Penmetsa and Pulugurtha

1

RISK DRIVERS POSE TO THEMSELVES DUE TO TRAFFIC VIOLATIONS

Praveena Penmetsa, M.Tech. Ph.D. Student, Department of Civil and Environmental Engineering The University of North Carolina at Charlotte, North Carolina 28223, United States of America Email: [email protected]

Srinivas S. Pulugurtha, Ph.D., P.E. Professor, Department of Civil and Environmental Engineering The University of North Carolina at Charlotte, North Carolina 28223, United States of America Email: [email protected]

The following is a pre-print, the final publication can be found in Traffic Injury Prevention, 18 (1): 6369, 2016 or http://www.tandfonline.com/doi/abs/10.1080/15389588.2016.1177637

ABSTRACT Objective: Violation of traffic rules is a major contributing factor of both crashes and fatalities in the United States. This study aims at examining crashes due to traffic violations and quantifying risk drivers pose to themselves by violating traffic rules. Method: The crash data from 2009 to 2011 was collected for the entire state of North Carolina. Descriptive statistics was carried out to identify who were committing the traffic violations that resulted in crashes. Multinomial regression model was then developed to examine the relation between different traffic violations and driver injury severity. Additionally, odds ratio was estimated to identify the likelihood (probability) of severe or moderate injury to the driver after committing a traffic violation that led to a crash. Results: Exceeding speed limit is 182 times more likely to result in severe driver injuries compared to driver injury in no traffic violation crashes. Passing on a curve (in no passing zones), driving under the influence of alcohol, going in wrong way, and aggressive driving are approximately 50 times more likely to result in severe driver injury. Disregarding stop sign is 20 and 4 times more likely to result in severe injury and moderate injury to the driver, respectively. Overall, most of the crashes that occurred due to traffic violations result in severe driver injuries. Conclusions: Often times, drivers perceive that violating traffic rules does not result in a crash or severe injuries. However, the results from this study show that a majority of the traffic violations lead to severe driver injury. The findings from this study serve as documented evidence to educate drivers about the risk they pose to themselves by violating traffic rules and encourage adaptation of safe driving behavior in order to contribute towards reaching the “zero traffic deaths” vision.

Keywords: Traffic violation, driver injury, crash, risk, multinomial logit model, odds ratio, speeding.

Penmetsa and Pulugurtha INTRODUCTION Though the fatality rate per vehicle miles travelled (VMT) reduced over the years, 32,719 people died in motor vehicle crashes in the United States during 2013. North Carolina has reported 1,299 traffic deaths and more than 100,000 injuries in 2013 (NHTSA, 2014). It stands fourth in the number of fatalities when compared to other states. To improve safety and mobility on the roadways, policy makers and transportation system managers laid a set of traffic rules. Crash reports indicate that violation of these traffic rules is a major contributing factor of both crashes and fatalities in the United States. Crashes involving speeding and driving under the influence of alcohol together accounted for 58% of the total fatalities in the United States (NHTSA, 2015). During 1999 and 2000, around 1,990 and 1,294 people were killed at intersections for not obeying traffic signals and failing to yield the right-of-way, respectively in the United States (Campbell et al., 2004). According to Insurance Institute for Highway Safety (IIHS), red light violations are the leading cause of urban crashes in the United States. Numerous studies were carried out in the past on examining the role of driver characteristics in violating traffic rules or at specific locations. Moyano-Diaz (1997) measured the attitude of people towards traffic violations in Santiago. The study concluded that men are more risk takers than women and the same was stated by Yagil (1998). Machin and Sankey (2008) indicated that excitement, altruism, risk taking attitude, and their own likelihood of involving in a crash accounted for 39% of young drivers’ speeding. Braitman et al. (2007) identified that failing to yield the right of way increases with age. Abdel-Aty et al. (2003) used ordered probit models to analyze driver injury severity at different locations and concluded that drivers’ traffic violation of rules was significant in causing severe injuries at signalized intersections. Waller et al. (1986), Stoduto et al. (1993), Li et al. (1997), Cunningham et al. (2002) and Drummer et al. (2004) studied the effect of alcohol on driver injury severity in crashes. Retting et al. (2003) analyzed the motor vehicle crashes at two-way stop controlled intersections. Stop sign violations accounted for 70% of the total crashes considered in their study. Drivers with age less than 18 and greater than 65 were found at fault in crashes that occurred due to violating stop signs. Pai and Saleh (2008) examined motorcyclists’ injury severity at intersections using ordered probit models and concluded that motorcyclists are seriously injured when a right-turning vehicle failed to yield the right of way. Shankar and Mannering (1996), Carson and Mannering (2001), Khorashadi et al. (2005), Islam and Mannering (2006), Savolainen and Ghosh (2008), Malyshkina and Mannering (2008) and Geedipally et al. (2011) used multinomial logit models for assessing driver or crash severity. Ayuso et al. (2010) examined the influence of traffic violations on the likelihood of resulting in a serious or fatal crash using data for Spain. Some of their conclusions sound contradictory and not applicable to all countries or locations. Additionally, traffic violations such as right turn on red, improper lane change, improper lane use, operating defective equipment, driving under the influence of alcohol, and driving under the influence of drugs were not considered in their study. Overall, driver injury severity has been extensively studied for various crash types and under different situations. However, literature documents no research on the extent to which drivers are injured in crashes due to different traffic violations. Unarguably, traffic violators not only pose risk to

2

Penmetsa and Pulugurtha themselves but also to other road users. As normal people care or weigh themselves more than others, educating and creating awareness about potential crash risk they pose themselves due to violating traffic rules may lead to a reduction in the number of crashes and contribute towards reaching the “zero traffic deaths” vision. Therefore, this study aims at quantifying the risk drivers pose to themselves (in terms of driver injury severity) by violating traffic rules.

DATA AND ANALYTICAL METHOD Crash data was collected for the entire state of North Carolina from 2009 through 2011. The data was provided by the Highway Safety and Information System (HSIS). There are a total of 460,427 reported crashes during the three-year study period. The crash data file contains information related to driver, crash, vehicle, road, and environmental characteristics. The dependent variable in this study is severity of driver injury (i.e., who committed traffic violation). HSIS defines 5 levels of severity; fatal, incapacitating injury, nonincapacitating injury, possible injury, and property damage only (PDO). For this study, the severity of driver injury was defined in 3 levels. They are severe injury (fatal + incapacitating injury), moderate injury (non-incapacitating injury + capacitating injury), and PDO. The independent variable in this study is the contributing factor of a crash. HSIS provides 32 types of contributing factors, out of which a few are of interest for this study. They are explained in Table 1. The crash records were deleted if the contributing factor for the crash is other than those that were mentioned in Table 1. This final data set with 250,951 crash records was used for descriptive analysis and statistical modeling. S No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

TABLE 1 Independent Variables Considered in this Study Categories Description No traffic violation® If the crash occurred not due to traffic violation Disregarded yield sign If the crash occurred due to disregarding yield sign Disregarded stop sign If the crash occurred due to disregarding stop sign Disregarded other traffic If the crash occurred due to disregarding other traffic signs signs Disregarded traffic signal If the crash occurred due to disregarding traffic signal Disregarded road If the crash occurred due to disregarding road markings markings Exceeded authorized If the crash occurred due to exceeding authorized speed limit speed limit Exceeded safe speed If the crash occurred due to exceeding safe speed limit for conditions limit for conditions Improper turn If the crash occurred due to making an improper turn If the crash occurred due to maneuvering a right turn Right turn on red on red Going wrong way If the crash occurred due to going wrong way Improper lane change If the crash occurred due to improper lane change Improper lane use If the crash occurred due to improper lane use Passed on hill If the crash occurred due to passing on hill Passed on curve If the crash occurred due to passing on curve Other improper passing If the crash occurred due to other improper passing Improper or no signal If the crash occurred due to other improper or no

3

Penmetsa and Pulugurtha

4 signal

18 19

Improper backing Improper parking

20

Followed vehicle closely

If the crash occurred due to improper backing If the crash occurred due to improper parking If the crash occurred due to vehicle following the lead vehicle too closely

Operated vehicle If the crash occurred due to erratic or aggressive erratically or driving aggressively Operated defective If the crash occurred due to operating a defective 22 equipment equipment 23 Alcohol If the crash occurred due to involvement of alcohol 24 Drugs If the crash occurred due to involvement of drugs Note: ® indicates reference variable for the categorical variable. 21

In statistical models, the dependent variable can be categorical or continuous in nature. If the response variable is discrete, the error term cannot be normally distributed. In such scenarios, logistic regression models are preferred to evaluate the relationship between categorical dependent variable and continuous/dichotomous/categorical independent variables. A base or reference variable should be defined for any kind of logistic regression model. For this study, crashes that occurred not due to traffic violations were taken as reference. Multinomial logistic regression is a form of logistic regression model used when the dependent variable has more than two categories and does not have any meaningful order. The mathematical form of multinomial logit model (MNL) is shown in Equation 1. The following assumptions are made when deriving the mathematical form of MNL. ●

The error terms are Gumbel distributed.

●

The error components are identically and independently distributed across alternatives.

●

The

error

components

are

identically

and

independently

distributed

across

individuals/observations.

Pr(i) =

exp(Vi ) ∑Jj=1 Vj

(1)

where, Pr(i) is the probability of a individual choosing alternative i among set of alternatives (1,2,…..J), Vj is the deterministic component of utility for alternative j, and, Vi is the deterministic component of utility for alternative i.

MNL model uses a linear prediction function f(i, n) to predict the probability of an individual/observation n choosing option i and is expressed as shown in Equation 2.

f(i, n) = βi * Xn where, βi is the regression coefficients and Xn is the explanatory variables.

(2)

Penmetsa and Pulugurtha

5

MNL predicts the probability of occurrence of a dependent variable using a set of given independent variables. Maximum likelihood estimate was used in estimating the coefficients of the variables. The coefficients estimated using the maximum likelihood are believed to be unbiased for large samples. To explain the extent of the effect of the independent variables on occurrence of the dependent variable, the odds ratio concept was used. Odds ratio is defined as the ratio of the probability of happening of an event to the probability of not happening of an event. There is a direct relationship between the coefficients obtained in the analysis and odds ratio (

p(x)

).

1−p(x)

ANALYSIS & RESULTS Table 2 shows the frequency of various traffic violations and distribution of injuries among those violations. TABLE 2 Frequencies of Traffic Violations and Driver Injury Severity Driver Injury Parameter Severe Injury Moderate Injury PDO 84 17,093 120,118 No traffic violation 1 52 236 Disregarded yield sign 31 1,227 2,125 Disregarded stop sign 5 143 511 Disregarded other traffic signs 11 2,052 5,832 Disregarded traffic signal 3 307 822 Disregarded road markings Exceeded authorized speed 227 2,100 1,713 limit Exceeded safe speed for 195 10,270 24,397 conditions 4 945 6,045 Improper turn 0 19 266 Right turn on red 190 3,394 5,062 Going wrong way 6 720 9,546 Improper lane change 5 295 1,006 Use of improper lane 0 0 16 Passed on hill 2 16 43 Passed on curve 4 203 1,340 Other improper passing 1 73 3,207 Improper backing 1 17 215 Improper parking 0 13 95 Improper or no signal 1 353 3,716 Followed vehicle closely Operated vehicle erratically or 186 5,195 5,780 aggressively 11 767 2,830 Operated defective equipment 188 3,858 4,807 Alcohol 8 444 503 Drugs Total 1,164 49,556 200,231

Total 137,295 289 3,383 659 7,895 1,132 4,040 34,862 6,994 285 8,646 10,272 1,306 16 61 1,547 3,281 233 108 4,070 11,161 3,608 8,853 955 250,951

Penmetsa and Pulugurtha

6

About 45% of the total crashes occurred due to violating traffic rules whereas more than 90% of the fatalities occurred due to some kind of traffic violation during the study period. Exceeding authorized speed limit followed by exceeding safe speed limit for conditions contributed to the highest number of fatalities. The third highest number of fatalities was caused due to going in the wrong way, which possibly leads to head-on collisions where the severity or damage is high. This is closely followed by driving under the influence of alcohol and aggressive driving. These five traffic violations contributed 85% of the fatalities. The total number of crashes due to exceeding safe speed limit for conditions is greater than the number of crashes due to exceeding speed limit. A majority of these crashes due to exceeding safe speed for conditions resulted in PDO crashes. Likewise, a majority of the improper lane changes resulted in PDO crashes. On the other hand, most of the crashes that occurred due to no traffic violation resulted in PDO. A considerable number of crashes occurred due to improper parking or improper backing of vehicles. There could be several factors that may make drivers to violate traffic rules. Figure 1 exhibits a comparison of crashes due to and not due to traffic violations under different lighting conditions and by gender. 70% 60% 50%

% of Crashes

50%

49%

49%

46%

40% 40%

36%

35% 65%

30% 50%

54%

64% 51%

60% 51%

20% 10% 0% Daylight

Dusk

Dawn

Dark-LR

Dark-NL

Male

Female

Lighting Conditions and Gender % of crashes not due to traffic violation

% of crashes due to traffic violation

FIGURE 1 Comparison of Light Conditions and Gender with % of Crashes due and not due to Traffic Violations Except during daylight, the percentage of crashes due to violation of traffic rules is less compared to the percentage of crashes due to no traffic violation. Figure 1 implies that drivers are more likely to not comply with traffic rules when they have good visibility of the roadway. There is a big difference between the percentage of crashes due to and not due to traffic violations during dawn and dark conditions. Male drivers are more likely to violate traffic rules and get involved in a crash compared to female drivers. This shows that female drivers are more likely to follow traffic rules compared to male drivers. In simple terms, male drivers are relatively more aggressive and risk takers. Figure 2 shows the distribution of percentage of crashes due to traffic violations among different age groups.

Penmetsa and Pulugurtha

7

80% 70%

68% 57%

% of Crashes

60% 50%

46%

45% 37%

40%

63%

30%

35% 65%

55%

54%

43%

20% 32% 10% 0% <=18 years

19-25 years

26-35 years

36-50 years

51-65 years

>65 years

Driver's Age % of crashes not due to traffic violation

% of crashes due to traffic violation

FIGURE 2 Percentage of Crashes due to Traffic Violations by Age From Figure 2, the percentage of crashes due to violation of traffic rules reduce with an increase in the age of drivers. Drivers with 35 to 65 years of age had a huge difference in the percentage (% crashes not due to traffic violations - % crashes due to traffic violations). However, this difference reduced for drivers with age greater than 65 years as often times they commit violations due to poor vision (Kline et al., 1992) and misjudgment of gaps. Drivers with age less than 18 years have a high percentage of crashes due to traffic violations, implying that young drivers are more risk takers and their perception of risk is different from adult and older drivers. The trend lines clearly depict the risk perception by age. Multinomial logit model was developed to examine the effect of different traffic violations on driver injury severity. Table 3 shows the results of the model developed. The reference dependent and independent variables are no traffic violations and PDO crash, respectively. TABLE 3 Multinomial Logit Model Results Severe Injury Moderate Injury Parameter Estimate p-value Estimate p-value -7.27 <0.01 -1.95 <0.01 Intercept 1.80 0.07 0.44 <0.01 Disregarded yield sign 3.04 <0.01 1.40 <0.01 Disregarded stop sign 2.64 <0.01 0.67 <0.01 Disregarded other traffic signs 0.99 <0.01 0.91 <0.01 Disregarded traffic signal 1.65 <0.01 0.97 <0.01 Disregarded road markings 5.24 <0.01 2.15 <0.01 Exceeded authorized speed limit 2.44 <0.01 1.09 <0.01 Exceeded safe speed for conditions -0.06 0.91* 0.09 <0.01 Improper turn -6.88 0.92* -0.69 <0.01 Right turn on red 3.98 <0.01 1.55 <0.01 Going wrong way -0.11 0.80* -0.63 <0.01 Improper lane change 1.96 <0.01 0.72 <0.01 Use of improper lane -7.01 0.98* -8.57 0.86* Passed on hill

Penmetsa and Pulugurtha

8

4.20 <0.01 0.96 Passed on curve 1.45 <0.01 0.06 Other improper passing -0.81 0.42* -1.83 Improper backing 1.90 0.76* -0.59 Improper parking -6.81 0.95* -0.04 Improper or no signal -0.96 0.34* -0.40 Followed vehicle closely Operated vehicle erratically or 3.83 <0.01 1.84 aggressively 1.72 <0.01 0.64 Operated defective equipment 4.02 <0.01 1.73 Alcohol 3.12 <0.01 1.83 Drugs Note: - indicates parameter not significant at a 95% confidence level.

<0.01 0.41* <0.01 0.02 0.90* <0.01 <0.01 <0.01 <0.01 <0.01

From Table 3, traffic violations such as improper turn, right turn on red, improper lane change, passing on a hill, improper backing, improper or no signal and following vehicle closely have higher probabilities of resulting in a PDO crash. The sign explains whether the violation increases or decreases the probability of resulting in a severe crash. However, the odds ratio quantifies the probability. Exceeding authorized speed limit, aggressive driving, driving under the influence of alcohol, passing on curve (in no passing zones), and exceeding safe speed for condition result in severe driver injuries. The log likelihood (Log L) value of the model developed was -119493.665. This value can be used to compare models, and the value itself is not meaningful. The Log L value of the constant only model is -131851.47. Rho-square (ρ2) can be used to some extent to estimate the overall goodness of fit of the model. It is computed as follows.

𝛒𝟐𝒄 = 1-

𝑳𝒐𝒈 𝑳 (𝑴𝒐𝒅𝒆𝒍 𝑫𝒆𝒗𝒆𝒍𝒐𝒑𝒆𝒅) 𝑳𝒐𝒈 𝑳 (𝑪𝒐𝒏𝒔𝒕𝒂𝒏𝒕 𝒐𝒏𝒍𝒚 𝒎𝒐𝒅𝒆𝒍)

(3)

A ρ2𝑐 close to 1 implies a perfect model. An effort was made to check whether the model developed was significantly different from the constant only model. The null hypothesis and alternative hypothesis statements are as follows.

Ho (null hypothesis)

= the model developed is same as the constant only model

HA (Alternative hypothesis)

= the model developed is significantly different from the constant only model

The null hypothesis can be rejected if Equation (4) is satisfied. -2(Log L of constant only model – Log L of developed model) > chi-square value

(4)

Substituting the Log L values in Equation (4) yielded 24715.61 on the left hand side. This is greater than the Chi-square value for one degree of freedom (3.841) at a 95% confidence level (since there is only one dependent variable, one degree of freedom is considered). Therefore, this model is significantly different from the constant only model.

Penmetsa and Pulugurtha Table 4 depicts different traffic violations and their likelihood of resulting in severe driver injury and moderate injury compared with driver injury in no traffic violation. TABLE 4 Estimated Odds Ratio for Each Traffic Violation Odds Ratio Parameter Severe Moderate PDO Injury Injury 1.00 6.06 1.55 Disregarded yield sign 1.00 20.86 4.06 Disregarded stop sign 1.00 13.99 1.97 Disregarded other traffic signs 1.00 2.70 2.47 Disregarded traffic signal 1.00 5.22 2.63 Disregarded road markings 1.00 189.50 8.62 Exceeded authorized speed limit 1.00 11.43 2.96 Exceeded safe speed for conditions 1.00 1.10 Improper turn 1.00 0.50 Right turn on red 1.00 53.67 4.71 Going wrong way 1.00 0.53 Improper lane change 1.00 7.11 2.06 Use of improper lane 1.00 Passed on hill 1.00 66.51 2.62 Passed on curve 1.00 4.27 Other improper passing 1.00 0.16 Improper backing 1.00 0.56 Improper parking 1.00 Improper or no signal 1.00 0.67 Followed vehicle closely Operated vehicle erratically or 1.00 46.02 6.32 aggressively 1.00 5.56 1.91 Operated defective equipment 1.00 55.97 5.64 Alcohol 1.00 22.74 6.20 Drugs Note: - indicates parameter not significant at a 95% confidence level. If a driver disregarded a yield sign, he/she is 6 times more likely to succumb to a severe injury when compared to their injury in no traffic violation crash. Non-compliance with stop sign has a higher probability of resulting in fatal and severe injury compared to disregarding yield sign. People often disregard stop signs; disregarding stop sign is 20 times more likely to result in a severe driver injury and 4 times more likely to result in a moderate driver injury compared to driver injury in crash occurred due no traffic violation. Disregarding signals has a lower probability of resulting in severe driver injury compared to crashes that occurred due to non-compliance with road signs such as yield sign, stop sign, and other signs (work zone signs, school zone, etc.). Among all the traffic violations, exceeding the authorized speed limit is 182 times more likely to result in a severe crash. Passing on a curve (in no passing zones), driving under the influence of alcohol, going in wrong way, and aggressive driving are approximately 50 times more likely to result in severe driver injury. Right turn on red violations, improper backing, and improper parking has higher probabilities of resulting in PDO compared to injuries in no traffic violations. Driving defective equipment is 5 times more likely to result in severe driver injury.

9

Penmetsa and Pulugurtha As mentioned previously, Ayuso et al. (2010) has carried out a similar kind of study using data for Spain. Their conclusions contradict the results of this study. Comparing among different traffic violations, Ayuso et al. (2010) stated that disregarding traffic signals, stop signs, and yield sign result in less severe crash in Spain, whereas in the state of North Carolina, United States these traffic violation has higher probabilities of resulting in severe and moderate driver injury. Similarly, traffic signs violations and erratic driving are more likely to result in severe or moderate driver injury in the state of North Carolina, United States while they result in less severe injuries in Spain.

CONCLUSIONS Drivers often perceive that violating a traffic rule does not end up in a crash or severe crash. However, the reality is far different than what drivers often perceive. A majority of the traffic violations have higher probabilities of resulting in severe driver injuries compared to injuries in no traffic violations. The associated risk varies by traffic violation. The findings from this study serve as documented evidence to educate and create awareness among the drivers about the risk of violating traffic rules to themselves. Educating drivers about the risk associated with various traffic violations could help them develop safe driving behavior, which would eventually improve safety on roads and contribute towards reaching the “zero traffic deaths” vision. The findings from this study could also be used by policy makers and transportation system managers to identify traffic violations that need to be immediately addressed so as to reduce both crashes and fatalities. Traffic rule violators increase the risk to other road users. They may be imposed with penalty points on their driver's license, asked to pay a fine, and may have their license revoked depending upon the type of violation committed (whether the traffic violation leads to a crash or not). It is important to feed the risk in terms of injury severity to themselves as well as other road users to define fine amount or penalty points. The integration of risk with enforcement penalties merits research and investigation.

REFERENCES Abdel-Aty M. Analysis of driver injury severity levels at multiple locations using ordered probit models. Safety Research. 2003;34(5):597-603. Ayuso M, Guillén M, Alcañiz M. The impact of traffic violations on the estimated cost of traffic accidents with victims. Accident Analysis & Prevention. 2010;42(2):709-717. Braitman KA, Kirley BB, Ferguson S, Chaudhary NK. Factors leading to older drivers' intersection crashes. Traffic Injury Prevention. 2007;8(3):267-274. Campbell BN, Smith JD, Nazm WG. Analysis of fatal crashes due to signal and stop sign violations. Washington, DC: National Highway Traffic Safety Administration (NHTSA), Report No. DOT HS 809 779. 2004. Carson J, Mannering F. The effect of ice warning signs on ice-accident frequencies and severities. Accident Analysis & Prevention. 2001;33(1):99-109.

10

Penmetsa and Pulugurtha Cunningham RM, Maio RF, Hill EM, Zink BJ. The effects of alcohol on head injury in the motor vehicle crash victim. Alcohol and Alcoholism. 2002;37(3):236-240. Drummer OH, Gerostamoulos J, Batziris H, Chu M, Caplehorn J, Robertson MD, Swann P. The involvement of drugs in drivers of motor vehicles killed in Australian road traffic crashes. Accident Analysis & Prevention. 2004;36(2):239-248. Geedipally S, Turner P, Patil S. Analysis of motorcycle crashes in Texas with multinomial logit model. Transportation Research Record. 2011;2265:62-69. Khorashadi A, Niemeier D, Shankar V, Mannering F. Differences in rural and urban driver-injury severities in accidents involving large-trucks: an exploratory analysis. Accident Analysis & Prevention. 2005;37(5):910-921. Kline DW, Kline TJ, Fozard JL, Kosnik W, Schieber F, Sekuler R. Vision, aging, and driving: The problems of older drivers. Journal of gerontology. 1992;47(1):27-34. Li G, Keyl PM, Smith GS, Baker SP. Alcohol and injury severity: reappraisal of the continuing controversy. Trauma and Acute Care Surgery. 1997;42(3):562-569. Machin MA, Sankey KS. Relationships between young drivers’ personality characteristics, risk perceptions, and driving behavior. Accident Analysis & Prevention. 2008;40(2):541-547. Malyshkina N, Mannering F. Effect of increases in speed limits on severities of injuries in accidents. Transportation Research Record. 2008;2083:122-127. Moyano-Díaz E. Evaluation of traffic violation behaviors and the causal attribution of accidents in Chile. Environment and Behavior. 1997;29(2):264-282. National Center for Statistics and Analysis. 2013 motor vehicle crashes: Overview. Washington, DC: National Highway Traffic Safety Administration (NHTSA). 2014. National Center for Statistics and Analysis. Speeding: 2013 data. Traffic Safety Facts. Washington, DC: National Highway Traffic Safety Administration(NHTSA), Report No. DOT HS 812 162. 2015. Pai CW, Saleh W. Exploring motorcyclist injury severity in approach-turn collisions at T-junctions: Focusing on the effects of driver's failure to yield and junction control measures. Accident Analysis & Prevention. 2008;40(2):479-486. Retting RA, Weinstein HB, Solomon MG. Analysis of motor-vehicle crashes at stop signs in four US cities. Safety Research. 2003;34(5);485-489. Savolainen P, Ghosh I. Examination of factors affecting driver injury severity in Michigan's singlevehicle-deer crashes. Transportation Research Record. 2008;2078:17-25. Shankar V, Mannering F. An exploratory multinomial logit analysis of single-vehicle motorcycle accident severity. Safety Research.1996;27(3):183-194. Stoduto G, Vingilis E, Kapur BM, Sheu WJ, McLellan BA, Liban CB. Alcohol and drug use among motor vehicle collision victims admitted to a regional trauma unit: demographic, injury, and crash characteristics. Accident Analysis & Prevention. 1993;25(4):411-420. Waller PF, Stewart JR, Hansen AR, Stutts JC, Popkin CL, Rodgman EA. The potentiating effects of alcohol on driver injury. Journal of American Medical Association. 1986;256(11):1461-1466. Yagil D. Gender and age-related differences in attitudes toward traffic laws and traffic violations. Transportation Research Part F: Traffic Psychology and Behaviour. 1998;1(2):123-135.

11