensate for the lack of such cues. The results also demonstrate that communication training may reduce the hazardous effects of cell phone conversations on driving performance.     

Should Cities Invest in Sheltering‐in‐Place Measures Against Chlorine Truck Attacks by Terrorists?

After an intentional release of chlorine in an office district, public responses such as sheltering‐in‐place could save many lives if rapid enough. However, previous work does not estimate how fast and effective such responses would be for several possible investments in attack detection, public alert, and building ventilation, nor whether such measures would be cost effective. We estimate public response times with investment options in place, and resulting changes in fatalities as well as system costs, including false alarm costs, and cost effectiveness in terms of cost per net death avoided. The measures do have life‐saving potential, especially if all response times are at or near the lower limits of the ranges assumed in this article. However, due to uncertainties, it is not clear that responses would be rapid enough to save many people. In some cases total fatalities would increase, since sheltering after chlorine vapor has already entered buildings can increase occupants’ chlorine exposure. None of the options considered have median cost per statistical life saved meeting a cost‐effectiveness threshold of $6.5 million across all of the chlorine exposure dose‐response and ingress‐delay models considered here, even if there were one attack per year in the area covered by the system. Given these and other issues discussed in this article, at this point investments to improve sheltering‐in‐place capability appear not to be robust strategies for reducing fatalities from chlorine attack in an office district.     

Standards, critical mass, and the formation of complex industries: A case study of the mobile Internet

This is one of the few papers to analyze multiple interface standards in a single industry and it finds that standard setting in such an industry is much more complex than those covered in the existing literature. In the mobile Internet, some of the interface standards initially required so-called integral design while others have been built on top of these “basic” interface standards. The former interface standards enable basic data connections between phones, services, and content while the latter ones connect the mobile phone with content and applications from other industries such as music, video, publishing, broadcasting, and payment. This paper shows that in connecting the mobile phone and other industries, each critical mass of phones, services, and content partly builds from previously created critical masses.     

Risk Perceptions of Mobile Phone Use While Driving

Numerous governments have introduced regulations governing the use of mobile (cellular) telephones while driving. Despite significant research into the "objective" risks, there is relatively little research into risk perceptions either in relation to other in-car distractions or with respect to the factors underlying such perceptions. The current article reports on two studies addressing these issues. Study 1 ( N = 199) found that whereas the use of hand-held sets is seen as one of the riskiest activities to perform while driving, the risks of using a hands-free kit are perceived to be relatively small. Study 2 ( N = 1,320) found that nearly half of all drivers in the sample with a mobile phone reported having used it while driving and that, overall, the probability of having an accident was perceived to be less for oneself than for one's peers, indicating an optimistic bias. Two factors underpinned risk perceptions, "impact" including perceived severity and equitability, and "controllability" including immediacy, detectability, and probability. While higher "impact" scores were associated with increased preferences for restrictions on the use of hand-held mobiles while driving, the "controllability" scores moderated this relationship such that when perceived "controllability" was low, restriction preferences were high irrespective of perceived "impact." However, when "controllability" was high, restriction preferences remained high when "impact" was high but were low when "impact" was low. Given the growing number of in-car technological innovations, it is suggested that regulators act strategically, rather than finding themselves developing a series of "hazard-specific" regulations, which may ultimately lack coherence.     

What Happens When Manufacturers Perform The Retailing Functions?

This study examines the effects of a relatively new channel structure on prices and sales in a large department store, which in recent years has switched the management of many of its product categories from a traditional retailer‐managed system to a manufacturer‐managed system. We find that the change caused overall retail prices to decrease. However, there was significant heterogeneity in the response across brands. In the cell phone category, brands with high market shares and inelastic demand did not change prices. In the watch category, the retail prices of relatively low‐end brands decreased while the prices of premium brands increased substantially after the switch. In addition to sales increases due to lower prices, we find that the channel structure change further caused sales to increase by 9–10% in the cell phone category and by 11–17% in the watch category. These results are consistent with previous theoretical predictions. We believe that our results provide important academic and managerial implications due to the increasing prevalence of manufacturer‐managed systems in the retail industry.     

The efficacy of using judgmental versus quantitative forecasting methods in practice

In an era where forecasts drive entire supply chains forecasting is seen as an increasingly critical organizational capability. However, business forecasting continues to rely on judgmental methods despite large advancements in information technology and quantitative method capability, prompting calls for research to help understand the reasons behind this practice. Our study is designed to contribute to this knowledge by profiling differences between firms identified as primary users of either judgmental or quantitative forecasting methods. Relying on survey data from 240 firms we statistically analyzed differences between these categories of users based on a range of organizational and forecasting issues. Our study finds large differences in forecast error rates between the two groups, with users of quantitative methods significantly outperforming users of judgmental methods. The former are found to be equally prevalent regardless of industry, firm size, and product positioning strategy, documenting the benefits of quantitative method use in a variety of settings. By contrast, the latter are found to have significantly lower access to quantifiable data and to use information and technology to a lesser degree.     

Physical injury risks associated with drinking water arsenic treatment.

We estimated the number of transportation deaths that would be associated with water treatment in Albuquerque to meet the EPA's recently proposed revisions of the Maximum Contaminant Level (MCL) for arsenic. Vehicle mileage was estimated for ion exchange, activated alumina, and iron coagulation/microfiltration water treatment processes to meet an MCL of 0.020 mg/L, 0.010 mg/L, 0.005 mg/L, and 0.003 mg/L. Local crash, injury, and death rates per million vehicle miles were used to estimate the number of injuries and deaths. Depending on the water treatment options chosen, we estimate that meeting an arsenic MCL of 0.005 mg/L will result in 143 to 237 crashes, 58 to 98 injuries, and 0.6 to 2.6 deaths in Albuquerque over a 70-year period, resulting in 26 to 113 years of life lost. The anticipated health benefits for Albuquerque residents from a 0.005 mg/L arsenic MCL, estimated using either a multistage Weibull or Poisson model, ranged from 3 to 80 arsenic-related bladder and lung cancer deaths prevented over a 70-year period, adding between 43 and 1,123 years of life. Whether a revised arsenic MCL increases or reduces overall loss of life in Albuquerque depends on the accuracy of EPA's cancer risk assessment. If the multistage Weibull model accurately estimates the benefits, the years of life added is comparable or lower than the anticipated years lost due to transportation associated with the delivery of chemicals, disposal of treatment waste, and operation of the water treatment system. Coagulation/microfiltration treatment will result in substantially fewer transportation deaths than either ion exchange or activated alumina.