category: literaturenote citekey: prechtidentifyingeffectsdrivingsecondary2017 title: "Identifying effects of driving and secondary task demands, passenger presence, and driver characteristics on driving errors and traffic violations – Using naturalistic driving data segments preceding both safety critical events and matched baselines" authors: "Precht, Lisa; Keinath, Andreas; Krems, Josef F." year: 2017 date: 2017-00-00 2017 doi: 10.1016/j.trf.2017.09.003 publication: "Transportation Research Part F: Traffic Psychology and Behaviour" url: "https://doi.org/10.1016/j.trf.2017.09.003" zotero_key: ST5Q62HQ zotero_storage: D25NV6G3 collections: transpordi_planeerimine folder: 001_artiklid firstAuthor: "Precht, Lisa"

status: converted

Transportation Research Part F

journal homepage: www.el sevier.com/locate/trf

Identifying effects of driving and secondary task demands, passenger presence, and driver characteristics on driving errors and traffic violations – Using naturalistic driving data segments preceding both safety critical events and matched baselines ^q

Lisa Precht a,⇑ , Andreas Keinath ^a , Josef F. Krems ^b

^a BMW Group, 80788 München, Germany
^b Technische Universität Chemnitz, 09107 Chemnitz, Germany

article info

Article history: Received 11 July 2017 Received in revised form 3 September 2017 Accepted 25 September 2017 Available online 13 October 2017

Keywords: Driver behavior Driver distraction Naturalistic driving Crash causation

abstract

Researchers have identified various factors that likely affect aberrant driving behaviors and therefore crash risk. However, it remains unclear which of these factors poses the greatest risk for committing either errors or violations under naturalistic driving conditions. This study investigated important variables contributing to driving errors and traffic violations based on naturalistic driving data from the second Strategic Highway Research Program (SHRP 2). The analyzed driving segments preceded both safety critical events and matched baselines. Results showed that intersection influence, high-risk visually distracting secondary tasks, and the severities of the safety critical events were the main factors associated with driving errors. The primary factors linked to violations were intersection influence, persistent individual differences in driver behavior, and the severities of the safety critical events. Furthermore, the number of aberrant driving behaviors in trip segments preceding crashes was higher than in the matched segments unrelated to safety critical events. However, the most common aberrant driving behavior types in the respective segment groups appeared to resemble each other. This suggests that crashes became more likely due to drivers committing more violations and errors overall as opposed to drivers making one certain type of error or violation.

1. Introduction

According to the World Health Organization, road traffic crashes lead to the loss of over 1.2 million lives and cause nonfatal injuries to as many as 50 million people around the world each year (World Health Organization, 2017). Road user behavior is a central factor in crash risk (Rowe, Roman, McKenna, Barker, & Poulter, 2015). A survey on crash causation (Singh, 2015) indicated that driving errors and violations were key in about 74% of the crashes. Therefore, these ''aberrant driving behaviors" are a crucial threat to road safety (Zhang, Chan, & Zhang, 2015). Environmental, situational, and individual factors may play a role in driving errors and traffic violations. The current study's focus was on how these factors contribute

E-mail address: Lisa.Precht@bmw.de (L. Precht).

^q The findings and conclusions of this paper are those of the authors and do not necessarily represent the views of the VTTI, SHRP 2, the Transportation Research Board, the National Academies, or the BMW Group, who funded the current work.

^⇑ Corresponding author.

to aberrant driving behaviors in naturalistic driving segments. In addition, this study examined possible differences in the number and types of aberrant driving behaviors between driving segments preceding safety critical events of varying intensity and matched driving segments unrelated to safety critical events.

1.1. Aberrant driving behaviors

A connection between aberrant driving behaviors and accidents has been well supported (e.g., Singh, 2015). Nevertheless, relatively little research exists that has systematically examined the nature of and particularly the factors contributing to errors and violations in naturalistic driving.

Errors and violations are distinct given their different psychological origins (Reason, Manstead, Stradling, Baxter, & Campbell, 1990). Errors are defined as unwitting deviations of action from intention or the failure of planned actions to achieve their intended consequence (Reason, 1990; Reason et al., 1990; e.g., unintentional lane deviation; signal violation, apparently did not see signal). In contrast, violations are ''deliberate deviations from those practices believed necessary to maintain the safe operation of a potentially hazardous system" (Reason et al., 1990, p. 1316; e.g., stop sign violation, intentionally ran stop sign at speed; signal violation, intentionally disregarded signal).

The following sections provide an overview of the potential risk factors examined in the current study for aberrant driving behaviors related to driving task demands, secondary task demands, passenger presence, and driver characteristics (see Precht, Keinath, and Krems, 2017b, for a more detailed review).

1.2. Potential risk factors for aberrant driving behaviors

1.2.1. Driving task demands

Evidence exists that the driving context can affect crash risk either directly or through moderating driver distraction effects. Environmental and situational factors, such as traffic density, driving through intersections, lighting and weather conditions were found to be associated with crash risk as well as the risk of driving errors (e.g., Johansson, Wanvik, & Elvik, 2009; Qiu & Nixon, 2008; Tian, Li, Chen, Chen, & Witt, 2013; Young, Salmon, & Lenne, 2013).

However, Young and Regan (2013) point out that drivers can engage in behavioral adaptation in the form of compensatory or self-regulatory behaviors, either to prepare themselves for distraction or to compensate for its effects. A recent naturalistic driving study (Precht et al., 2017b) observed that drivers indeed consider the driving context in several ways. For example, drivers committed fewer violations during demanding traffic conditions or when driving through interchanges.

Therefore, it still remains unclear how the driving context affects aberrant driving behaviors under real-world conditions and to what extent drivers adapt their behavior to minimize crash risk, for example with respect to committing violations.

1.2.2. Secondary task demands

Inattention and driver distraction due to secondary task engagement are among the leading causes of motor vehicle accidents (e.g., Klauer, Dingus, Neale, Sudweeks, & Ramsey, 2006; Klauer et al., 2014). It is assumed that a large number of these accidents are the result of driving errors caused by the driver being distracted (e.g., Staubach, 2009). Driving distractions can stem from competing visual, manual, or cognitive resources (Strayer & Turrill, 2014). All of these distraction types have been associated with detrimental effects on driving performance. However, whereas there is broad agreement on the negative impact of visual and visual-manual distraction (e.g., Carsten & Merat, 2015), researchers still debate the effects of cognitive load from secondary task engagement on driving errors and crash risk.

Some findings suggest that cognitive distraction impairs driving performance due to a diversion of attention (e.g., Horrey & Wickens, 2006; Strayer et al., 2013). However, a recent naturalistic driving study found that cognitively distracting secondary tasks, such as passenger conversation and talking on the phone, did not impact driving errors (Precht et al., 2017b). These results were found despite factoring in the visually distracting subtask talking while looking at the passenger. Moreover, cognitive distraction can mitigate driving error risk caused by fatigue (Jellentrup, Metz, & Rothe, 2011). Therefore, it is not surprising that cognitively distracting tasks have been associated with a decreased crash and near-crash risk in several naturalistic driving studies (Hickman, Hanowski, & Bocanegra, 2010; Klauer et al., 2006; Olson, Hanowski, Hickman, & Bocanegra, 2009; Simmons, Hicks, & Caird, 2016; Victor et al., 2015).

In contrast, activities requiring the driver to glance away from the road are known to carry the highest crash risk (e.g., Dingus et al., 2016; Simmons et al., 2016). Congruent with this finding, numerous investigations have suggested a connection between visual/visual-manual distraction and driving errors (e.g., Olson et al., 2009).

Even though driver distraction and error causation have been thoroughly investigated, the impact of cognitive distraction compared to visual and visual-manual distraction on driving performance in real-world traffic remains controversial. In addition, many observable distractions related to secondary tasks are not purely cognitive, but are combinations of cognitive and manual distraction (e.g., talking on the phone) or of cognitive and visual distraction (e.g., passenger interactions). Due to the high crash risk associated with visual distraction, further examination is particularly needed for cognitively distracting tasks coupled with visually distracting subtasks such as passenger interaction.

1.2.3. Passenger presence

It is generally accepted that people change their behavior in social presence (Bond & Titus, 1983). Several investigations have reported that the presence of one or more passengers influences driving performance. However, the respective findings of how it influences performance are contradictory.

For instance, numerous studies have shown safer driver behaviors with the presence of passengers, such as fewer traffic violations (e.g., Precht et al., 2017b; Rosenbloom & Perlman, 2016), driving errors (e.g., Ross, Jongen, Brijs, Brijs, & Wets, 2016) and lower crash risk (e.g., Vollrath, Meilinger, & Krüger, 2002). Conversely, some findings have indicated that passenger presence correlates with an increase in driving errors and crash risk, as well as a decrease of hazard avoidance due to cognitive distraction, particularly for young drivers (for a review, see Durbin, McGehee, Fisher, & McCartt, 2014).

Consequently, more examination is required regarding the effects of passenger presence on aberrant driving behaviors under naturalistic driving conditions.

1.2.4. Driver characteristics

There is a large body of evidence that driver characteristics, such as driver age, experience, and sex greatly impact both errors and violations. De Winter and Dodou's (2010) meta-analysis of 174 studies using the Driver Behaviour Questionnaire (DBQ; Reason et al., 1990) found that age negatively correlated with self-reported errors and violations, driving experience positively correlated with self-reported errors and violations, and that males reported fewer errors and more violations than female drivers. Moreover, a recent naturalistic driving study (Guo et al., 2016) reported that the crash risk associated with secondary task engagement varied substantially by age group. Guo et al. (2016) found that secondary-task-induced distraction posed a consistently higher threat for drivers younger than 30 and above 65 when compared with middle-aged drivers.

In addition, evidence exists of permanent individual differences in driver behavior due to temperament and personality factors. A 24-year follow-up study conducted by Summala, Rajalin, and Radun (2014) concluded that individual differences in driver behavior persist for decades. An example of a personality factor possibly causing permanent individual differences is sensation seeking. It has been found to predict risky driving behavior, violations, as well as accidents in numerous investigations (e.g., Bachoo, Bhagwanjee, & Govender, 2013; Dahlen & White, 2006).

In conclusion, numerous investigations suggest that driver characteristics have a great impact on the way in which drivers behave. However, Precht et al.'s (2017b) very recent naturalistic driving study did not find any impact on errors or violations from driver characteristics such as sex, age, experience, and sensation seeking. Therefore, further analysis is warranted to establish how driver characteristics affect aberrant driving behaviors under naturalistic conditions.

1.3. Aberrant driving behaviors prior to safety critical events

There is a well-supported connection between aberrant driving behaviors and accidents (e.g., Singh, 2015). In contrast to vehicle or environmental factors, drivers have been found to be the critical reason in as many as 94% of crashes. Aberrant driving behaviors were discovered in about 74% of these crashes (Singh, 2015). However, little evidence is available whether drivers who crashed committed more violations and errors overall during the preceding trip, finally resulting in the crash, or whether a particular single error or violation was the cause. Furthermore, it remains unclear whether the frequency or type of aberrant driving behaviors preceding the crash affects crash severity. Therefore, it is necessary to examine whether naturalistic driving trip segments preceding crashes or near-crashes of varying severity differ from those not preceding a safety critical event (SCE) with regard to driving errors and violations to approach a complete model of crash causation.

1.4. Investigative approach

It must be noted that a major part of the previous findings on the cause of aberrant driving behaviors stem from questionnaire studies, relying on unverifiable self-reported error and violation frequencies. Thus it is essential to investigate the complex interrelationship between potential risk factors for aberrant driving behaviors and their combined effect on errors and violations under real driving conditions. Compared to driving simulation and field experiments, naturalistic driving studies (NDSs) offer researchers an invaluable opportunity to study driver performance and behavior in the real world with actual consequences and risks (Carney, McGehee, Harland, Weiss, & Raby, 2015). NDSs collect information on driver behavior by obtaining objective in situ driving data continuously for an extended time period. This enables the analysis of environmental, situational, and individual factors recorded in the seconds prior to real live errors or violations. Currently, the second Strategic Highway Research Program (SHRP 2; Campbell, 2012) is the largest and most comprehensive naturalistic driving study, therefore representing the current gold standard for driver distraction data. The analysis of this data will improve understanding of potential risk factors for aberrant driving behaviors and their actual effects on driver behavior in real traffic.

1.5. The current study – aims and hypotheses

The objective of the present investigation was to examine variables contributing to driving errors and traffic violations based on naturalistic driving data. Previous research outlines various factors likely to affect aberrant driving behaviors and therefore crash risk. However, it remains uncertain which of these factors poses the greatest risk under naturalistic driving conditions and whether observed risks change when considering several contributing variables. For this reason, the effects of driving and secondary task demands, passenger presence, and driver characteristics on aberrant driving behaviors were analyzed in the current study using 11,352 six-second trip segments of SHRP 2 video footage. In addition, the current goal was to examine these risk factors independent of driver impairments such as driver drowsiness or driving anger. Furthermore, this study analyzed possible differences in the number and types of aberrant driving behaviors between naturalistic driving segments preceding safety critical events of varying severity and matched naturalistic driving segments unrelated to safety critical events.

All the following hypotheses have been previously examined in simulation, field, or questionnaire studies. Furthermore, a very recent naturalistic driving study investigated part of the hypotheses in a smaller dataset and included the impairment of driving anger (Precht et al., 2017b). The current paper attempts to replicate and extend results of this work. However, to the best of the authors' knowledge, this study is the first to systematically analyze the causation of errors and violations in naturalistic driving segments preceding crashes, near-crashes, and matched baselines. It therefore provides new insights into the variables that contribute to dangers in real traffic which can only be revealed by the unique data of large and comprehensive NDSs.

Consistent with previous research, we hypothesized that...

(1) ...high driving task demands, such as driving in rain, high-density traffic, or through intersections would be associated with aberrant driving behaviors by increasing the number of errors.
(2) ...drivers would commit fewer violations during high driving task demands, showing adaptation of their behaviors to the driving context.
(3) ...visually and visual-manually distracting secondary tasks, which have been found to considerably increase crash risk, would result in more errors.
(4) ...there would not be a connection between cognitively distracting secondary tasks and error occurrences.
(5) ...drivers would commit fewer violations in the presence of one or more passenger(s).
(6) ...drivers' past frequency of committed violations would predict their current frequency of committed violations.
(7) ...no other driver characteristics examined would be associated with aberrant driving behaviors.
(8) ...the number of aberrant driving behaviors in the trip segments preceding safety critical events would be higher than in the matched naturalistic driving segments unrelated to safety critical events.

2. Method and materials

2.1. Data

The current study analyzed naturalistic driving data from the second Strategic Highway Research Program (SHRP 2). The SHRP 2 is the largest study of its kind, including data from 35 million vehicle miles and 5.4 million trips. Data from instrumented vehicles was collected from more than 3500 U.S. participants during a 3-year period (Dingus et al., 2016; Victor et al., 2015). Recorded data consisted of driving parameters such as speed, acceleration, braking, all vehicle controls, forward radar, and lane position. In addition, video views forward, to the rear, on the driver's face, and on the dashboard were captured (Blatt et al., 2015). A more detailed description of the data in SHRP 2 can be found at the study's data access website ''InSight" (see Fig. 1 for a summary of the data and the sampling procedure).

2.1.1. Test and control group data

The subset of SHRP 2 data used in this study consisted of 11,352 six-second segments, extracted from 556 different trips. The trip segments of the test group were defined as the duration of a crash or near-crash as well as the two minutes preceding the safety critical event (beginning two minutes before the crash or near-crash started, and ending when the incident ended). The trip segments of the test group included crash and near-crash events selected based on the following data specifications (defined in the SHRP 2 event data dictionary, available at https://insight.SHRP2nds.us/):

(1) The ''event severity" is ''crash" or ''near-crash" (preference given to crashes)
(2) The crash or near-crash was caused due to the ''fault" of the ''subject driver"
(3) There were no additional secondary tasks (''secondary task 2/3") present
(4) All ''surface conditions" were included, except ''snowy", ''icy", or ''oily"
(5) All ''event natures 1" were included, except ''conflict with parked vehicle" or ''other"
(6) All ''incident types 1" were included, except ''backing, fixed object"

In addition to meeting all of the above requirements, the set of events was identified and provided to the extent possible in the below distribution, with up to 50 events meeting each description. When there were more than 50 crash events for a given category available, events from as many different drivers as possible were selected. Crashes were prioritized over near-crashes.

a. Driver impairments = ''none apparent"
b. ''Secondary task 1" is one of the following (up to 50 each):

Fig. 1. Data and sampling procedure summary.

i. Interaction with passenger in adjacent seat
ii. Cell phone, talking hand-held
iii. Cell phone, texting
iv. Cell phone, browsing
v. Other external distraction
vi. Adjusting/monitoring radio
vii. Adjusting/monitoring other devices integral to the vehicle
viii. No secondary task (and minimal secondary tasks for two minutes preceding the crash or near-crash)

The secondary task 1 specifications were intended to ensure a sufficiently large sample of the secondary task demand risk factor, including less frequent secondary tasks such as adjusting other devices integral to the vehicle. Furthermore, the secondary task 2/3, surface condition, and driver impairment specifications were to minimize the possibility that multiple factors or factors other than those considered in the current study models simultaneously affected drivers. Moreover, participants had to have caused all crashes and near-crashes to increase the comparability of the different SCEs. Finally, the event nature and incident type specifications were intended to increase the probability that the crash/near-crash occurred after two minutes of driving.

The control group data consisted of one matched trip segment for each test group trip segment. The trip segments of the control group were represented by 6-s segment durations matching the criteria described below, plus/minus one minute

Table 1 Total number of SCEs in test group divided into crash severity and secondary tasks preceding the events.

Secondary task preceding the SCEa	Most severe	Police-reportable crash	Minor crash	Low-risk tire strike	Crash (total)	Not a crash (near crash)
Interaction with passenger in adjacent seat	3	2	17	28	50	0
Cell phone, talking hand-held	2	1	9	4	16	24
Cell phone, texting/browsing	1	6	11	2	20	30
External distraction	4	1	31	14	50	0
Adjusting/monitoring radio	1	2	5	8	16	9
Adjusting/monitoring other devices integral to the vehicle	0	0	2	7	9	4
No secondary task	3	0	11	36	50	0

^a Secondary task in which the driver engaged - beginning at any point during the 5 s prior to the ''precipitating event" time through the end of the conflict.

(beginning one minute before the matched 6 s, and ending one minute after). The control group segments were as similar as possible to the associated test group segments, except for an absence of a safety critical event occurrence. Thus, the Virginia Tech Transportation Institute (VTTI) matched the baselines individually and manually for each event. The baseline segments were matched for the following prioritized criteria (defined in the SHRP 2 event data dictionary, available at https://insight.:) SHRP2nds.us/)::)

(1)	Priority 1: Same participant
(2)	Priority 1: Secondary task 1
(3)	Priority 1: Secondary task 2 and 3 (i.e., no additional secondary tasks)
(4)	Priority 1: Driver impairment (i.e., none apparent)
(5)	Priority 2: Number of front seat passengers
(6)	Priority 2: Presence/Absence of rear seat passenger(s)
(7)	Priority 2: Weather
(8)	Priority 2: Lighting
(9)	Priority 2: Traffic density
(10)	Priority 2: Relation to junction
(11)	Priority 2: Locality
(12)	Priority 2: Intersection influence
(13)	Priority 2: Construction zone
(14)	Priority 3: Number of rear seat passengers
(15)	Priority 3: Surface condition
(16)	Priority 3: Traffic flow
(17)	Priority 3: Alignment
(18)	Priority 3: Weekday vs. weekend

The current analysis is part of a larger project investigating variables contributing to crash risk. A second study analyzed the 5 s prior to the crash's impact time/near-crash's proximity time (of the crashes and near-crashes included in the current study and additional crashes preceded by drowsiness), using the crash trifecta concept, to assess contributing factors in at-fault crash and near-crash events associated with driver distraction, driver drowsiness, or neither of the two: see Precht, Keinath and Krems (submitted for publication) for more information.

2.1.2. Final data sample

The requested data specifications resulted in 278 trip segments used as the test group in this study. These segments included 248 two-minute trip segments and 30 trip segments of less than two minutes (M = 0.92, SD = 0.57; crash/nearcrash occurred before the driver drove for two minutes). Furthermore, the test group data included 211 crash (N = 26 most severe and police-reportable crashes; N = 86 minor crashes; N = 99 low-risk tire strikes) and 67 near-crash events (the SHRP 2 event data dictionary contains the definition of the crash severities, available at https://insight.SHRP2nds.us/). Table 1 summarizes the descriptive statistics of the SCEs and the secondary tasks preceding these events. Furthermore, the matched control group segments originated completely from baseline events or previously non-coded data (N = 278).

2.2. Variables and data coding

Fig. 2 summarizes the variables investigated in this study and the study approach.

2.2.1. Dependent variables

Appendix A shows a list of the coding scheme used including all variables, definitions, and examples. Aberrant driver behavior was the main dependent variable and was operationalized through the variables driving error and traffic violation. The majority of driver behaviors were coded on the basis of the original ''SHPR 2 Researcher Dictionary for Video Reduction Data" (Version 3.4, pp. 46–51; available at https://insight.SHRP2nds.us/). In addition, driver behaviors, such as forcing to side or blocking (based on the Driving Anger Expression Inventory; Deffenbacher, Lynch, Oetting, & Swaim, 2002) were included into the coding scheme.

2.2.2. Independent variables

The independent variables in this study included: (1) secondary task demands, (2) driving task demands, (3) driver characteristics, (4) presence of passenger(s), and (5) crash severity (see Appendix A and InSight for all variable definitions). The secondary task demands were operationalized through the secondary tasks performed by the drivers, which were coded based on the original ''SHRP 2 Researcher Dictionary for Video Reduction Data" (Version 3.4, pp. 53–60; available at https://insight.SHRP2nds.us/). To account for the proportion of visual distraction in the passenger related secondary tasks

Fig. 2. Summary of variables and study approach. Aberrant driving behaviors were analyzed as a function of driving and secondary task demands, driver characteristics, passenger presence, and crash severity.

''Passenger in adjacent seat - interaction" and ''Passenger in rear seat – interaction", these original SHRP 2 categories were replaced by the following categories: ''Passenger in adjacent/rear seat – talking without looking at passenger", ''Passenger in adjacent/rear seat – talking while looking at passenger", and ''Passenger in adjacent/rear seat – any interaction but talking".

The driving task demands analyzed consisted of weather (''no adverse conditions", ''fog", ''mist/light rain", ''raining", ''snowing", ''rain and fog"), lighting (''dawn", ''daylight", ''dusk", ''darkness, lighted", ''darkness, not lighted"), traffic density (''Level-of-service A1 to B"/''smooth traffic", ''Level-of-service C to D"/''moderate traffic", ''Level-of-service E to F"/''heavy traffic"), and intersection influence (''traffic signal", ''stop sign", ''uncontrolled", ''interchange", ''parking lot", ''other", ''no intersection influence"). All driving task demand variables were coded based on the original ''SHRP 2 Researcher Dictionary for Video Reduction Data" (except for the slightly modified variables interchange and parking lot).

The driver characteristics assessed included age, sex, sensation seeking, driving experience (''driver mileage last year"), and the drivers' past behavior (''number of moving/traffic violations in past three years", ''number of crashes in past three years"), all collected via a ''Sensation Seeking Scale Survey", a ''Driver Demographic Questionnaire", and a ''Driving History Questionnaire" (see https://insight.SHRP2nds.us/ for more information). Passenger presence contained the two categories ''no passenger present" and ''one or more passengers present". Finally, crash severities examined included the categories ''most severe", ''police-reportable crash", ''minor crash", ''low-risk tire strike", ''near-crash", and no crash (i.e., matched baseline).

2.2.3. Control variables

Driver emotions were coded as control variables given that emotion-based distraction, especially anger, can impact driving performance (e.g., Precht, Keinath, & Krems, 2017a). Four categories of driver emotions (''happy", ''angry/frustrated", ''sad", ''surprised") were coded based on a scheme developed by Fitch, Hanowski, Burnett, and Crundall (2013; see Appendix A, 1.1–1.7) to ensure that factors outside of those considered in the current study models did not simultaneously affect drivers. For the same reason, the driver impairment drowsy, sleepy, asleep, fatigued was coded based on the original ''SHRP 2 Researcher Dictionary for Video Reduction Data".

2.2.4. Data coding procedure

The following four video views were coded for each trip segment: (1) forward, (2) rear, (3) on the driver's face, and (4) a view of the driver's interactions with the vehicle's steering wheel and center console. The video coding occurred at VTTI's secure data enclave. The data reduction software tool used was developed by VTTI, and the data was coded in collaboration with two professional VTTI data reductionists. This included coding the start and end time for each dependent and independent variable. Two of the raters coded different variables and all segments. One VTTI data reductionist coded the original SHRP 2 driver behavior and secondary task categories, whereas the other coded all variables in approximately 30% of the trip segments to examine inter-rater reliability. Cohen's kappa was calculated based on the variables driver behavior, categorization of driver behavior, secondary task, driver emotion, and intersection influence, all of which were coded independently in 190 two-minute segments. The resulting kappa values indicated an appropriate level of reliability (see Table 2).

During the data reduction procedure, coded driver behaviors were categorized according to Reason et al.'s (1990) differentiation between a violation and an error. Classifying aberrant driving behaviors as either deliberate or a misjudgment was mostly straightforward with the aid of video recordings and continuous data. The driver's direction of gaze, relation to other vehicles, adherence to traffic laws, and driving data (e.g., speed, acceleration, deceleration) were several of the indicators evaluated. Some driver behaviors such as ''failed to signal" and ''exceeded speed limit" were categorized as a violation unless there was a reason to think it was an error, as these driver behaviors represent traffic laws. In addition, the classification of driver behaviors as either deliberate or a misjudgment is already inherent to a large proportion of the SHRP 2 driver behavior categories (e.g., ''signal violation, apparently did not see signal", ''signal violation, intentionally disregarded signal", ''signal violation, tried to beat signal change"). When a driver behavior considered unsafe could not be assigned to either violation or error according to Reason et al.'s (1990) definitions, it was categorized as unclassified unsafe driving behavior (behavior was neither error nor violation but still unsafe, or it was ambiguous whether behavior was deliberate or a misjudgment). Driver behaviors not identified as a violation, error, or unsafe were categorized as ''other driver behavior". Other driver behaviors and unclassified unsafe driving behaviors were excluded from the analyses.

Finally, the two-minute trip segments were recoded to allow for an appropriate analysis of the variables and their interrelations. Given the wide variation of when the variables occurred during the trips, each six-second interval of the 556 trip segments was coded with regard to the frequency of occurrence of the variables analyzed in the current study. Therefore, 11,352 six-second trip segments were included in the analyses.

2.3. Data analysis

2.3.1. Test and control group checks

Prior to the main analysis, checks of the test and control groups were performed. Linear Mixed Models (LMM) were applied to verify whether environmental and situational factors, as well as secondary tasks (which occurred at least once simultaneously with a driving error), were evenly distributed in the test and control groups. The total time in which these variables occurred throughout the trip segments was compared. Statistical calculations were performed in IBM SPSS v.20.0. (IBM Corp. Released, 2011). Upon using visual inspection as well as a Kolmogorov-Smirnov test, it was determined that none of the variables were normally distributed. Therefore, the dependent variables were rank-transformed and the LMMs were applied on the rank-transformed data (Conover & Iman, 1981). Variable's mean rank replaced any ties occurring during the rank transformation. As at least two trip segments were sampled from the same driver, this variable was treated as a random effect to take the autocorrelation of the variables within driver into account. Fixed ''test group" (categorical variable with ''control group" as the reference group) effects for all models were tested.

2.3.2. Analysis of aberrant driving behaviors

To evaluate the factors contributing to traffic violations and driving errors, the frequency of these driver behaviors was analyzed as a function of secondary and driving task demands, driver characteristics, passenger presence, and crash severity. As a result, the dependent variables violation and error consisted of discrete counts, making the assumption of distribution normality inappropriate (e.g., Coxe, West, & Aiken, 2009). More appropriate distributions for this data type include the Poisson and negative binomial distribution. In addition, as mentioned, at least two trip segments were sampled from the same driver. Thus there may have been correlations between driver behaviors exhibited by the same driver, which would violate the independence assumption of traditional regression models (e.g., Hubbard et al., 2010). Generalized Linear Mixed Models (GLMM) with a log link function were applied to take account of the Poisson/negative binomial distributed data with repeated measures. Applying GLMM techniques addresses elements such as nested count data. They combine the properties of both linear mixed models (which incorporate random effects and therefore the autocorrelation of driver behavior within each driver in the current study) and generalized linear models (which analyze non-normal data by using link functions, such as Poisson or binomial distributed data; e.g., Bolker et al., 2009).

The glmmADMB package (Fournier et al., 2012; Skaug, Fournier, Bolker, Magnusson, & Nielsen, 2016) in R 3.3.1 (R Core Team, 2016) was used to investigate how environmental, situational, and driver factors as well as crash severity associated with aberrant driving behaviors. The response variables consisted of the coded error and violation frequencies in each six-

Table 2 Inter-rater reliabilities.

Variable	Sample size of comparisons	Cohen's kappa value
Driver behavior	N = 486	k = 0.93
Categorization of driver behavior	N = 411a	k = 0.95
Secondary task	N = 1346	k = 0.88
Driver emotion	N = 429	k = 0.85
Intersection influence	N = 812	k = 0.88

^a Inter-rater reliability of ''categorization of driver behavior" could only be calculated from the driver behaviors consistently coded by both raters and excluded the category ''no driver behaviors".

second segment and followed a Poisson distribution. Eight separate GLMMs were fit (four separate models predicting errors and four separate models predicting violations, see below). The driver was treated as a random effect to consider the autocorrelation of driver behavior within driver. Fixed effects included: driving task demands (traffic density; intersection influence; weather; lighting), driver characteristics (age group; sex; driver mileage last year; sensation seeking score; number of moving/traffic violations in past three years; number of crashes in past three years), secondary task demands (''talking/singing, audience unknown"; ''Passenger in adjacent seat – talking without looking at passenger"; ''Passenger in adjacent seat – talking while looking at passenger"; ''Cell phone, talking/listening, hand-held"; ''Cell phone, texting/browsing/dialing, handheld"; ''Cell Phone, locating/reaching/answering"; ''Cell phone, holding"; ''External distraction"; ''Other non-specific internal eye glance"; ''Adjusting/monitoring climate control"; ''Adjusting/monitoring radio"; ''Adjusting/monitoring other devices integral to vehicle"; ''Reaching for object, other"; ''Object in vehicle, other"), passenger presence, and crash severity.

The lighting categories dawn and daylight were combined as dawn occurred too infrequently to be analyzed separately. Also combined were the crash severity categories most severe and police reportable crash. Furthermore, the analyses could only include secondary tasks occurring at least once simultaneously with a driving error and at least 20 times in the complete data set given the impossibility of fitting models with all secondary tasks observed in the current study (see Table B1). The fixed effects of the error and violation predicting GLMMs were kept consistent to allow for comparability.

To observe each set of variable's unique contribution in explaining the aberrant driving behaviors, four separate models predicting errors and four separate models predicting violations were applied. In order to control for the driving task demands, these variables were entered at first. The second model comprised driving task demands and driver characteristics, whereas the third model additionally included passenger presence and crash severity. Finally, the fourth model contained all four sets of predictor variables.

No considerable correlations existed between the predictor variables. However, as the different sets included related variables such as passenger related factors, all variable sets were estimated in various model versions. The significance level of the remaining variables and respective coefficient magnitudes were stable when one or more of the related variables were excluded from the analyses.

3. Results

3.1. Test and control group checks

Because the test-group included 30 trip segments of less than two minutes, the total duration of the test group segments (551.4 min, M = 1.98, SD = 0.39) differed from that of the matched baseline segments (583.8 min, M = 2.10, SD = 0.00).

3.1.1. Driver impairments in test and control group

Overall, drivers exhibited anger/frustration for 4.78 min (M = 0.01, SD = 0.09) of the 1135.2 min analyzed in the current study (see Fig. 3), making it unlikely that this variable affected driver performance. Crash and near-crash events directly caused all coded surprises and other emotions (e.g., fear). Drivers exhibited the emotion variable happy for 45.03 min (M = 0.08, SD = 0.23). In line with Precht et al.'s (2017b) findings, this variable was not significantly associated with errors or violations and thus was not included in the analyses presented in Section 3.5. The variable drowsy, sleepy, asleep, fatigued was never exhibited in the current study.

3.1.2. Driving and secondary task demands in test and control group

Tables B2 and B1 summarize the descriptive statistics for the driving task demands and the secondary tasks coded in the current study (see Appendix A for all variable definitions). Thirty-six LMMs tested whether the test and control groups had an even distribution of environmental and situational factors, as well as secondary tasks occurring at least once simultaneously with a driving error. The total time in seconds in which the variables occurred throughout the trip segments was compared. The Bonferroni adjustment method was used (a⁰ = a/36 = 0.001) to correct for the number of comparisons. The results showed significant differences between the durations of the variables smooth traffic, daylight, and no adverse weather conditions across the test and control group. These variables represent ''optimal" driving conditions, which would not be expected to impair driving performance and were the most common driving task demand variables in this dataset (see Table B2). The other analyzed variables, including such which could be expected to impair driving performance (e.g., heavy traffic, darkness, adverse weather conditions) did not differ significantly between the test and control group (see Table 3). Therefore, the risk of aberrant driving behavior occurrences due to driving and secondary task demands was comparable for both groups. As possibly risk-inducing variables did not differ significantly between the test and control group, the different durations of the most common variables smooth traffic, daylight, and no adverse weather conditions originated most probably due to the test and control group segment's different total durations (see Section 3.1).

3.2. Driver characteristics

The 250 drivers (127 female) were between 16 and 89 years of age (N = 67 were between 16 and 19 years; N = 89 were between 20 and 29 years; N = 49 were between 30 and 65 years; N = 45 were between 65 and 89 years). Participants' vehi-

Fig. 3. Total duration of exhibited emotion variables in minutes as a function of test and control group, related to total data sample.

Table 3 Results of LMM analyses for models predicting driving and secondary task demands.

Model		Fixed Effect	F-value (df)	p-value
Driving task demands
Traffic density	Level-of-service A1 to B (smooth)	Test Groupa	55.09 (1, 320.83)	0.000
	Level-of-service C to D (moderate)	Test Groupa	2.67 (1, 298.87)	0.103
	Level-of-service E to F (heavy)	Test Groupa	1.10 (1, 302.55)	0.295
Intersection influence	Traffic signal	Test Groupa	0.42 (1, 298.08)	0.518
	Stop sign	Test Groupa	0.07 (1, 293.02)	0.794
	Uncontrolled	Test Groupa	7.09 (1, 311.69)	0.008
	Interchange	Test Groupa	0.30 (1, 287.91)	0.582
	Parking lot	Test Groupa	0.74 (1, 299.51)	0.391
	Other	Test Groupa	1.28 (1, 312.67)	0.259
	No intersection influence	Test Groupa	0.73 (1, 296.81)	0.393
Lighting	Dawn	Test Groupa	0.00 (1, 296.20)	0.997
	Daylight	Test Groupa	51.09 (1, 296.38)	0.000
	Dusk	Test Groupa	0.49 (1, 326.05)	0.482
	Darkness lighted	Test Groupa	0.83 (1, 300.72)	0.363
	Darkness not lighted	Test Groupa	1.04 (1, 299.51)	0.308
Weather	No adverse conditions	Test Groupa	51.04 (1, 291.84)	0.000
	Adverse conditions	Test Groupa	0.48 (1, 297.67)	0.490
Secondary task demands
Talking/singing, audience unknown		Test Groupa	0.72 (1, 301.62)	0.396
Dancing		Test Groupa	0.67 (1, 288.46)	0.414
	Passenger in adjacent seat – talking without looking at passenger	Test Groupa	2.88 (1, 303.15)	0.091
	Passenger in adjacent seat – talking while looking at passenger	Test Groupa	0.85 (1, 305.60)	0.357
	Passenger in rear seat – talking without looking at passenger	Test Groupa	0.00 (1, 344.41)	0.992
Cell phone, talking/listening, hand-held		Test Groupa	0.71 (1, 304.08)	0.400
Cell phone, texting/browsing/dialing, hand-held		Test Groupa	0.01 (1, 303.10)	0.913
Cell phone, locating/reaching/answering		Test Groupa	0.19 (1, 301.71)	0.667
Cell phone, holding		Test Groupa	0.71 (1, 303.75)	0.401
External distraction		Test Groupa	0.00 (1, 309.20)	0.963
Adjusting/monitoring climate control		Test Groupa	0.71 (1, 330.30)	0.399
Adjusting/monitoring radio		Test Groupa	3.73 (1, 298.82)	0.054
Adjusting/monitoring other devices integral to vehicle		Test Groupa	0.22 (1, 340.18)	0.643
Reaching for object, other		Test Groupa	0.11 (1, 318.04)	0.740
Object in vehicle, other		Test Groupa	0.30 (1, 306.76)	0.584
Eating/drinking		Test Groupa	0.10 (1, 553.00)	0.750
Lighting/smoking/extinguishing cigar/cigarette		Test Groupa	0.21 (1, 553.00)	0.649
Biting nails/cuticles		Test Groupa	0.00 (1, 553.00)	1.000
Removing/adjusting jewelry		Test Groupa	0.22 (1, 416.66)	0.643

^a The reference group for the categorical variable ''test group" is ''control group". Significant effects are indicated in bold.

cles included 176 ''Cars", 24 ''Pickup Trucks", 46 ''Crossover SUVs", and 4 ''Vans/Minivans", their self-reported mileage within the ''last year" ranged from 0 to 100,000 (M = 12,464, SD = 10,273), and their sensation seeking scores ranged from 2 to 35 (M = 17.16, SD = 6.98; possible value range of sensation seeking score: 0–40). Furthermore, 144 participants claimed to have never received a moving or traffic violation in the past three years, 62 stated that they received one moving or traffic violation in the past three years, and 44 reported to have received two or more. A total of 165 participants stated that they were not involved in a crash that was, or could have been, police-reported in the past three years, 67 participants claimed to have experienced one police-reported crash in the past three years, and 18 participants noted two or more police-reported crashes.

3.3. Secondary tasks and passenger presence

Table B1 summarizes the descriptive statistics for the secondary task variables effectively coded in the current study (see Appendix A for all variable definitions). It must be noted that the majority of the secondary tasks were coded following the SHRP 2 reduction standard. Therefore, these tasks were coded as interrupted when an interruption lasted for 5 s or longer. Thus, the respective reported task durations can include interruptions of a maximum of 5 s. Exceptions to this were the secondary tasks related to passenger interactions (all interruptions for ''passenger in adjacent seat" and ''passenger in rear seat" interactions were coded when the interruption was caused by another passenger interaction task, e.g., ''talking while/without looking at passenger") and the secondary tasks related to external distractions (were coded as interrupted when interruptions lasted for 1 s or longer). These specifications ensured the coding of more precise glance durations and therefore more precise assessment of the visual distraction included in these tasks. Fig. 4 overviews the ten most commonly performed secondary tasks in this study. A total of 155 of the 556 coded trips included one or more passengers.

3.4. Aberrant driving behaviors

Table B3 summarizes the descriptive statistics for the aberrant driving behavior variables coded in the current study (see Appendix A for all variable definitions). Of all coded aberrant driving behaviors, 508 were categorized as a violation and 375 were categorized as an error. Fig. 5 shows the percentage of six-second segments containing errors and violations in relation to the total number of these segments in the respective crash severity categories/matched baselines. Error and violation percentages were higher in the trip segments preceding crashes compared to those not preceding a safety-critical event (matched baselines). Conversely, there was not a difference in the percentages of aberrant diving behaviors between the trip segments preceding near-crashes and the matched baseline trip segments. Moreover, Fig. 6 depicts the total number for each of the five most common aberrant driving behaviors (related to total trip segments) divided into the respective crash severities/matched baselines. Although there were differences in error and violation percentages between the crash and matched baseline segments, the most common aberrant driving behavior types resembled each other in the respective segment groups.

Fig. 4. Total number for each of the ten most common secondary tasks effectively coded in the current study in relation to total trip segments.

Fig. 5. Percentage of six-second segments in which aberrant driving behaviors occurred in relation to the total number of these segments in the respective crash severity categories/matched baselines.

3.5. Predictors of aberrant driving behaviors

3.5.1. Predictors of errors

Table 4 presents the findings from the four GLMMs predicting errors. Regarding the analyzed driving task demands, the complete model (Model 4) shows that only intersection influence was significantly associated with driving errors. Errors significantly decreased by an average factor of 0.74 when drivers waited at/passed a traffic light compared to not driving through intersections. In contrast, all other intersection influence types investigated (stop sign, uncontrolled, interchange, parking lot, other) significantly increased the driving error occurrence compared to not driving through intersections.

Neither the analyzed driver characteristics nor the presence of one or more passengers were significantly associated with errors.

Regarding examined crash severities, errors significantly increased by a factor of 1.73 when trip segments preceded most severe/police-reportable crashes, by 2.57 when they preceded minor crashes, and by 3.32 when they preceded low-risk tire strikes compared to those preceding no safety critical event (matched baseline). Trip segments preceding near-crashes contained proportionately about the same number of errors as matched baseline segments.

Finally, there were several secondary task demands that were significantly associated with driving errors. Errors significantly decreased by an average factor of 0.61 whenever drivers talked while looking at the passenger. Moreover, errors sig-

Fig. 6. Total number for each of the five most common aberrant diving behaviors (related to total trip segments) divided into the respective crash severities/matched baselines.

nificantly decreased by a factor of 0.10 whenever drivers looked at, handled, held, or manipulated an object, other than those listed in other categories. In contrast, whenever drivers pressed buttons or activated a touch screen on a cell phone to create/ send a text message, browse, or dial, errors significantly increased by an average factor of 1.71. Furthermore, errors significantly increased by an average factor of 1.31 whenever drivers focused on an external distraction. Additionally, in situations where drivers adjusted or monitored other devices integral to the vehicle (e.g., sun visors) and whenever they reached for an object not described in any other category, errors significantly increased by the factors 2.09 and 2.98 respectively.

3.5.2. Predictors of violations

Table 5 presents the findings from the four GLMMs predicting violations. Regarding the analyzed driving task demands, the complete model (Model 4) shows that only the factors intersection influence and weather were significantly associated with violations. Violations significantly decreased by an average factor of 0.65 when drivers waited at/passed a traffic light when compared to not driving through intersections. In contrast, violations significantly increased by an average factor of 6.00 while waiting at/passing a stop sign, by 1.75 while passing through an uncontrolled intersection, by 1.34 while driving in a parking lot, and by 2.52 while driving through other intersections (e.g., roundabouts) when compared to not driving through an intersection. Moreover, violations significantly increased by a factor of 1.58 when participants drove in rainy/ adverse weather compared to nice weather.

Most of the driver characteristic variables analyzed were not significantly associated with violations. However, when participants stated having had two or more moving or traffic violations in the past three years, the actual number of committed violations significantly increased by a factor of 1.76 compared to those who had stated that they have had no moving or traffic violations in the past three years.

Passenger presence was not significantly associated with violations. In terms of the analyzed crash severities, violations significantly increased by a factor of 2.40 when trip segments preceded the most severe/police-reportable crashes, by 1.93 when they preceded minor crashes, and by 1.55 when they preceded low-risk tire strikes compared to those not preceding a safety critical event (matched baseline). Trip segments preceding near-crashes contained proportionately about the same number of violations as matched baseline segments.

Regarding secondary task demands analyzed, violations significantly decreased by an average factor of 0.65 whenever drivers pressed buttons or activated a touch screen on a cell phone to create/send a text message, browse, or dial and by 0.11 whenever drivers looked at, handled, held, or manipulated an object, other than those listed in other categories. On the other hand, violations significantly increased by an average factor of 1.82 whenever drivers ate or drank.

4. Discussion

The current study investigated variables associated with driving errors and traffic violations. In addition, this study analyzed whether driving segments preceding safety critical events of varying severity differed from matched driving segments unrelated to safety critical events, with regard to the number and types of aberrant driving behaviors. Therefore, 11,352 sixsecond trip segments of SHRP 2 video footage were examined to assess the effects of environmental, situational, and driver factors on aberrant driving behaviors. To the best of the authors' knowledge, this comprehensive analysis based on naturalistic driving data is the first of its kind.

In summary, the driving task demand intersection influence, high-risk visually/visual-manually distracting secondary tasks, and the severities of the safety critical events subsequent to the analyzed trip segments were particularly associated with driving errors. The primary factors associated with violations under naturalistic driving conditions included the driving task demands intersection influence and weather, the number of violations drivers previously committed, and the severities of the safety critical events subsequent to the analyzed trip segments. Some secondary task demands were also associated with violations committed in the current study.

Error and violation percentages were higher in the trip segments preceding crashes compared to those not preceding a safety-critical event. However, the most common aberrant driving behavior types in the respective segment groups appeared to resemble each other. These results suggest that crashes became more likely due to drivers committing more violations and errors overall as opposed to committing one certain type of error or violation.

4.1. Driving task demands

Contrary to previous findings that environmental and situational factors affect crash risk as well as driving error risk (e.g., Johansson et al., 2009; Qiu & Nixon, 2008; Tian et al., 2013), the driving task demands investigated in the current study failed to majorly impact driving error frequency. Driving in high-density traffic, in dark conditions, or in rain was not associated with the number of errors. However, this coincides with Precht et al.'s (2017b) findings. Only the driving task demand intersection influence was associated with errors. They were significantly increased when drivers passed a stop sign/an uncontrolled intersection, when traveling through an interchange/in a parking lot, and when passing other types of intersections such as roundabouts. Thus the hypothesis that high driving task demands would increase the number of errors was only partially supported (Hypothesis 1).

Moreover, there was no support for the hypothesis that drivers would adapt their behaviors to the driving context indicated by committing fewer violations during high driving task demands (Hypothesis 2). Violations significantly increased while driving through intersections with stop signs/uncontrolled intersections, while driving in a parking lot, and while driving through other intersection types such as roundabouts when compared to not driving through an intersection. Furthermore, violations significantly increased when participants drove in adverse weather compared to nice weather. Unlike Precht et al. (2017b), there was not any evidence in the current study of drivers committing fewer violations during demanding traffic conditions or when passing through interchanges. What may partly explain this contradictory result is that Precht et al.'s (2017b) data sample included angry drivers. Some of the most common violations committed in Precht et al.'s (2017b) study were illegal passing on right, sudden or improper braking, and following too closely, all of which occurred much less frequently in the current dataset. Angry drivers possibly commit more risky violations overall compared to those not exhibiting emotions and therefore could have a greater need for behavioral adaptation to driving task demands.

4.2. Secondary task demands

Not surprisingly, distracting activities requiring drivers to gaze away from the forward roadway and those associated with a particularly high crash risk (Dingus et al., 2016) were associated with the highest probability of driving errors, supporting the respective hypothesis (Hypothesis 3). This result is consistent with earlier studies (e.g., Olson et al., 2009; Precht

Table 4 GLMMs predicting errors.

		Model 1 (AIC = 4430.8)		Model 2 (AIC = 4444.2)		Model 3 (AIC = 4325.4)		Model 4 (AIC = 4297.5)
Fixed effect		Exp(b)	z-value	Exp(b)	z-value	Exp(b)	z-value	Exp(b)	z-value
Driving task demands
Traffic density	Moderate Traffica	0.71	1.48	0.71	1.48	0.83	0.76	0.81	0.83
	Heavy Traffica	0.74	0.44	0.79	0.35	1.03	0.04	0.95	0.07
IntersectionInflb	Traffic Signalc	0.78*	1.97	0.78	1.88	0.74*	2.28	0.74*	2.28
	Stop Signc	1.67**	3.27	1.67**	3.27	1.54**	2.71	1.58**	2.87
	Uncontrolledc	3.01***	6.61	3.03***	6.64	2.67***	5.88	2.68***	5.87
	Interchangec	1.45	1.10	1.44	1.07	1.87	1.83	2.02*	2.05
	Parking Lotc	1.42**	2.60	1.42**	2.59	1.41*	2.53	1.45**	2.67
	Otherc	2.35***	3.40	2.37***	3.44	2.13**	2.93	2.12**	2.91
Weather	Adverse Conditd	0.73	1.57	0.74	1.49	0.85	0.76	0.86	0.72
Lighting	Duske	1.02	0.05	1.01	0.03	0.87	0.40	0.94	0.19
	Darkness Lightede	1.00	0.01	0.98	0.12	1.00	0.02	1.04	0.22
	Darkness not Lige,f	1.76	1.69	1.77	1.72	1.12	0.32	1.24	0.58

Driver characteristics
Age group	16–19g	–	–	1.17	0.74	1.30	1.22	1.27	1.11
	20–29g	–	–	0.84	0.82	0.89	0.54	0.85	0.76
	65–89g	–	–	0.88	0.57	0.88	0.56	0.88	0.54
Sex	Maleh	–	–	1.13	0.85	1.04	0.24	1.06	0.38
Miles last year		–	–	1.00	0.38	1.00	0.58	1.00	0.52
Sensation seeking		–	–	0.99	1.12	0.99	1.02	0.99	1.00
Num. violationsi	1j	–	–	1.08	0.48	1.11	0.60	1.14	0.75
	2 or morej	–	–	1.28	1.23	1.39	1.59	1.35	1.49
Num. crashk	1l	–	–	0.97	0.18	1.03	0.17	1.00	0.02
	2 or morel	–	–	1.36	1.16	1.41	1.25	1.43	1.31
Passenger presence & crash severity
Passenger presm	1 Pass Presentn	–	–	–	–	1.10	0.66	1.13	0.72
Crash severity	MostSeverePolio,p	–	–	–	–	1.68*	2.46	1.73*	2.56
	Minor Crasho	–	–	–	–	2.51***	6.89	2.57***	7.02
	LowRiTireStrikeo,q	–	–	–	–	3.34***	9.74	3.32***	9.59
	Near-Crasho	–	–	–	–	1.22	0.98	1.13	0.59

Secondary task demands
Talking/Singingr		–	–	–	–	–	–	0.82	0.64
PassAdjs		–	–	–	–	–	–	1.14	0.78
PassAdjLookt		–	–	–	–	–	–	0.61*	2.15
CellTalku		–	–	–	–	–	–	1.17	0.81
CellTextv		–	–	–	–	–	–	1.71***	3.41
CellLoc/Reachw		–	–	–	–	–	–	0.72	0.97
CellHoldingx		–	–	–	–	–	–	0.88	0.52
ExternalDistry		–	–	–	–	–	–	1.31*	2.25
IntEyeGlancez		–	–	–	–	–	–	0.20	1.59
ClimateContraa		–	–	–	–	–	–	0.94	0.10
Radioab		–	–	–	–	–	–	1.31	1.28
DevicesIntegralac		–	–	–	–	–	–	2.09***	3.49

Table 4 (continued)

	Model 1 (AIC = 4430.8)		Model 2 (AIC = 4444.2)		Model 3 (AIC = 4325.4)		Model 4 (AIC = 4297.5)
Fixed effect	Exp(b)	z-value	Exp(b)	z-value	Exp(b)	z-value	Exp(b)	z-value
OtherObjReachad	–	–	–	–	–	–	2.98*	1.97
OtherObjVehicleae	–	–	–	–	–	–	0.10**	3.07
Eating/Drinking	–	–	–	–	–	–	1.37	0.87

^a The reference group for this categorical variable is ''smooth traffic".
^b Intersection influence.
^c The reference group for this categorical variable is ''no intersection influence".
^d The reference group for the categorical variable ''adverse condition" is ''no adverse condition".
^e The reference group for this categorical variable is ''dawn-daylight".
^f Darkness not lighted.
^g The reference group for this categorical variable is ''30–64".
^h The reference group for this categorical variable is ''female".
ⁱ Number of moving/traffic violations in past three years.
^j The reference group for this categorical variable is ''zero violations".
^k Number of crashes in past three years.
^l The reference group for this categorical variable is ''zero crashes".
^m Presence of passengers.
ⁿ The reference group for the categorical variable ''one or more passengers present" is ''no passenger present".
^o The reference group for this categorical variable is ''matched baseline".
^p Most severe and police-reportable crashes.
^q Low-risk tire strike.
^r Talking/singing, audience unknown.
^s Passenger in adjacent seat talking without looking at passenger.
^t Passenger in adjacent seat talking while looking at passenger.
^u Cell phone, talking/listening, hand-held.
^v Cell phone, texting/browsing/dialing, hand held.
^w Cell Phone, locating/reaching/answering.
^x Cell phone, holding.
^y External distraction.
^z Other non-specific internal eye glance.
aa Adjusting/monitoring climate control.
ab Adjusting/monitoring radio.
ac Adjusting/monitoring other devices integral to vehicle.
ad Reaching for object, other.
ae Object in vehicle, other.
*** p < 0.001.
** p < 0.01. * p < 0.05.

et al., 2017b). Interestingly, violations significantly decreased whenever drivers pressed buttons or activated a touch screen on a cell phone to create/send a text message, browse, or dial. This suggests that drivers engaged in behavioral adaptation to compensate for the distraction associated with this particularly risky secondary task.

As hypothesized, cognitive distractions were not associated with impaired driving performance (Hypothesis 4). Even passenger conversations did not impact driving errors despite factoring in its visually distracting subtask looking at the passenger. This finding corroborates recent naturalistic driving study results showing no relation between cognitively distracting secondary tasks and driving errors (Precht et al., 2017b). In fact naturalistic driving studies show that cognitively distracting secondary tasks correspond to decreased crash and near-crash risk (Hickman et al., 2010; Klauer et al., 2006; Olson et al., 2009; Simmons et al., 2016; Victor et al., 2015).

The observed significant decrease in errors whenever drivers talked while looking at the passenger or looked at, handled, held, or manipulated an object, other than those listed in other categories, could be because drivers performed these tasks more often while stationary (e.g., at traffic lights). This corroborates Precht et al.'s (2017b) results that the visually distracting subtask talking while looking at the passenger occurred more frequently in parking lots, at traffic lights, during nice weather and during nighttime (considered a low-traffic period). This also supports that drivers factor in the driving context when deciding to engage in risky behaviors.

4.3. Passenger presence

The current study's results did not correspond with recent findings that drivers display safer behaviors when accompanied by passengers (Precht et al., 2017b; Rosenbloom & Perlman, 2016), that passenger presence decreases driving errors (e.g., Ross et al., 2016), or that passenger presence increases errors due to cognitive distraction (e.g., Durbin et al., 2014). Therefore, the hypothesis that drivers would commit fewer violations in the presence of one or more passenger(s), which was based on Precht et al.'s (2017b) findings, was not supported (Hypothesis 5). The current study's contradictory result could be explained by the different violation types (discussed above in Section 4.1) committed in this study compared to Precht et al.'s (2017b) study. Passenger presence possibly only affects certain violation types, such as following too closely, which are considered particularly risky.

4.4. Driver characteristics

Analogous to Summala et al.'s (2014) findings based on a 24-year follow-up study and Precht et al.'s (2017b) study, the current investigation's results indicate that individual differences in driver behavior persist over time. Indeed, as hypothesized, drivers' number of previous committed violations (i.e., self-reported number of moving/traffic violations in past three years) predicted drivers' current violation frequency (Hypothesis 6).

Moreover, differing from previous findings mainly based on questionnaire data (e.g., de Winter & Dodou, 2010), driver characteristics such as sex, age, experience, and sensation seeking were not associated with either violations or errors. Nevertheless, this result is identical to Precht et al.'s (2017b) findings based on naturalistic driving data. Given that recent naturalistic driving studies suggest that there are, for example, driver age effects on crash risk (e.g., Guo et al., 2016), it is likely that driver characteristics more highly impact crash causation compared to error and violation occurrences as crashes and aberrant driving behaviors differ in their origination.

Table 5 GLMMs predicting violations.

		Model 1 (AIC = 6128.1)		Model 2 (AIC = 6137.3)		Model 3 (AIC = 6076.6)		Model 4 (AIC = 6064.0)
Fixed effect		Exp(b)	z-value	Exp(b)	z-value	Exp(b)	z-value	Exp(b)	z-value
Driving task demands
Traffic density	Moderate Traffica	0.83	0.96	0.83	0.97	0.84	0.91	0.87	0.73
	Heavy Traffica	0.00	0.01	0.00	0.04	0.00	0.02	0.00	0.02
IntersectionInflb	Traffic Signalc	0.61***	4.36	0.62***	4.27	0.62***	4.16	0.65***	3.77
	Stop Signc	5.92***	19.54	5.92***	19.56	5.97***	19.41	6.00***	19.37
	Uncontrolledc	1.88***	3.92	1.89***	3.96	1.74***	3.44	1.75***	3.47
	Interchangec	1.05	0.18	1.04	0.15	1.18	0.57	1.22	0.69
	Parking Lotc	1.23	1.75	1.23	1.75	1.27*	2.04	1.34*	2.44
	Otherc	2.76***	5.00	2.75***	4.98	2.56***	4.58	2.52***	4.48
Weather	Adverse Conditd	1.82***	3.36	1.84***	3.42	1.61**	2.59	1.58*	2.48
Lighting	Duske	1.24	0.57	1.19	0.48	0.99	0.02	1.09	0.22
	Darkness Lightede	0.96	0.23	0.94	0.40	0.90	0.66	0.96	0.24
	Darkness not Lige,f	1.88	1.86	1.83	1.79	1.34	0.83	1.50	1.14
Driver characteristics
Age group	16–19g	–	–	1.51	1.58	1.46	1.45	1.57	1.71
	20–29g	–	–	0.83	0.73	0.82	0.74	8.67	0.54
	65–89g	–	–	1.18	0.60	1.08	0.28	1.05	0.18
Sex	Maleh	–	–	1.02	0.13	0.97	0.16	0.95	0.27
Miles last year		–	–	1.00	0.17	1.00	0.24	1.00	0.29
Sensation seeking		–	–	1.00	0.23	1.00	0.26	1.00	0.22
Num. Violationsi	1j	–	–	1.00	0.02	0.99	0.04	1.02	0.10
	2 or morej	–	–	1.72*	2.19	1.73*	2.23	1.76*	2.29
Num. crashk	1l	–	–	0.93	0.35	0.94	0.30	0.95	0.25
	2 or morel	–	–	1.09	0.27	1.02	0.06	1.08	0.23
Passenger presence & crash severity
Passenger presm	1 Pass Presentn	–	–	–	–	1.13	0.96	1.15	1.06
Crash severity	MostSeverePolio,p	–	–	–	–	2.34***	4.72	2.40***	4.77
	Minor Crasho	–	–	–	–	1.83***	5.85	1.93***	6.33
	LowRiTireStrikeo,q	–	–	–	–	1.53***	4.41	1.55***	4.53
	Near-Crasho	–	–	–	–	1.20	1.15	1.21	1.19
Secondary task demands
Talking/Singingr		–	–	–	–	–	–	0.70	1.75
PassAdjs		–	–	–	–	–	–	0.86	1.11
PassAdjLookt		–	–	–	–	–	–	0.86	0.85
CellTalku			–	–	–	–	–	0.93	0.42
CellTextv		–	–	–	–	–	–	0.65**	3.07
CellLoc/Reachw		–	–	–	–	–	–	1.05	0.21
CellHoldingx		–	–	–	–	–	–	0.71	1.68
ExternalDistry		–	–	–	–	–	–	1.04	0.37
IntEyeGlancez		–	–	–	–	–	–	1.08	0.27
ClimateContraa		–	–	–	–	–	–	0.97	0.09
Radioab		–	–	–	–	–	–	0.96	0.25
DevicesIntegralac			–	–	–	–	–	0.99	0.07

Table 5 (continued)

	Model 1 (AIC = 6128.1)		Model 2 (AIC = 6137.3)		Model 3 (AIC = 6076.6)		Model 4 (AIC = 6064.0)
Fixed effect	Exp(b)	z-value	Exp(b)	z-value	Exp(b)	z-value	Exp(b)	z-value
OtherObjReachad OtherObjVehicleae Eating/Drinking		– – –	– – –	– – –	– – –	– – –	2.18 0.11* 1.82	1.24 2.97 2.22

^a The reference group for this categorical variable is ''smooth traffic".
^b Intersection influence.
^c The reference group for this categorical variable is ''no intersection influence".
^d The reference group for the categorical variable ''adverse condition" is ''no adverse condition".
^e The reference group for this categorical variable is ''dawn-daylight".
^f Darkness not lighted.
^g The reference group for this categorical variable is ''30–64".
^h The reference group for this categorical variable is ''female".
ⁱ Number of moving/traffic violations in past three years.
^j The reference group for this categorical variable is ''zero violations".
^k Number of crashes in past three years.
^l The reference group for this categorical variable is ''zero crashes".
^m Presence of passengers.
ⁿ The reference group for the categorical variable ''one or more passengers present" is ''no passenger present".
^o The reference group for this categorical variable is ''matched baseline".
^p Most severe and police-reportable crashes.
^q Low-risk tire strike.
^r Talking/singing, audience unknown.
^s Passenger in adjacent seat talking without looking at passenger.
^t Passenger in adjacent seat talking while looking at passenger.
^u Cell phone, talking/listening, hand-held.
^v Cell phone, texting/browsing/dialing, hand held.
^w Cell Phone, locating/reaching/answering.
^x Cell phone, holding.
^y External distraction.
^z Other non-specific internal eye glance.
aa Adjusting/monitoring climate control.
ab Adjusting/monitoring radio.
ac Adjusting/monitoring other devices integral to vehicle.
ad Reaching for object, other.
ae Object in vehicle, other.
*** p < 0.001.
** p < 0.01. * p < 0.05.

4.5. Aberrant driving behaviors prior to safety critical events

The hypothesis that the number of aberrant driving behaviors in the trip segments preceding safety critical events would be higher than in the matched naturalistic driving segments unrelated to safety critical events was supported for the segments preceding crashes (Hypothesis 8). Furthermore, there was a slight decrease in the percentage of six-second segments containing errors with increasing crash severity (9.10% in low-risk tire strikes, 7.95% in minor crashes, and 8.09% in most severe and police-reportable crashes), whereas the opposite trend occurred for the percentage of six-second segments containing violations (11.28% in low-risk tire strikes, 12.89% in minor crashes, and 14.45% in most severe and police-reportable crashes). This result could suggest that violations more often result in high severity crashes than driving errors.

In contrast, the percentage of aberrant diving behaviors did not differ between trip segments preceding near-crashes and matched baseline trip segments. Furthermore, even though the percentages of errors and violations differed between the crash and matched baseline segments, the most common aberrant driving behavior types in the respective segment groups appeared to resemble each other. These observations indicate that committing a high number of aberrant driving behaviors impacts crash risk more than single specific errors or violations.

4.6. Limitations

This study contained some limitations. For example, driver impairments strongly related to aberrant driving behaviors, such as driver fatigue and intoxicated driving, could not be considered in the current analyses. Furthermore, the categories of the investigated variables (e.g., sensation seeking score, number of crashes in past three years, adverse weather conditions) were not necessarily representatively distributed among the participants and the investigated trip segments. It is therefore clear that this investigation is only a step towards a complete model of the various variables that contribute to, or mitigate aberrant driving behaviors and therefore crash risk.

In addition, the current study only included 26 driving segments preceding most severe and police-reportable crashes. It would therefore be worthwhile to analyze aberrant driving behaviors in a similar dataset which, however, includes a larger number of driving segments preceding such crashes.

Moreover, the effects of driving task demands and driver characteristics on aberrant driving behaviors found in this study as well as by Precht et al. (2017b) were quite different from part of the findings in previous investigations regarding crash risk. Even though aberrant driving behaviors and accidents relate to each other, they represent very different events and differ in their origination. Future research should examine whether driving task demands and driver characteristics have a higher impact on crash causation compared to error and violation occurrences, as well as why that may be the case. In addition, it is necessary to further investigate how specific aberrant driving behaviors cause crashes and near-crashes and what role, for example, the timing relation between errors, violations, and crashes plays. Furthermore, research is needed regarding the interrelationships between specific aberrant driving behaviors, such as speeding or illegal passing, and environmental, situational, and individual factors.

Overall, further work is needed to validate this study's findings. Wide-ranging studies are essential to reveal mechanisms behind emerging driving errors and traffic violations to develop measures for reducing aberrant driving behaviors and therefore crash risk.

4.7. Practical implications

Consistently with earlier studies, the current analysis found that several distracting activities requiring drivers to gaze away from the forward roadway were associated with driving errors. In contrast, the examined cognitive distractions were not associated with impaired driving performance. Whether or not the cognitive load induced by natural language (used when talking on the phone/with a passenger) resembles the load induced by the language used when interacting with a voice-recognition system (e.g., command language) is controversial. Nevertheless, developing human–machine interfaces towards auditory interfaces including implementations of voice control for functions, which would otherwise induce visual/ manual distraction, could contribute to an increase in road safety. The protective effect of such systems should be the greater, the more similar their voice interaction is to natural conversation. Therefore, automobile manufacturers should develop voice-recognition systems which adapt to the driver and the driving context, comparable to a human passenger.

Another factor which was strongly associated with aberrant driving behaviors in this study was intersection influence. Intersections are a substantial part of the road system and represent a situation in which road users converge and potentially conflict with each other (Young et al., 2013). Many earlier investigations already observed that driving through intersections contributes to driving errors and crashes due to the increased complexity and need to make multiple decisions, often under time pressure and at considerable speed (e.g., Young et al., 2013). The findings of the current study suggest that intersections remain a particularly dangerous part of the road system. Therefore, further development and improvement of intersections and traffic guidance, including factors such as intersection design, traffic signal phasing, and environmental conditions, remains crucial, offering vast potential for an increase of road safety.

5. Conclusion

The current analysis provides insight into the emergence of aberrant driving behaviors under real-world driving conditions. The main factors associated with driving errors consisted of driving through intersections, engaging in high-risk visually/visual-manually distracting secondary tasks, and the severities of the safety critical events subsequent to the analyzed trip segments. On the other hand, the primary factors associated with violations under naturalistic driving conditions were driving through intersections, weather conditions, persistent individual differences in driver behavior, and the severities of the safety critical events following the analyzed trip segments.

Moreover, observations from the current study indicated that a large number of aberrant driving behaviors corresponded with a higher impact on crash risk compared to single specific errors or violations.

Interestingly, several effects noted in previous studies on aberrant driving behaviors were not observed in this naturalistic driving data investigation. For example, driving task demands such as driving in high-density traffic, in the dark, or in the rain were not associated with the number of driving errors. In addition, most of the examined driver characteristics did not impact the number of aberrant driving behaviors coded in the current study, with the exception of drivers' past behavior.

Acknowledgements

The authors would like to thank the Virginia Tech Transportation Institute employees for their very helpful cooperation, especially Sarah Robinson and Lisa Goodell for the data reduction.

Funding

The BMW Group supported this work.

Appendix A

Used coding scheme

ia b Va r #	le ia b le Va r na m e	ia b le de f in it ion Va r	les d h int Ex am p an s
1	ive io Dr Em t r o n	ive io io f in it io Dr Em t Re t De r o n ac n ns
1. 1 1. 2	Un b le De ine to ter a m l / ho Ne No Em ion S ut ot ra wn	Ca l l w ha ion he dr ive is ho ing ot te t e ot t nn m r s w he dr ha ht fac do le T ive ig i str t s r s a a e, es no m lau h, do t g tu g es no es re	or
1. 3	Ha p p y	he dr ive i les lau hs he dr ive T T tu r s m or g r g es re ite t ex c m en	in s
1. 4	/ d An Fr tra te g ry us	he dr ive low / br in k l T r er s sq ue ez es ey e ow s, w r for he d he dr ive len he h is / he h he T T r t t e a r c c s ee dr ive l ls ( h w i de it h br t r y e op en s m ou w ey e ow low d ) he dr / T ive in tu er e r g es re s an g er fru ion he dr ive ise h is / he l ip T str at r r a s r u p p er ig ht l ip t en s s	ing s or
1. 5	Sa d	T he dr ive ha dr br ( ise inn r s oo p y ey e ow s ra s er br low br ) he dr ive T ut ey e ow s, er s o er ey e ow s r fro by low ing he f h is / he t te wn s er ou r c or ne rs o	l ip r s
1. 6	Su ise d rp r	T he dr ive 's br ise T he dr ive 's h t r ey e ow s r a r m ou he dr T ive tu op en s r g es re s
1. 7	he Ot r	ion l r ion ha do f it int Em ot ct t t t ot a ea es no o a ny lea de f in in he ion teg P t No te t ca or y se e s s ec on Ex l Lo ce g	he r he t
2	ive ha io Dr Be r v r	ive be ha io ha in lu de ha he dr ive Dr t t t t r v rs c w d i d.	r	L. Precht et al. / Transportation Research Part F 51 (2017) 103–144
2. 0	ive ha ior No Dr Be r v s	Su b j h ic le dr ive in t v t ec e r e ng ag es no ap p ar en be ha ior lat d t ing i bu ing nt t to v s r e e o c au s or co r h h. cra s or ne ar -cr as	he t
2. 1	lee lee fat ig d Dr ow sy s p y, as p, ue ,	b j h ic le dr ive h i b its bv iou ig Su t v ec e r e x o s s ns o be ing lee ire d, is l ly lee h i t tu as p or or ac a as p w dr de d for f t he dr iv ing ing iv ing g ra p er m an ce o , h is ho l d lso be de d lee T Dr s u a co as ow sy s p y, , lee fat ig d de ive irm Dr Im t. as p, ue un r r p a en	f le k. ta s

Va ia b le r #	Va ia b le r na m e	Va ia b le de f in it ion r	Ex les d h int am p an s
2. 2	Ex de d d l im it ce e sp ee	Su b j h ic le l ing d g ha t v tra at ate r t ec e ve a sp ee re n he d d l b le d im it. In Va ia Sp Zo t te p os sp ee r ee ne s, h is is lat ive he d l im it in f fec he t to t t a t t re sp ee e im f t he t t. e o ev en	Ov 1 0 m h bo d d l im it. te er p a ve p os sp ee
2. 3	de d fe d bu d l Ex im it t n ot ce e sa sp ee sp ee	b h le l d los Su j ic ing t v tra at e t ec e ve a sp ee c o o r de he d d l im it, bu i l l t fas r t te t s t t t un p os sp ee oo o int in fe dr iv ing iro ive t g nt m a a a s a en v nm en n c ur re iro l c d it ion ( he f f ic ta t tra en v nm en on s e.g ., w ea r, , l ht ). ig ing	du d ha low Ex ing it ion ire s t t m r co n ay re q u s er ds h he f f ic itu ion t tra at etc sp ee su c as w ea r, s ,
2. 4	iv ing low ly be low d l im it Dr s : sp ee	b j h ic le l ing d h low Su t v tra at ec e ve a sp ee m uc er ha he d d l im it he h ig he ds t t te n p os sp ee w n r s p ee iat ar e a p p ro p r e.	ha de h u de d s d l im it. M 1 0 m e t te or n u n r p n r p os p ee
2. 5	iv ing low ly in lat ion he f f ic Dr to ot r t ot s re ra : n be low d l im it sp ee	b j h ic le l ing h low ha he Su t v tra t ot ec e ve m uc s er n r h ic les in f f ic ( bu bs ia l ly tra str t n ot ta nt ve ea m su be low he d d l im it ). t te p os sp ee
2. 6	l leg l p ht I ing ig a as s on r	b h le de l be ly d l leg l ly Su j ic i i t v te ec e ra an a p as se s he h ic le in he lan im d iat ly he ot t to t an r v e e m e e ig ht f t he he h ic le. ot r o r v e	h b le l b le he f f ic T is ia is ica in t a tra va r no p p av y
2. 7	I l leg l p ing ( he ha ig ht ) ot r t a as s n on r	Su b j h ic le he h ic le in fe t v t ec e p as se s a no r v e an un sa i l leg l m ( he ha he ig ht ). ot r t t or a an ne r n on r	Ex ing do b le l ine ing ig ht str . p as s ac ro ss u g o a , hr h lan ing ho l de ing in Pa t tu ou g rn e, us s u r. ss a ig h h ic le ( ) lan l d be H Oc Ve H O V N O T cu p an cy e, w ou i de d i l leg l p ing i f t he dr ive is he t co ns re a as s ev en r , ly he h le. in ic t on p as se ng er ve
2. 8	he im fe ing Ot r p ro p er or un sa p as s	b j h ic le he h ic le in Su t v t ec e p as se s a no r v e an im inc lu de d in iou ot p ro p er m an ne r n p rev s ies teg ca or	ing lan d w it h l im ite d ig ht Ex tw . p as s on o- e r oa s d ist he h ic le ot nt an ce or r v e p re se
2. 9	los fro f / be h d he Cu ing in in in tt to nt ot o c e o r , h ic le ve	b h le lan f a he h le Su j ic ic t v te t to ec e en rs e o no r v e o los ly he fro f t ha h ic le. b j h ic le Su to t nt t v t v c e o e ec e lan f a he h ic le los ly he te t to to t en rs e o no r v e o c e ba k o f t ha h le. ic t v c e
2. 1 0	k ing fro lan M tu a rn m w ro ng e	b j h ic le le ft ig ht fro lan Su t v tu ot ec e rn s or r m a e n int de d for k ing ha t t t en m a ur n.	k ing lan Ex tu . m a rn ac ro ss es
2. 1 1	D i d he h ic le du ing lan ha t s ot no ee r v e r e c ng e o r m erg e	Su b j h ic le lan int lan t v te ec e en rs a e o r m erg es o a e ho be f a he h le los by it ing ic ut t aw ar e o no r v e c e w

ia b le Va r	ia b le Va r na m e	ia b le de f in it ion Va r	les d h int Ex am p an s
#
		ha is lre dy l ing in ha lan t t tra t t a a ve e.
2. 1 2	Dr iv ing in he h ic le 's b l in d ot r v e zo ne	Su b j h ic le is l ing los he h ic le t v tra e t t ec e ve c o a no r v e
		h a ha he dr f t he he h le in ive ic is t t t ot su c ay r o r v e w
		d be b le it. b j h ic le Su t e te to to t v no xp ec a se e ec e
1 3	ive dr iv i f ic d ire d	int in h is lat ive it ion for lea t m t at st 5 s m us a a re p os Su b h ic le dr ive is dr iv in fu l	Ex h i b it d
2.	Ag ing ing cte g re ss sp ec m en ac , , ion t ac s	j ing ly t v ec e r a p ur p os e h de d for ive it ion int t re ss m an ne w ac s en a	ing . e x ro a rag e.
		ag g r, i f ic ip ien t. sp ec re c
2. 1 4	Ag ive dr iv ing he ot g re ss r ,	Dr ive is dr iv ing in ive ot r an ag g re ss m an ne r n	Ex k les dr iv ing it ho d ire d ing ut cte . r ec s w m en ac
		de i be d in iou ies In lu de k les ate sc r p rev s c g or c s r ec s	ion h ive d, ing in d t ac s, su c as ex ce ss sp ee w ea v an
		d dr iv ing ''s " ty an p or	f t f f ic i lg ing t o ta at ou ra ,
2. 1 5	i de f r d, k ing W t o rta ro ng s o oa no ve	b j h ic le is l ing he i de f t he Su t v tra t ec e ve on w ro ng s o
		d w it h int f p ing k ing t o ta ro a no en as s or ov er
2. 1 6	l low los Fo to	he h le. ic ot an r v e b h ic le is l fe d ist Su t v tra at
	ing ly o c e	j ing ec e ve an un sa an ce ( los ) be h in d he lea d v h ic le. to t o c e e
2. 1 7	Fa i le d ig l / Im ig l to s na p ro p er s na	Su b j h ic le fa i le d t ly ig l its int t v t ec e o p ro p er s na en	Ex ha d lan de it ho a t ut . c ng e es or m a ur n w
		by ig l ing l l. l ies lan d Ap t s at to no na a p p ne	ig l ing d r ig ht ig l w he k ing Ex tu s na . u se rn s na n m a a
		d de ive b j Su t s t m an eu ve rs, no u n e va s m an eu ve rs. ec	le ft tu rn
		h ic le fa i le d ly ig l its int by to t ve p ro p er s na en
		l ly. h p lan d m ig ing inc Us it ct s na or re e w ne an eu ve rs,
		d de ive t s no u n ev as m an eu ve rs.
2. 1 8	i de ig ht / le ft Im tu tu p ro p er rn , w r rn	b j h ic le d ig ht fro he in it ia l t l Su t v tu t ec e rn e r m rav e h, i ly h ing int he le ft at o t	ing int ing f f ic ing int Ex . t tra tu ur n o o nc om rn o , lan le ft. Ex ing int lan . t
		p un ne ce ssa r en cro ac d lan d b h le d le ft j ian Su j ic t t v tu a ac en e o r m e ec e rn e	wr on g e o n ur n o w ro ng e o n ht b. ig ing r or g o ov er cu r
		fro he in it ia l t l p h, i ly t at m rav e un ne ce ssa r
		h ing int he ig ht d j lan o t t en cro ac r a ac en e o r
		ho l de / b. s u r cu r
2. 1 9	ig ht / le ft Im tu t c tu p ro p er rn cu or ne r o n r rn ,	b j h ic le d ig ht fro he in it ia l t l Su t v tu t ec e rn e r m rav e	ing int lan ig ht ing Ex . tu rn o w ro ng e o n r or g o ov er ,
		h, i ly h ing int he ig ht at o t p un ne ce ssa r en cro ac r	b. int d j lan ing f f ic Ex ut t tra cu r . c s o a ac en e o r o nc om ,
		d j lan ho l de / b. Su b j h ic le t t v a ac en e o r s u r cu r ec e	ing int lan le ft ing tu rn o w ro ng e o n or g o ov er
		d le ft fro he l t l p h, in it ia tu t at rn e m rav e i ly h int he le ft d o t t	d ian m e
		ing j un ne ce ssa r en cro ac a ac en lan d ian e o r m e
2. 2 0	Im he tu ot p ro p er rn r ,	Su b j h ic le d le ft ig ht fro he in it ia l t v tu t ec e rn e or r m
		l p h in fe de i be d in tra at ot ve an un sa m an ne r n sc r
		iou ies ate p rev s c g or

Va ia b le r #	Va ia b le r na m e	Va ia b le de f in it ion r	Ex les d h int am p an s
2. 2 1	Im ba k ing d i d t s p ro p er c no ee ,	Su b j h ic le le d in it ho t v tra ut ec e ve re ve rse w bt f t he d in ing iew ing o a a p ro p er v o su rro un s be h in d t he h ic le. ve	Ex d i d he k m irr he ba k ing t c no c or s w n c
2. 2 2	Im ba k ing he ot p ro p er c r ,	Su b j h ic le le d in in fe t v tra ec e ve re ve rse an un sa de be d i in iou ies ot ate m an ne r n sc r p rev s c g or	Ex ba k ing int f f ic ba k ing int o t sta te c ra c up on er ,
2. 2 3	fro ke d it ion Im sta rt p ro p er m p ar p os	b j h ic le d fro ke d p it ion in Su t v ec e m ov e m a p ar os an fe un sa m an ne r.	d i d he k m irr Ex t c no c or s.
2. 2 4	D isr de d f f ic hm atc eg ar o er or w an	Su b j h ic le dr ive d i d ice be t v t n ot ec e r no or o y an f f ic f t he law f f ic d s de ing i tra to o er o or g ua r erv p ro v i da in f f ic f lo d t he dr iv ing k. tra ta g u nc e w an s	Ex lat e t ct. . u na wa re or o r ea
2. 2 5	S ig l v io lat ion ly d i d ig l t t s na ap p ar en no ee s na ,	Su b j h ic le dr ive d i d ice d t hu t v t n ot ec e r no an s d iso be d ( ly d iso be d ) f f ic ig l. a t y e or ne ar y e ra s na	Ex lat e t ct. . u na wa re or o r ea
2. 2 6	ig l v io lat ion int ion l ly d isr de d ig l S t na en a eg ar s na ,	b j h ic le dr ive f f ic ig l bu Su t v a t t ec e r s aw ra s na fu l ly d isr de d its in ion ( f dr ive I str t p ur p os e eg ar uc r ing be l low l ig ht be for it ha d s t to at wa ry a y e e p se d, de l v lat d t be l int ''S ig io ion ie ig tr at o r e co na o s na , ha ) ". c ng e	dr ive d t f f ic l ig ht bu de d Ex t p r s aw a re ra ro ce e hr h he int ion dr ive fa i le d t t t ct to ou g er se r o s p , be for k ing ig ht d. e m a a r on re
2. 2 7	S ig l v io lat ion ie d be ig l c ha tr to at na s na ng e ,	Su b j h ic le dr ive ler d o inu d a t v ate t t a ec e r a cc e r c on e d int de d t hr h int ion s t ct sp ee en o p as ou g an er se be for he f f ic l t d d. ig e t tra s na ur ne re	hr h ( d hr h ) Ex t t at te te to o t . w en ou g or m p g ou g int ion h i le l ig ht l low ct er se w w as y e
2. 2 8	ig io lat ion ly d i d Sto t t s sto p s n v ap p ar en no ee p , ig s n	b j h ic le dr ive d i d ice d t hu Su t v t n ot ec e r no an s d iso be d ly d iso be d ig to y e or ne ar y e a s p s n.	lat Ex e t ct. . u na wa re or o r ea
2. 2 9	Sto ig io lat ion int ion l ly ig t sto at p s n v en a ra n p s n , d sp ee	Su b j h ic le dr ive ig bu t v sto t ec e r s aw a p s n fu l ly dr hr h he int ion e t t ct at p ur p os e ov ou g er se a d g ha h. 1 5 m ate r t sp ee re n p	Ex fu l ly ig it ho de ler ing to ut at . p ur p os e ra n s p s n w ce be low d f h. 1 5 m a sp ee o p
2. 3 0	Sto ig io lat ion ''ro l l ing " sto p s n v p ,	Su b j h ic le dr ive d i d let t v t c e t ec e r no om o a co m p e ( d w be low ig in im 1 5 sto at to p a s p s n m um sp ee as h, bu bo h ). 0 m t a m p ve p
2. 3 1	Ot he ig ( ie l d ) io lat ion ly d i d n Y t ot r s n e.g v ap p ar en ., , ig se e s n	Su b j h ic le dr ive d i d ice d t hu t v t n ot ec e r no an s d iso be d f f ic ig ( he ha ig ). a t ot r t to y e ra s n n a s p s n	ie l d, On Y Do No t En te W e.g r, e ay .,
2. 3 2	he ig ( ie l d ) io lat ion int ion l ly Ot Y t r s n e.g v en a ., , d isr de d eg ar	b j h ic le ive f f ic ig ( he ha Su Dr t v a t ot r t ec e r s aw ra s n n ig ) bu fu l ly d iso be d ha ig to t p t t s a s p s n ur p os e y e n.
2. 3 3	Ot he ig io lat ion r s n v	Su b j h ic le dr ive d iso be d f f ic ig in t v a t ec e r y e ra s n a de be d i in iou ies ot ate m an ne r n sc r p rev s c g or
2. 3 4	ig d ing io lat ion No n- s ne cro ss v	b j h ic le dr ive de d t hr h Su t v ec e r p ro ce e ou g a n on ig d int ion in fe ct s ne er se an un sa m an ne r.	d i d he k t f f ic he ing dw Ex t c te no c ra w n en r ro a ay fro dr ive k ing lot m wa y or p ar
2. 3 5	R ig ht f-w in lat ion he h ic le to ot -o ay er ro r re r v e or it ion fa i lur t r p er so n, ap p ar en ec og n e	Su b j h ic le dr ive de he inc de is ion t v t ct ec e r m a or re c d ing ho ha d he ig ht f-w ( h is / he t reg ar w r -o ay r o wn h ic le he h ic le de ian ) du ot str e t ve or an r v e or p e o a isu de d ing f t he itu ion rst at m n an o s	Ex d i d he h ic le. t s ot no ee r v e

Va ia b le r #	Va ia b le r na m e	Va ia b le de f in it ion r	Ex les d h int am p an s
2. 3 6	R ig ht f-w in lat ion he h ic le to ot -o ay er ro r re r v e or de is ion fa i lur t p er so n, ap p ar en c e	Dr ive de he inc de is ion d ing ho t ct r m a or re c re g ar w ha d t he ig ht f-w ( h is / he h ic le r -o ay r o wn ve or he h ic le de ian ) du im ot str e t an r v e or p e o p ro p er ly is f t he itu ion at an a s o s	Ex d i d he h ic le ior ion bu t to t t se e o r v e p r ac is j dg d g m u e ap
2. 3 7	R ig ht f-w in lat ion he h ic le to ot -o ay er ro r re r v e or he kn ot p er so n, r o r u n ow n ca us e	Su b j h ic le dr ive de inc de is ion t v ct ec e r m a or re c d ho ha d he ht f-w ( h / he ing ig is t reg ar r -o ay r o wn w h ic le he h ic le de ian ) for ot str ve or an r v e or p e an kn for de i be d in ot un ow n re as on or re as on s n sc r iou ies ate p rev s c g or	R ig ht f-w in lat ion he h ic le to ot -o ay er ro r re r v e or he kn ot p er so n, r o r u n ow n ca us e
2. 3 8	d de im br k ing Su n or p ro p er a	b j h ic le br ke d d de ly in fe Su t v ec e a su n or an un sa in he dw bu d i d t t t c e t m an ne r ro a ay no om o a , let ( i.e d in d ica d i d n dr to to ot to co m p e s p sp ee r op ., ). ze ro	f t he d de br k ing lea ds d ire ly ing I ct to sto su n a p p ( d in d ica ), de d de ''s to s t sp ee r g oe o z er o co as u n or im ing dw " sto p ro p er p p on ro a ay
2. 3 9	d de im ing dw Su sto n or p ro p er p p on ro a ay	b j h ic le d ( d in d ica dr d Su t v sto to ec e p p e sp ee r op p e ) it ho le ing in fe to ut ze ro w am p w ar n or an un sa in he dw t m an ne r ro a ay	ha d lat br k ing / de ly he dr ive Ex r or e a co on w n r d in d ica de ''s d de to s t sp ee r g oe o z er o- -co u n or im br k ing " o he ise T he ly im t t e t p ro p er a rw on o de br k ing d ing for l d be sto nt co a an p p on e e ve w ou he he br k ing do lea d d ire ly he 't t ct to t w n a es n ing ( b j br ke he b it lat ha sto t t s t p p su ec a s, n a er o d de ly ). sto su n p
2. 4 0	k ing in im da loc ion Pa at r p ro p er or ng er ou s	b j h ic le ke d ( d w it h he int f Su t v sto t t o ec e p ar p p e en in ing d ) in loc ion int de d for sto at t re m a p p e a no en k ing p ar	ho l de f Ex Int sta te . s u r o er
2. 4 1	lur d he d l ht Fa i im ig e t o a s	b h le l h h h be Su j ic ing it ig t v tra ec e ve am s w iva d he d l ig ht it ho d im ing he t te ut t ac on a s, w m l ht he iat
2. 4 2	Dr iv ing it ho l ig ht it h in f f ic ien l ig ht ut t w s o r w su s	ig s w n ap p ro p r e. Su b j h ic le l ing it h n he d l ig ht ( t v tra ec e ve w o a s o n or in f f ic ien he d l ig ht ) he he itu ion ire t n t at su a s w s re q u s h l ht for fet
2. 4 3	Av i d ing de ian str o p e	ig ing su c sa y. Su b j h ic le dr ive be ha d in t v ec e r ve a m an ne r de d d f l h de int i ict it ian to str en av o co n a p e w
2. 4 4	i d ing he h ic le Av ot o r v e	b j h ic le dr ive be ha d in Su t v ec e r ve a m an ne r int de d i d f l ict it h he h ic le. to ot en av o co n w an r v e
2. 4 5	Av i d ing im l o an a	Su b j h ic le dr ive be ha d in t v ec e r ve a m an ne r de d d f l h l. int i ict it im to en av o co n w an an a
2. 4 6	Ap fam i l iar ity it h dw t u p ar en n w ro a ay	b j h ic le dr ive be ha d in fe Su t v ec e r ve an un sa ly du fam i l iar ity it h t e t m an ne r, ap p ar en o a n un w he d f f ic loc l ing itu ion ity t tra at su rro un s or a	d d ing Ex U- te tu et . r ep ea rn s, re a m ap s, p ap er s, c.

Va ia b le r #	Va ia b le r na m e	Va ia b le de f in it ion r	Ex les d h int am p an s
2. 4 7	Ap fam i l iar ity it h v h ic le t u p ar en n w e	Su b j h ic le dr ive be ha d in fe t v ec e r ve an un sa ly du fam l h i iar ity it t e t m an ne r, ap p ar en o a n un w he h ic le. t ve	Ex fam i l iar ity it h d isp lay d ls nt . u n w s a n co ro
2. 4 8	Ap l ine ien dr iv ing t g p ar en en er a xp er ce	Su b j h ic le dr ive be ha d in fe t v ec e r ve an un sa ly du lac k o f e h ien it t e t m an ne r, ap p ar en o xp er ce w he dr iv ing k. t ta s	Ex hy -fo d dr iv ing ly iou ut p er cu se ov er ca s , et m an eu ve rs, c.
2. 4 9	ing i de Fo to rc s	he dr ive ies for he dr ive he i de T r t to ot r t o t r ce an r s f t he d o ro a
2. 5 0	Bu m p	he dr bu he dr bu h T ive ive 's it t r m p s a no r r m p er w h is / he r o wn
2. 5 1	Ho / F las h L ig ht rn s	T he dr ive be h is / he ho f la he l ig ht t r ep s r rn or s s s a he dr ive ot an r r	W he lea ly d ing ig l. t u n c r no se as a w ar n s na
2. 5 2	loc k ing B	ive b loc ks he dr ive fro do ing ha he / Dr ot t r an r r m w he do dr ive he dr ive E.g ts to ts ot s w an r p rev en r rs fro ing m p as s
2. 5 3	l Un int ion La De iat ion t en a ne v	b h le dr fa ls Su j ic ive i int in iat t v to ec e r m a a ap p ro p r e d fe lan it ion d int ion l ly dr i fts t an sa e p os an un en a ds d / lan l ine to wa r an or ov er on e o r m or e e s.
2. 5 4	Ot he r	Su b j h ic le dr ive in he be ha ior t v ot ec e r e ng ag es r v de be d i in iou ies t ate no sc r p rev s c g or
2. 5 5	kn Un ow n	de ine he be ha ior ( ) d in by Ca ot te t nn rm v s en g ag e he b j h ic le dr ive du l im ita ion in t t v e t t su ec e r o s i de iew l ig ht ing isu l o bs ion tru ct v o v s, , v a s, or l im ite d ive t p er sp ec	f t he i de is iss ing he is Ex Pa rt t o v o m or re in f f ic ien in for ion in he i de ke t at t o t su m v o m a a de ina ion te t rm
3	iza io f ive ha io Ca te t Dr Be g or n o r v r	iza io f ive ha io Ca te t Dr Be Er g or n o r v r a s ro r, V io la io Un las i f ie d Un fe ( fe be ha io t n, c s sa un sa v r bu it he io la io ), he be ha io t n t t e r e rro r n or v n o r v r
3. 1	Er ro r	he bs d be ha ior ly T t o erv e v w as ap p ar en a is j dg int ion l. t, t m u em en un en a	int ion l lan de iat ion ig l v io lat ion E.g t un en a e v ; s na ., , ly d i d ig l. t t s ap p ar en no ee s na
3. 2	V io lat ion	T he bs d be ha ior ly de l i be t te o erv e v w as ap p ar en ra	E.g Ho / F las h l ig ht he lea ly d t u rn s, w n c r no se as a ., l. ing ig wa rn s na
3. 3	las i f ie d fe Un Un c s sa	he be ha ior fe, ho it is i b le T t p v w as un sa we ve r no os s lua he he he be ha ior to te t r t ev a w v w as l o l. int ion int ion t t un en a r en a
3. 4	he ha ior Ot Be r v	he ies d v io lat ion do ly T teg t a to ca or er ro r a n no p p he bs d be ha ior t o erv e v

ia b le Va r #	ia b le Va r na m e	ia b le de f in it ion Va r	les d h int Ex am p an s
4	da k Se Ta co n ry s	bs b le dr ive in f t he O t er va r e ng ag em en an y o l is d da ks be in in in te ta t a t se co n ry s g n g a ny p o , du in he im is io in lu de Ev D t t t tra t r g en e. c ns c no n dr iv in la d lan fro he d ire io te t t g re g ce s a wa y m c n f v h ic le in lu de ks t. Do t ta o e m ov em en es no c s ha it ica l t he dr iv in k, h t t a t ta re cr o g s su c as do he ks irr / b l in d he ks te t c sp ee m e r c c m or sp o c , , iva in ip / he d l ig ht h i ft in t t ac g w er s a s, or s g g ea rs O he it ica l t ks in lu de d, in lu d in t r n on -cr as ar e c c g d io d j be lt d j tm ts, t tm ts, ra a us en se a a us en in do d j d iso d irr tm ts, w w a us en an v r a n m or d j ha he is low l im it tm ts. No te t t t a us en re no er fo k du io r t t as ra n.	de ks in d i f t he in io Co ta te te te t s as rru p rru p n las fo 5 s lo les he ise d. ts t ta te r or ng er un s o rw s ( du io da d ). de l l S H R P 2 r Co t ta e c n s n r a in io fo in d j ''P A Se " te t t t rru p ns r as se ng er ac en a d ''P in Re Se " in io i f t he t te t an as se ng er ar a ra c ns in io is he in io te t t te t rru p n an o r p as se ng er ra c n k. de he k in d i f t he Co ta t ta te te s s as rru p in io las fo lo fo k in te t ts 1 s ''L rru p n r or ng er r oo g io h in i de / de ian / im l / t p nt tr a re v us cr as or c p e s an a b j l t he h ic le / is d by D t e xt t tra te o ec er na o ve c io / he l d is io O ". tru t t xt tra t co ns c n r e er na c n
4. 0	da ks No Se Ta co n ry s	he b h le dr d T j ic ive is in t v t e su ec e r no ng ag e an y bs b le da ks d is ive he ta at te nt to t o erv a se co n ry s an dr iv ing k. ta s
4. 1	Ta l k ing / ing ing d ien kn s au ce un ow n ,	Su b j h ic le dr ive is ing l ip i f t l k ing t v ec e r m ov s a s a or ing ing he int ion is be l iev d t be it h a t t t s er ac no e o w , h is inc lu de h ist l ing h is T T teg p as se ng er ca or y s w do inc lu de he dr ive l k ing teg t t r t to ca or y es no a a de he kn de he ian i str ot ty ts t p e or r ow n p ar ou h ic le, h ic h ho l d be de d he iat t ve w s u co as ap p ro p r e l d ist ion ter ct ex na ra	T h is Se da Ta k inv lve ing ing it h d io, co n ry s o s s w ra l k ing l f, ta to etc se
4. 2	Da ing nc	Su b j h ic le dr ive is ing h is / he he d, t v ec e r m ov r a rm s, a he bo dy ing ly in im it h t he be ot t s t at or r p ar ee m e w f m ic. o us	ing ha ds / f in ing he l, ta ste e.g p p n g er s o n er w e ., bo b b ing he d, ir dr ir ita ''a " o ''a ". a um s r g u r
4. 3	Re d ing a	Su b j h ic le dr ive is d ing ia l t ha is in t v ate t ec e r re a m r he h le, bu f t he h le ( ic ic i.e t t n ot t o t ve a p ar ve no ., d ing l s ig k d isp lay ). te nt sta re a ex rn a ns or ce er c ,	T h is l d be d ing d ire ion ia l, ct ate co u re a s, p ap er m r ka f r d ho be d ing I ing p ac g ea a p ne nu m r, re co r as d ia l ing l l p ho ce ne
4. 4	it ing W r	Su b j h ic le dr ive is it ing it h / i l t v ec e r w r w a p en p en c ing lu b let sty a t or us a s o n a	ive l d be it ing iec f p k ing Dr r c ou w r on a p e o ap er m a , b let te a t etc no s o n a ,
4. 5	d l k it ho Pa in j ing t s t – ta t sse ng er a ac en ea w u lo k in to o g p as se ng er	fro b le b le, bu A is is i is i nt at t v t se p as se ng er or no v he b j h ic le dr ive is lea ly l k ing it h t t v ta su ec e r c r w a in he d j / fro h i le l k ing t t nt at. W ta p as se ng er a ac en se l he dr do loo k t ds he ist ing ive t t t or en r es no ow ar , p as se ng er	h d f t he dr l k he Us is ist ion i ive is ing e t ct ta to t ra r fro bu is isu l ly d ist d; nt at t t v cte se p as se ng er no a ra inc lu de l ist ing En ire ip t to t tr s a p p ar en en p as se ng er f i le be d t loo k for de f i t m m or se g en ay us e o ev nc e o i de h is d ist ion lon he Co r t ct t p as se ng er ns ra as g as dr ive d bo h in in he h ic le t t r a n p as se ng er re m a ve ( i f t he is i d l ing ). to ev en ca r s p s o r

ia b le Va r #	ia b le Va r na m e	ia b le de f in it ion Va r	les d h int Ex am p an s
4. 5. 1	Pa in d j l k ing it h lo k in t s t – ta sse ng er a ac en ea w o g to p as se ng er	A fro is is i b le is i b le, bu nt at t v t se p as se ng er v or no he b j h ic le dr ive is lea ly l k ing it h t t v ta su ec e r c r w a in he d j / fro h i le l k ing W t t nt at. ta p as se ng er a ac en se l ist ing he dr ive loo ks ds he t to t or en r wa r , p as se ng er	Us h is d ist ion i f t he dr ive is l k ing he e t ct ta to t ra r fro d is isu l ly d ist d; nt at cte se p as se ng er an v a ra inc lu de l ist ing ire ip En t to t tr s a p p ar en en p as se ng er f i le be d t loo k for i de f t m or se g m en ay us e o ev nc e o Co i de h is d ist ion lon he r t ct t p as se ng er ns ra as g as dr d bo h he h le ive in in ic t t r a n p as se ng er re m a ve ( i f t he is i d l ing ). to ev en ca r s p s o r
4. 2 5.	in d j int ion bu Pa t s t – t t sse ng er a ac en ea an y er ac l k ing ta	A fro is is i b le is i b le, bu nt at t v t se p as se ng er v or no he b j h ic le dr ive is lea ly int ing it h t t v t su ec e r c r er ac w in he d j / fro h is l d T t t nt at. a p as se ng er a ac en se co u be ing ds ing d o fro tu to to g es r wa r m ov wa r r a wa y m , he h ing ke h ing fro t to ta et p as se ng er or re ac so m m , h he ive ing et to t or g so m p as se ng er	f t he dr ive is l k ing h i le int ing de he I ta t t r w er ac co as , iat ''ta l k ing " c En ire ip f i le ate t tr ap p ro p r e g or y. or be d t loo k for i de f t m se g m en ay us e o ev nc e o i de h is d ist ion lon he Co r t ct t p as se ng er ns ra as g as dr ive d bo h in in he h ic le t t r a n p as se ng er re m a ve ( f t he d l ). i is i ing to ev en ca r s p s o r
4. 6	in l k ing it ho lo k in Pa at - t t to sse ng er re ar se a w u o g p as se ng er	is is i b le is i b le, bu A at t v t re ar se p as se ng er v or no he b j h ic le dr ive is lea ly l k ing it h t t v ta su ec e r c r w a in he W h i le l k ing t at. ta p as se ng er re ar se or l ist ing he dr ive do loo k t ds he t t t en r es no ow ar , p as se ng er	h is d ist ion i f t he dr ive is l k ing he Us e t ct ta to t ra r bu is isu l ly d ist d; at t t v cte re ar se p as se ng er no a ra inc lu de l ist ing En ire ip t to t tr s a p p ar en en p as se ng er f i le be d t loo k for i de f t m or se g m en ay us e o ev nc e o i de h is d ist ion lon he Co r t ct t p as se ng er ns ra as g as dr ive d bo h in in he h ic le t t r a n p as se ng er re m a ve ( f t he d l ). i is i ing to ev en ca r s p s o r
4. 6. 1	in l k ing it h lo k in Pa at - t to sse ng er re ar se a w o g p as se ng er	is is i b le is i b le, bu A at t v t re ar se p as se ng er v or no he b j h ic le dr ive is lea ly l k ing it h t t v ta su ec e r c r w a in he W h i le l k ing t at. ta p as se ng er re ar se or l he dr loo ks ds he ist ing ive t to t en r wa r p as se ng er ,	h is d ist ion i f t he dr ive is l k ing he Us e t ct ta to t ra r d is isu l ly d ist d; at cte re ar se p as se ng er an v a ra inc lu de l ist ing En ire ip t to t tr s a p p ar en en p as se ng er f i le be d t loo k for de f i t m or se g m en ay us e o ev nc e o i de h is d ist ion lon he Co r t ct t p as se ng er ns ra as g as dr ive d bo h in in he h ic le t t r a n p as se ng er re m a ve ( f t he d l ). i is i ing to ev en ca r s p s o r
4. 6. 2	in int ion bu l k ing Pa at t t t sse ng er re ar se - a ny er ac a	is is i b le is i b le, bu A at t v t re ar se p as se ng er v or no he dr ive is lea ly int ing it h a in t t r c r er ac w p as se ng er he T h is l d be ing d o t at. to re ar se co u m ov wa r r a wa y fro he h for he ing t t at m p as se ng er or re ac re ar se , / h ing ke h ing fro ive to ta et p as se ng er re ac so m m or g	f t he dr ive is l k ing h i le int ing de he I ta t t r w er ac co as , iat l k ing ire ip f i le ''ta " c En ate t tr ap p ro p r e g or y. or be d t loo k for i de f t m se g m en ay us e o ev nc e o de h d lon he Co i is ist ion r t ct t p as se ng er ns ra as g as dr ive d p in in he h ic le ( i f t r a n as se ng er re m a ve ev en
4. 7	M ing b j in h ic le t ov o ec ve	h ing he et to t so m p as se ng er An b j in i de he h ic le, h ic h is be ing t t t o ec s ve w no he l d by he dr ( ) ( f p ) ive i is in t nt r o r p as se ng er s re se ion it he du he ion f t he h ic le ot e t o t ot m e r m o ve or , du he hr ing he b j e t t t t t. o a no r p as se ng er ow o ec	he is i d l ing ). t to ca r s p s o r Ex b j fe l l o f f s he dr ive d ha d a t t w to t . o ec ea n r s p p e r f f ic l ht ig a t ra

ia b le Va r #	ia b le Va r na m e	ia b le de f in it ion Va r	les d h int Ex am p an s
4. 8	in h ic le In ct se ve	ion it h in in he h ic le ( Int t ct t er ac w an y se ve e.g ., ing in ing bo dy i d in att at ct, to ct, sw se m ov av o se loo k ing d ing loc in ). try to ate ct ar ou n se
4. 9	in h ic le Pe t ve	int ion it h in he h ic le bu An t et t t y er ac w a p ve l k ing inc lu d ing ho l d ing ing ing ta ett et. p or m ov p , , ,	ly de i f a im l / is is i b le int in On et at co n a p v so m e p o he ip f i le i f t he is h ist / it h t he t tr nt t w or re or y co ex dr ive d he dr ive is h i b it ing be ha ior ha t s t t r a n r ex v ha he h le. iat ing in ic e t et t ar e a p p ro p r o a p ve v
4. 1 0	l k ing Ta to et p	ive is l k ing Dr ta to et r p
4. 1 1	O b j dr d by dr ive t ec op p e r	Su b j h ic le dr ive is in it ia l ly ho l d ing t v ec e r h ing d dr it d t he dr ive he et r t so m an op s an n im d iat ly ic ks it ba k u k ing he dr ive 's ta t m e e p c p, r ion fro he dr iv ing k. att t t ta en aw ay m s	h is de he ''re h ing " T teg t ca or y su p er se s o r ac ies in he itu ion f a b j be ing teg t at t ca or s o n o ec dr d d im d iat ly iev d. tr op p e an m e e re e
4. 1 2	Re h ing for b j he t, ot ac o ec r	Su b j h ic le dr ive he for b j t v t n ot ec e r r ea c s an o ec de be d he lu de b i in In j in ot ate ts sc r an y r c g or y. c s o ec sto tm ts. rag e c om p ar en	On he dr ive ha f in is he d r h ing for he b j t t t ce r s ea c o ec d ha ha d ( f n be d for de d it in i ing int ot m an s n ov e en ), he it be b j in h ic le, he ''o " t t ot us ag e n co m es ec ve r, lon it do 't f it int f t he he ot as g as es n o a ny o r ies ( ing dr in k ing ) teg t etc ca or e.g ea ., , ,
4. 1 3	b h le, he O j in ic t ot ec ve r	b h le dr lea ly loo k Su j ic ive is ing t v at ec e r c r , ha d l ing ho l d ing ip lat ing b j t n or m an u an o ec , , ( is i b le ) h ing loc d in he h ic le, he t t ate t ot v or no or ve r ha ho l ist d in he ies t t ot ate n se e r c g or
4. 1 4	l l p ho l d ing Ce Ho ne ,	b j is ho l d ing l l p ho bu Su t t n ot ec a ce ne ip lat ing it. l d be ho l d ing it in ha d, lap Co m an u u n or , he t so m e o r w ay
4. 1 5	l l p ho l k / l ha d- he l d Ce Ta ing ist ing ne en n , ,	b h le dr l k ha d he l d Su j ic ive is ing t v ta ec e r on a n ho ha ho i f l ist ing to to p ne or s p ne up ea r a s en a ho ion it ing for he t t p ne co nv er sa or w a p er so n y ar e l l ing ic k u he ho I f dr ive ha to t ca p p p ne r s a n he ds he dr be bs d iec ive et, t t ea rp e o r a r m us o erv e l k ing d ly. ta te re p ea
4. 1 6	Ce l l p ho Ta l k ing / l ist ing ha ds -fr ne en n ee , ,	Su b j h ic le dr ive is l k ing l ist ing t v ta ec e r or en on a ho ha ds -fr de h a he ds ing ice et, p ne us a n ee su c s a a v in- h ic le int d s ha ds -fr ke te ste ve eg ra y m or n ee sp ea r , ho f dr ive ha iec he ds he I et, t p ne r s a n ea rp e o r a dr ive be bs d l k ing d ly. t ta te r m us o erv e re p ea	I f t h is be l ia b ly de ine d c de teg ot ter ca or y ca nn re m o l k / d kn ''T ing ing ing ien " as a s au ce un ow n ,

Va ia b le r #	Va ia b le r na m e	Va ia b le de f in it ion r	Ex les d h int am p an s
4. 1 7	Ce l l P ho Te ing / Br ing / Ce l l P ho D ia l ing xt ne ow s ne , , ha d- he l d / d l ha d- he l d k ke ia ing ing ic n n us q u s y	Su b j h ic le dr ive is ing bu t v tto ec e r p re ss ns or a h he l l p ho d / to t to te uc sc re en on ce ne cr ea an or d / b j h ic le dr ive is Su a t t m t v se n ex es sa g e. ec e r ing bu h he l l tto a t t p re ss ns or ou c sc re en on ce ho br he int ho to t t o p ne ow se er ne r p ne l lso lu de ds ica ion M inc ice t ap p s. ay a vo co m m an ( ir i ). / b j h ic le dr ive is h ing S Su t v e.g ec e r p us ., be bu l l p ho h s tto to to nu m r ns on a c e ne or uc cre en d l a be br / he k s h lse ia ing et nu m r o r ow se c c om e on he ir l l p ho ( h is l d lso inc lu de d ing t t ce ne w ou a re a he ho be fro he f p ). / Su b j t et t p ne nu m r m a s o ap er ec h ic le dr ive is h ing ic k ke bu ( tto ve r p us q u y ns e.g ., d d l ) ia sp ee
4. 1 8	l l ho ia l ing ha ds -fr ing ice Ce P D ne n ee us vo , iva d ftw t te ac so ar e	ive ks int iva d l l p ho Dr t te r s p ea o o p en or ac ce ne , he ds in- h ic le int d de ice for he et, te t a or ve eg ra v f d ia l ing it h lon ior de lay f n p ur p os e o w g, p r o o k ing int de ice ( i.e l i ke ly in ior t t sp ea o v m os no p r ., ion ) d n ha bu t e t tw tto co nv er sa an o m or n o ne or o n ( h be in ) ho iec to p re sse s e.g p us g on p ne ea rp e, ., , he ds h le d de in- ic int et, te ste a or ve eg ra sy m ar e m a f ir st.
4. 1 9	Ce l l ho ing / h ing / ing P Lo t ne ca re ac an sw er ,	Su b j h ic le dr ive is lan ing f in d c l l p ho t v to ec e r g c e ne , h ing ds h is / he l l p ho d / to re ac wa r r c e ne an or , f l ho bu ip ing ing tto n t p p ne op en or p re ss a o a ns we r l l. a c a	f m ha d ist ion ha ( dr ive I e t ct or n on e ra p p en s e.g r ., loo ks for ho he for it d t he it ), p ne re ac s an n a ns we rs , he las fra be l d be he las in t t t t s tep m e n um r w ou h is ( ing l l p ho ). On t se q ue nc e e.g an sw er ce ne ce ., ho is dr ive 's ion ha lea ly at t p ne r ea r o r c on ve rsa s c r be de he iat ''ta l k ing " c t ate g un co as ap p ro p r e g or y. ,
4. 2 0	l l p ho he Ce ot ne r ,	b j h ic le dr ive is int ing it h l l Su t v t ec e r er ac w a c e ho in ( loo k ing l l p ho at p ne so m e m an ne r e.g a c e ne ., j ho l d ing it, bu i ly ip lat ing t t n ot or us ne ce ssa r m an u he l l p ho ), do f it in ion in t t t ce ne an wa or ac es no y y he ot ate an y r c g or y.	lu de lug ing ho int ha lea ing In c s p g p ne o c rg er c n , ing he ds ut t et, et sc re en p on a c. ,
4. 2 1	b let de ice ing / h ing Ta Lo t v ca re ac ,	Su b j h ic le dr ive is do ing h ing it h t v t ec e r an y w a b let de ice ( he lan d for ta ta rts to v re ac s o r s g ce ar ou n lec ic b let de ice ( i d ). Pa tro ta an e n v e.g .,
4. 2 2	b let de ice ing Ta Op at v er ,	b j h ic le dr ive is ing bu Su t v tto ec e r p re ss ns on or ing he h he lec ic b let t to t tro ta us uc sc re en on e n de ice v

ia b le Va r #	ia b le Va r na m e	ia b le de f in it ion Va r	les d h int Ex am p an s
4. 2 3	b let de ice iew ing Ta V v ,	Su b j h ic le dr ive is ho l d ing d loo k ing t v at ec e r an an lec ic b let de ice bu ing tro ta t n ot e n v p re ss an y , bu tto ns
4. 2 4	b let de ice he Ta Ot v r ,	b j h ic le dr ive is int ing it h Su t v t ec e r er ac w an lec ic b let de ice in tro ta ot e n v so m e m an ne r n de i be d in he ies ot ate sc r r c g or	lu de lug ing b let int ha lea ing In ta c s p g o c rg er c n , he ds ho l d ing it ho ip lat ing et, ut etc sc re en a w m an u , ,
4. 2 5	d j ing / ito ing l im l A t ate nt us m on r c co ro	b j h ic le dr ive int it h in- h ic le Su t v ts ec e r er ac w ve l im l s it he by h ing he ate nt ste to t c co ro y m e r uc l im l bu lan ing he l im ate nt tto at t ate c co ro ns g c c , l o da h bo d, d l l j ing im nt t ate nt co ro n s ar or a us c co ro nt ve s.
4. 2 6	A d j ing / ito ing d io t us m on r ra	Su b j h ic le dr ive int it h in- h ic le t v ts ec e r er ac w ve d io / d io it he by h ing he d io ste to t ra au sy m e r uc ra bu da h bo d he l, ing tto ste ns on s ar or er w e or lan ing he d io da h bo d. at t g c ra on s ar
4. 2 7	In ing / iev ing C D ( im i lar ) rt tr se re or s	Su b j h ic le dr ive ic ks C D, t v tte ec e r p up ca sse or , he de ( he ha lay ) ic ice M P 3 p ot sto ot r t r m us rag e n er v in h ic le d / in it int d io, rts ve an or se o r a p re sse s a ny bs bu de ice lay / nt tto to et to su eq ue ns g v p in d / fas for d d t he lay dr ive t rew wa r an n p or r , bu de d t he lac j ice it tto n t t p re sse s o e ec an n p es v he in h ic le. so m ew re ve
4. 2 8	A d j ing / ito ing he de ice int l t t ot us m on r r v s eg ra o h ic le ve	Su b j h ic le dr ive int it h t v ts ec e r er ac w a fac in l le d de ice he ha ho tu sta ot r t t m an u re r- v n se l ist d in he ies it he by h ing ot ate to e r c g or e r uc or , lan ing he de ice inc lu de dr iv ing Do at t t g c v es no it ica l t ks h ig l, w ip tu cr as su c as rn s na er s, , he d l ht h ft, do ig i ete a s, g ea r s sp ee m r.	lu de int ion it h be lt, do loc ks In t at c s er ac w se or , in do ls, Na ig ion iso nt at ste w w co ro v sy m su n v rs, , iew irr inc lu de iev ing Do etc t tr re ar v m or es no re , b j in i de ts sto tm ts. o ec s rag e c om p ar en

Va ia b le r #	Va ia b le r na m e	Va ia b le de f in it ion r	Ex les d h int am p an s
4. 2 9	Lo k ing iou h inc i de / de ian / at nt str o p rev s c ra s or p e im l / b j l t he h ic le / D ist d by t e xt o t cte an a o ec er na ve ra ion / he l d ist ion Ot tru ct xt ct co ns r e er na ra	Su b j h ic le dr ive is loo k ing i de f t he t v ts ec e r ou o h ic le in he d ire ion f w ha is bv iou ly t ct t ve o o s an i de im i lar inc i de / de ian it he nt nt str ac c or s p e e r o n he i de f t he d o in fro f t he ( i.e ing t nt s o ro a r o m us a ., l k o i d ing b i ke d l ig ht ) / im l at cro ss w a r r a a re an a he he de f t he d ( h l d be it i is t t t e r o n s o ro a ou no w d for im l c ing he d ) / b j he t t o n t us e an an a ro ss ro a o ec i de f t he d ( bo ). / Su b j h ic le dr ive t v s o ro a e.g a x ec e r ., is loo k ing i de f t he h ic le in he d ire ion f ts t ct ou o ve o ion ion l d be A str t tru ct a c on uc zo ne co ns zo ne w ou de f in d a he f a ba l, in ha d s t e p re se nc e o rre p er so n a r ha ion ip h ic les / Su b j t, tru ct t o t co ns eq u m en r v e ec h le dr loo k de f t he h le for ic ive is ing i ic ts ve r ou o ve de i be d in iou ies t ate p ur p os es no sc r p rev s c g or or , for kn he lan is t an un ow n re as on w n g ce no i de d be f t he dr iv ing k. to t o ta co ns re p ar s	Co de he k a int d i f t he int ion t ta te t s s er ru p er ru p las for 1 s lon ts or g er
4. 3 0	h ing for foo d- lat d dr in k- lat d ite Re ac re e or re e m	b j h ic le dr ive is loo k ing for h ing for Su t v ec e r or re ac ite lat d t ing dr in k ing at an y m re e o e or
4. 3 1	Ea ing / Dr in k ing t	Dr ive / dr in ks at r e s
4. 3 2	h for / Re ing ig ig ett ac c ar c ar e	b h le dr he lan Su j ic ive t v ta rts to ec e r r ea c s o r s g ce d for ig / ig lat d ite ett ar ou n c ar c ar e o r r e e m s.	he ha d a d be d w h t he On ite is in ing it t ce m n n m ov e int de iat t t ate en o u se co as ap p ro p r e u sa g e c g or y , ( l ig ht ing ). e.g .,
4. 3 3	L ig ht ing / Sm k ing / Ex ing is h ing ig / ig t ett o u c ar c ar e	Su b j h ic le dr ive is in f t he t v tag ec e r so m e s e o f l ht / / ha l ig ing ig ig it ett p ro ce ss o c ar c ar e s a ig / ig it he in he ir h ha d / ett t t ut c ar c ar e e r m ou or n p s h is / he ig / ig ha ds it t ett to ou r c ar c ar e, n so m eo ne lse it he in do to t t e or sse s ou w w.
4. 3 4	h ing for l bo dy lat d ite Re ac p er so na -re e m	, b j h ic le dr ive is h ing for ite Su t v ec e r re ac an y m lat d l hy ien he lt h, do to t. re e p er so na g e, a or a rn m en	h ing for b, br h, ke de l Ex nt . re ac co m us m a up ra zo r, a , f lo len las ( ly be ing nt t t c nt ss, co ac se s, g se s no ur re ), ha ( ly be ing ). On he t t c nt t wo rn no ur re w or n ce ha d a d be d w h t he ite is in ing it int t t m n n m ov e en o de iat ate us e, co as ap p ro p r e u sa g e c g or y.
4. 3 5	Co b ing / br h ing / f ix ing ha ir m us	Su b j h ic le dr ive is d j ing b ing / t v t ec e r a us or co m , br h ing ha ir, for ic k ly ip ing ha ir t t us ex ce p q u sw ou f e d le l f i ir ing tw o y es or o
4. 3 6	Ap ly ing ke p m a -u p	b j h ic le dr ive is in f a ly ing Su t v tag ec e r so m e s e o p p bo dy du bo dy ct to an y p ro	lot ion ke l ip ba lm fu Ex m a -u p, p er m e ., , ,
4. 3 7	ha S ing v	b h le dr l h a Su j ic ive is ing ian it t v ec e r us an ap p ce y w b la de ha ir fro bo dy to re m ov e m	( lec l ) Ex ic tr ra zo r e or m an ua .,

ia b le Va r #	ia b le Va r na m e	ia b le de f in it ion Va r	les d h int Ex am p an s
4. 3 8	h ing / f lo ing h Br te et us ss	b j h ic le dr ive is ing l ian Su t v to ec e r us an y ap p ce br h, f lo he ise lea h h. ot tee t t us ss or rw c n or m ou	inc lu de h br h, f lo hp ic k, Ex s t t to ot et oo us ss, c. .,
4. 3 9	B it ing i ls / ic les t na cu	Su b j h ic le dr ive is b it ing i ls ic les t v t ec e r na or cu
4. 4 0	ing / d j ing lot h ing Re t m ov a us c	b j h ic le dr ive is ing d j ing Su t v t ec e r re m ov a us or , , ing lot h ing inc lu d ing j ke h irt ut t ts, ts, p on c ac s p an , , ho ks ha lov kt ies d ts, s es so c g es ne c an sc arv es , , , ,
4. 4 1	/ d lry Re ing j ing j t m ov a us ew e	b h le dr d Su j ic ive is ing j ing t v t ec e r re m ov or a us j lry inc lu d ing he atc ew e w s. ,	k lac br let he Ex ing ing atc r s, ne c es ac e s, w s, ea rr s o r ., , he ier ing ot r p c s.
4. 4 2	ing / ins ing / d j ing len Re t t nt t m ov er a us co ac se s o r las g se s	Su b j h ic le dr ive is ing in ing t v rt ec e r re m ov or se len ( ) fro ( ) ing / k ing nt t ut t ta co ac s es m ey e s or p on f f / d j ing las las t o a us g se s o r s un g se s.
4. 4 3	he l hy ien Ot r p er so na g e	b j h ic le dr ive is d in he Su t v t ec e r en g ag e so m e o r l hy ien iv ity ( ies ) de i be d in ct t p er so na g e a no sc r iou ies ate p rev s c g or	he k ing l f in irr it ho he Ex ut t c c on es e m or w ., d ing ks ing h ing f ta try to et et t o p re ce s g so m ou , 's on e ey e.
4. 4 4	he i f ic int l e lan Ot r n on -sp ec er na y e g ce	b j h ic le dr ive lan fro he Su t v t ec e r g ce s a wa y m d ire ion f t l a h ing in i de he b j ct t s et t t o rav e om s su ec h ic le, bu de ine i f ic lan t c t te ve an no rm a sp ec g ce loc ion at
4. 4 5	he kn da k Ot ta r ow n se co n ry s	b j h ic le dr ive is d in iza b le Su t v ec e r en g ag e a r ec og n da k t ha is l ist d in he ies ta t t ot ate se co n ry s no e r c g or P lea de f in in he No ion he Ex l Lo t te t t se e s s ec on ce g	lu de he he h ic le is l ing in In t tra c s c as es w re ve ve d t he dr ive is loo k ing he for d o t t rev er se an r ou wa r r i de in do ( he ha i de irr ), he ha ot r t t r t s w ws n s m or s ra n he dw be h in d he h ic h is he t t t ro a ay ca r, w no w d ire ion f t l. ct o rav e
4. 4 6	Un kn ( da k p ) ty ta nt ow n p e se co n ry s re se	Su b j h ic le dr ive is lea ly d ist d fro he t v cte t ec e r c r ra m dr k, bu he f ic d iv ing i ist ion is ta t t ct s sp ec ra kn un ow n.
4. 4 7	kn Un ow n	Ca de ine he he he b j h ic le ot te t r t t v nn rm w su ec e dr ive is d in da k du ta e t r en g ag e a se co n ry s o l im ita ion in i de iew l ig ht ing isu l t s v o v s, , v a bs ion l im ite d ive tru ct t o s, or p er sp ec	f t he i de is iss ing he is Ex Pa rt t o v o m or re in f f ic ien in for ion in he i de ke t at t o t su m v o m a a de ina ion te t rm
5	Tr f f ic De ity a ns	T he lev l o f t f f ic de ity Ba d ire ly t e ra ns se en on be f v h ic les in he b j 's nt t t nu m r o e p re se su ec l lan d he lan in he b j 's tra t t t ve e a n o r es su ec d ire io f t l, d he b i l ity f t he b j t t t c n o ra ve an a o su ec h ic le dr ive be lan d r t tw ve o m an eu ve r ee n es an lec he dr iv in d. ia b le d z In Va Sp t t se g sp ee r ee on es , i de du d s d l im it be in d ica to to co ns r a re ce p ee an r f t f f ic de ity ( ia b le d l im it f o ra ns e.g a v ar sp ee o ., 3 0m h In ho l d be in d a te ta te te te p on an rs s u rp re s du io in l s ds ). 5 0 % t tra a re c n ve p ee

134

Appendix A (continued)

Variable#Variable name Variable definition Examples and hints 5.1 Level-of-service A1: Free flow, no lead traffic/ Level-of-service A2: Free flow, leading traffic present/Level-of-service B: Flow with some restrictions/ LOS A1 represents ^a free flow traffic situation when the subject vehicle has no leading traffic in any lane (following traffic may or may not be present). Individual users are unaffected by the traffic stream. Freedom to select desired speeds and to maneuver within the traffic stream is at the highest level possible./LOS A2 represents ^a free flow traffic with ^a leading vehicle present in at least one lane. However, individual drivers are still virtually unaffected by the presence of others in the traffic stream. Freedom to select desired speeds and to maneuver within the traffic stream is extremely high. The general level of comfort and convenience provided to the motorist, passenger, or pedestrian is excellent./LOS B is still in the range of stable flow, but the presence of other users in the traffic stream begins to be noticeable. Freedom to select desired speeds is relatively unaffected, but there is ^a slight decline in the freedom to maneuver within the traffic stream from LOS A. The level of comfortand convenience provided is somewhat less than at LOS A, because the presence of others in the traffic stream begins to affect individual behavior. Driving-related decisions are made with virtually no need to consider the presence of other vehicles (due to the lowest traffic density)./Ex. 1: If more than 1 lane is present in the direction of travel, then LOS A2 may apply if there is ^a lead vehicle in the subject's lane but no vehicles in the adjacent lane preventing the driver from passing the lead vehicle./Driving-related decisions are made with ^a small need to consider the presence of other vehicles (due to ^a fairly low traffic density). Ex. 2: If >1 through lane is present, then LOS B may apply if there is ^a lead vehicle as well as an adjacent vehicle preventing the driver from easily passing OR if there are adjacent vehicles on both sides. However, this situation should be transient. The subject driver should not be ''boxed" in for a more than afew seconds. LOS B would also apply if several vehicles are present in the mid-range vicinity, even if they are not directly in front of or adjacent to the subject. Driving speeds are still at or near the speed limit and are not persistently affected by surrounding traffic. 5.2 Level-of-service C: Stable flow, maneuverability and speed are more restricted/Level-of-service D: Unstable flow - temporary restrictions LOS C is still in the range of stable flow, but marks the beginning of the range of flow in which the operation of individual users becomes significantly Driving-related decisions are made with ^a definite need to consider the presence of other vehicles, with ^a good chance of mishap if such

substantially slow driver

affected by interactions with others in the traffic stream. The selection of speed is now affected by the presence of others, and maneuvering within the traffic stream requires substantial vigilance on the part of the driver. The general level of comfort and convenience declines noticeably at this level./ LOS D represents ^a high-density, but stable flow. Speed and freedom to maneuver are severely restricted, and the driver or pedestrian experiences ^a generally poor level of comfort and convenience. Small increases in traffic flow will generally cause operational problems at this level.

considerations are not made (due to ^a medium traffic density).Ex. 1: If only 1 through lane is present, LOS C may apply if subject has ^a lead vehicle AND another car is following the subject. OR, if subject is following multiple vehicles. In either case, the speed is significantly controlled by leading traffic, but the prevailing speed is not more than 10 mph below the speed limit.Ex. 2: If >1 through lane is present, LOS C may apply if the subject is ''boxed in" by lead and adjacent vehicles and this condition is not transient (e.g., it persists as the vehicles travel for some time). LOS C would also apply if multiple vehicles are present in the near-range vicinity, and travel speeds are

5.3 Level-of-service E: Flow is unstable, vehicles are unable to pass, temporary stoppages, etc./Level-ofservice F: Forced traffic flow condition with lowspeeds and traffic volumes that are below capacity

LOS E represents operating conditions at or near the capacity level. All speeds are reduced to ^a low, but relatively uniform value. Freedom to maneuver within the traffic stream is extremely difficult, and it is generally accomplished by forcing ^a vehicle or pedestrian to ''give way" to accommodate such maneuvers. Comfort and convenience levels areextremely poor, and driver or pedestrian frustration is generally high. Operations at this

moderately affected (but are not more than 10 mph below the posted speed limit).Ex. 3: If the subject is preparing to exit, merge, change lanes, etc. in an LOS C environment, there will be multiple vehicles posing potential hazards and requiring careful monitoring by the subject. The maneuver will be more difficult, but will generally be completed without incident./Driving-related decisions are made with urgent need to consider the presence of other vehicles, with ^a great likelihood of mishap if such considerations are not made (due to ^a fairly high traffic density).Ex. 1: If only 1 through lane is present, LOS D may apply if subject is following another car AND another car is following the subject. OR, if subject is following multiple vehicles. In either case, the speed is significantly controlled by leading traffic, and prevailing speed is more than 10 mph below the speed limit.Ex. 2: If >1 through lane is present, LOS D may apply if the subject is persistently ''boxed in" by lead vehicles and adjacent vehicles, AND the prevailing travel speed is determined by surrounding traffic and is more than 10 mph below the posted speed limit.Ex. 3: If the subject is preparing to exit, merge, change lanes, etc. in an LOS D environment, there will be multiple vehicles posing potential hazards and requiring careful monitoring. The maneuver will not be easy and will likely involve braking, accelerating, or excessive steering on the part of the subject or other vehicles.

Driving-related decisions are made with an urgent need to consider the presence of other vehicles, with ^a great likelihood of mishap if such considerations are not made/freedom to execute maneuvers is severely restricted such that drivers must be aggressive in maneuvering (due to ^a very high traffic density).Ex. 1: If only 1 through lane is present, LOS E may apply if subject is following multiple cars AND multiple cars are following the

L. Precht et al. / Transportation Research Part F 51 (2017) 103–144

Variable#Variable name Variable definition Examples and hints level are usually unstable, because small increases in flow or minor perturbations within the traffic stream will cause breakdowns./LOS F represents forced or breakdown flow. This condition existswherever the amount of traffic approaching ^a point exceeds the amount which can traverse the point. Queues form behind such locations. Operations within the queue are characterized by stop-and-go waves, and they are extremely unstable. Vehicles may progress at reasonable speeds for several hundred feet or more, then be required to stop in ^a cyclic fashion. LOS F is used to describe the operating conditions within the queue, as well as the point of the breakdown. It should be noted, however, that in many cases operating conditions of vehicles or pedestrians discharged from the queue may be quite good. Nevertheless, it is the point at which arrival flow exceeds discharge flow, which causes the queue to form, and level-ofservice F is an appropriate designation for such points. subject. The speed is significantly controlled by leading traffic, and the prevailing speed is reduced to less than half the posted speed limit.Ex. 2: If >1 through lane is present, then LOS E may apply if the subject is persistently ''boxed in" by lead vehicles and adjacent vehicles, AND the prevailing travel speed is determined by surrounding traffic and is less than half the posted speed limit.Ex. 3: If the subject is preparing to exit, merge, change lanes, etc. in an LOS E environment, there will be multiple vehicles posing potential hazards and requiring careful monitoring by the subject. The maneuver will be ''forced" and will likely involve braking, accelerating, or excessive steering on the part of both the subject and other vehicles./Traffic flow and related driving decisions are based entirely on the presence and actions of other vehicles (due to the highest traffic density).Ex. 1: Regardless of the number of travel lanes, LOS F represents ''traffic jam" or ''stop and go" conditions. Ex. 2: If the subject is preparing to exit, merge, change lanes, etc. queues will be forming or present either in the subject's desired lane and/or in the subject's destination lane. The maneuver will be ''forced"and will involve braking, accelerating, or excessive steering on the part of both the subject and other vehicles.5.4 Unknown Cannot determine the traffic density due to limitations in video views, lighting, visual obstructions, or limited perspective. Ex. Part of the video is missing or there is insufficient information in the video to make adetermination.6 Lighting Lighting condition. If inside ^a tunnel or parking facility, code the conditions inside the facility, regardless of the lighting conditions outside. 6.1 Dawn The time of day during the Event is sunrise. It is just starting to get light, but is still mostly dark. Most vehicles will drive with headlights. 6.2 Daylight The Event occurs in daylight, such as occurs in after dawn but before dusk.6.3 Dusk The time of day during the Event is sunset. It is just before full darkness. Most vehicles will

drive with headlights.

b le Va ia r #	b le Va ia r na m e	b le de f in Va ia it ion r	les d h Ex int am p an s
6. 4	kn l ht d Da ig r es s, e	da k du he bu he dw It is ing Ev is t t, t t r r en ro a ay f f ic ien ly l ig ht d. ( h ic le be i de Ve t ts su e m ay ou or in i de k ing l. ) str tu tu s a p ar uc re or nn e	ht d dw lu de lam l l a L ig inc tre et e ro a ay s s p s a s w e s l ig ht ing ing fro bu ine i de d t ha t co m m s sse s p ro v he i l lum ina he dw lso t te t y ro a ay a
6. 5	Da kn l ig ht d t r es s, no e	It is da k du ing he Ev d t he dw is t t, t r r en an ro a ay no l ig ht d. ( h ic le be i de in i de k ing Ve ts e m ay ou or s a p ar l. ) str tu tu uc re or nn e
6. 6	Un kn ow n	Ca de ine he l ig ht ing d it ion du ot te t e t nn rm co n s o l im ita ion in i de iew l ig ht ing isu l t s v o v s, , v a bs ion l im ite d ive tru ct t o s, or p er sp ec	Ex Pa f t he i de is iss ing he is rt t o v o m or re in f f ic ien in for ion in he i de ke t at t o t su m v o m a a de ina ion te t rm
7	he W t ea r	he d it io f t he f in i de W t Ev t. I ea r c on n o en s a l o k in fa i l ity de he d it io tu t nn e r p ar g c co co n ns , in i de he fa i l ity d les f t he he t t s c re g ar s o w ea r , d it io i de ts co n ns ou
1 7.	A dv Co d it ion No er se n s	he dv he ic d it ion he T at t t re ar e n o a er se m os p r co n s a im f t he Ev ( d it ion de i be d in he t t ot e o en no co n s sc r r ies ). teg ca or
7. 2 7. 3	Fo g M ist / L ig ht Ra in	he is fog is i b le he im f t he T Ev at t t t. re v e o en T he is ist in he ir l ig ht in he im f t at t t re m a or ra e o he Ev t t. en
7. 4	in ing Ra	is in ing d i ly he im f t he ( de It Ev Co ste at t t t. ra a e o en d in Su fac Co d it ion ) t r we oa r e n	he k for ip C c w er us e.
7. 5	Sn ing ow	It is ing he im f t he Ev ( Co de at t t t. sn ow e o en sn ow lu h d in fac d it ion ) Su Co or s s on ro a r e n
7. 6	lee ing S t	is lee ing he im f t he ( de ice It Ev Co t at t t t. s e o en on d in Su fac Co d it ion ) ro a r e n
7. 7	Ra in & Fo g	It is bo h r in ing d fog he im f t he Ev t at t t t. a an g y e o en ( de d in fac d it ion ) Co Su Co et w ro a r e n
7. 8	/ lee Sn S & Fo t ow g	is bo h ing d fog he im f t he It t at t t sn ow an g y e o Ev ( Co de d in Su fac Co d it ion ) t. en sn ow on ro a r e n
7. 9	he Ot r	he f a dv he T is ic e t at re so m p e o er se m os p r y d it ion de i be d in he nt t ot co n p re se no sc r r , ies he im f t he teg at t t Ev t. ca or e o en ,	b low d, b low ( fa l l Ex ing ing ing t . s m og sa n sn ow no , fro ky ), in d, ha i l, d / du ke st, m s cro ssw sa n sm o
7. 1 0	Un kn ow n	Ca de ine he he he im f t he ot te t t t t t nn rm w ea r a e o du l im ita ion in i de iew l ig ht ing Ev t e t t en o s v o v s, , isu l o bs ion l im ite d ive tru ct t v a s, or p er sp ec	Ex Pa f t he i de is iss ing he is rt t o v o m or re in f f ic ien in for ion in he i de ke t at t o t su m v o m a a de ina ion te t rm

Variable#Variable name Variable definition Examples and hints

Intersection Influence A judgment call as to whether the subject vehicle's safe movement, travel path, and travel speed, are under the influence of an intersection at the time of the event (at any time between Event Start through Event End). This can include the subject or other involved vehicle(s) accelerating or decelerating in relation to an intersection or intersecting trafficway, accelerating or decelerating prior to ^a turn onto ^a new roadway or into ^a parking lot or driveway, waiting in ^a queue of traffic, moving between through lanes and turn lanes or through lanes and acceleration/deceleration lanes, yielding to oncoming or cross traffic, etc. Note that ^a ''Yes" option can be coded here even if Relation to Junction is Non-junction if the vehicle(s) are too far from the intersection to code Relation toJunction categories but are still being influenced in ^a manner described here by an intersection (e.g., ^a longer queue of traffic at ^a signal, or ^a long process of deceleration prior to ^a turn).

8.1 Yes, Traffic Signal Subject vehicle's safe movement along the roadway (e.g., speed, travel path, lane changes) is influenced by ^a traffic signal at the time of the event. This can include accelerating or decelerating in ^a queue of traffic or in response to signal color changes or in preparation for ^a turn at such an intersection, moving between through lanes and turn lanes, yielding to oncoming or cross traffic, etc. Does not include green lights that have no impact on the subject's travel during the event (to be coded as ''No"). Includes parking lot and driveway entrance/exits that are control by ^a traffic signal.

| ia
b
le
Va
r
# | ia
b
le
Va
r
na
m
e | ia
b
le
de
f
in
it
ion
Va
r | les
d
h
int
Ex
am
p
an
s | | |-------------------------------|--------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------------|--| | 8.
2 | Sto
S
ig
Ye
s,
p
n | Su
b
j
h
ic
le
's
fe
lon
he
t v
t a
t
ec
e
sa
m
ov
em
en
g
dw
(
d,
l p
h,
lan
ha
tra
at
ro
a
ay
e.g
sp
ee
ve
e c
ng
.,
in
f
lu
d
by
ig
he
im
f t
he
sto
at
t
t
en
ce
a
p
s
n
e o
h
is
inc
lu
de
ler
ing
de
ler
ing
T
at
at
ca
n
ac
ce
or
ce
f t
f
f
ic
in
ion
for
at
a t
q
ue
ue
o
ra
or
p
re
p
ar
ur
n,
m
be
hr
h
lan
d t
lan
l
d
ie
tw
t
ee
n
ou
g
es
an
ur
n
es
, y
ing
f
f
ic
lu
de
k
In
s t
etc
on
co
m
or
cr
os
ra
c
s p
ar
,
d
dr
ive
its
he
he
b
j
is
he
n t
t
t
an
wa
y
ex
w
su
ec
ex | )
is
es
t.
ev
en
in
a
ing
ov
ing
to
ing
lot
it
ing | | | 8.
3 | l
le
d
Ye
Un
nt
s,
co
ro | h
ic
le
d
is
b
j
ig
t t
sto
ve
an
su
ec
o a
p
s
n.
b
h
le
fe
lon
he
Su
j
ic
's
t v
t a
t
ec
e
sa
m
ov
em
en
g
dw
(
d,
l p
h,
lan
ha
tra
at
ro
a
ay
e.g
sp
ee
ve
e c
ng
.,
in
f
lu
d
by
int
ion
int
ing
ct
ct
en
ce
an
er
se
or
er
se
f
f
ic
he
ha
dr
k
ive
ing
tra
ot
r t
-w
ay
n
a
wa
y
or
p
ar
,
/
it,
ha
is
l
le
d
(
in
he
tra
t
t
t c
tro
t
en
nc
e
ex
no
on
b
j
's
d
ire
ion
f t
l
)
by
f t
t
ct
ty
su
ec
o
rav
e
an
y
p
e o
l
de
ice
T
h
is
inc
lu
de
ler
ing
nt
at
co
ro
v
ca
n
ac
ce
de
ler
f t
f
f
ic
ing
in
in
at
ce
a
q
ue
ue
o
ra
or
p
re
p
ar
for
ing
be
hr
h
lan
a t
tw
n t
ur
n,
m
ov
ee
ou
g
es
an
lan
ie
l
d
ing
ing
f
f
ic
to
s t
es
, y
on
co
m
or
cr
os
ra
,
Ge
l
ly
i
l
l o
ly
be
de
d
i
f c
f
f
ic
tra
ne
ra
w
n
co
ro
ss
or
ing
f
f
ic
is
ing
he
b
j
tra
t
on
co
m
m
ov
ac
ro
ss
su
ec | )
is
es
lot
f
f
ic
ra
or
ion
at
d t
ur
n
etc
's
t | | | 8.
4 | ha
Ye
Int
s,
er
c
ng
e | int
de
d
l p
h
d
ir
ing
tra
at
en
ve
an
re
q
u
so
m
e s
p
ee
ing
fro
he
b
j
ste
t
t.
er
re
sp
on
se
m
su
ec
b
h
le
fe
lon
he
Su
j
ic
's
t v
t a
t
ec
e
sa
m
ov
em
en
g
dw
(
d,
l p
h,
lan
ha
tra
at
ro
a
ay
e.g
sp
ee
ve
e c
ng
.,
in
f
lu
d
by
int
ha
(
inc
lu
d
ing
en
ce
an
er
c
ng
e a
re
a
d
it
)
he
im
f t
he
tra
at
t
t
en
nc
e a
n
ex
ra
m
p
s
e o
h
lu
de
ler
de
ler
T
is
inc
ing
ing
at
at
ca
n
ac
ce
or
ce
f t
f
f
ic
ior
ing
ing
to
q
ue
ue
o
ra
or
p
r
m
erg
m
ov
,
be
hr
h
lan
d
ler
ion
/
tw
t
at
ee
n
ou
g
es
an
ac
ce
de
ler
ion
lan
ie
l
d
ing
f
f
ic
in
d
j
at
to
tra
ce
es
, y
a
ac
lan
lat
d t
ing
by
he
b
j
t
t o
es
re
e
o m
erg
su
ec
r a
h
ic
le,
et
ve
c. | d
or
)
is
es
t.
ev
en
in
a
t
en
he
t
no
r | |

ia b le Va r #	ia b le Va r na m e	ia b le de f in it ion Va r	les d h int Ex am p an s
8. 5	k ing lot ive / it Ye Pa Dr En Ex tra s, r wa y nc e ,	b j h ic le fe lon he Su 's t v t a t ec e sa m ov em en g dw ( d, l p h, lan ha ) is tra at ro a ay e.g sp ee ve e c ng es ., in f lu d by k ing lot dr ive / tra en ce a p ar or wa y en nc e ha l le d ( he b d it t is in j 's ire ion t t c tro t t ct ex no on su ec f t l ) by f t f f ic l de ice ty nt o rav e an y p e o ra co ro v ( lt ho h he inv lve d v h ic les ot t a ug r o e m ay or m ay no ha ). h lu de he b ig T is inc j sto t t o ve a p s n ca n su ec r he inv lve d v h ic le ( ) de ler ing in f f ic in ot at tra r o e s ce ion int k ing lot dr ive at to tu p re p ar rn o a p ar or wa y, ing int f a k ing lot dr ive tu ut rn o o r o o p ar or wa y, l d f f ic ie ing ing ior to s t to y on co m or cr os ra p r ing in it ing k ing lot dr ive tu et rn or ex a p ar or wa y, c. T h is inc lu de dr iv ing k ing in teg ca or y s or p ar
8. 6	he Ye Ot s, r	k ing lot p ar b j h ic le fe lon he Su 's t v t a t ec e sa m ov em en g dw ( d, l p h, lan ha ) is tra at ro a ay e.g sp ee ve e c ng es ., in f lu d by he f int ion t r t ct en ce so m e o y p e o er se or f f ic l d he int ing ist in ct tra t ot er se wa no e r y ies he im f t he teg at t t t. ca or e o ev en	h is inc lu de in f lu lat d l ks T to ca n en ce s r e e cr os sw a , i lro d c ing f f ic irc les / d- bo tra ut et ra a ro ss s, c ro un a s, c.
8. 7	No	Su b j h ic le 's fe lon he t v t a t ec e sa m ov em en g dw ( d, l p h, lan ha ) is tra at ro a ay e.g sp ee ve e c ng es ., in f lu d by f int ion t ty ct no en ce an y p e o er se or int ing f f ic he im f t he ct tra at t t t. er se wa y e o ev en T h is be de d e i f t he b j is in t m ay co ve n su ec or ne ar f f ic de d int ion int ing i ct ct tra an er se or er se wa p ro y v ha he b j d, l p h, 's t t t t tra at et ot su ec sp ee ve c. ar e n be ing ive ly f fec d du ing he t te t t. ac a r ev en
8. 8	Un kn ow n	Ca de ine he he he b j is be ing ot te t r t t nn rm w su ec in f lu d by int ion du l im ita ion in ct e t t en ce an er se o s i de iew l ig ht ing isu l o bs ion tru ct v o v s, , v a s, or l im ite d ive t p er sp ec	Ex Pa f t he i de is iss ing he is rt t o v o m or re in f f ic ien in for ion in he i de ke t at t o t su m v o m a a de ina ion te t rm

Appendix B

Tables B1–B3.

Table B1 Total number and duration (M, SD) of effectively coded secondary tasks (related to two-minute segments).

| Secondary task
Total
Total duration
Total
Total
Total duration
Total duration in
number
in s
number,
number,
in s, SCEs only
s, baselines only
SCEs only
baselines
only
Talking/singing, audience unknown
97 (0.2, 0.5)
1425.95 (2.6, 11.1)
49 (0.2, 0.5)
48 (0.2, 0.6)
824.65 (3.0, 12.0)
601.31 (2.2, 10.1)
Dancing
10 (0.0, 0.2)
27.33 (0.0, 0.4)
3(0.0, 0.1)
7(0.0, 0.2)
9.87 (0.0, 0.4)
17.45 (0.1, 0.5)
Reading
1 (0.0, 0.0)
22.26 (0.0, 0.9)
0
1(0.0, 0.1)
0
22.26 (0.1, 1.3)
Passenger in adjacent seat – talking
541 (1.0, 2.6)
11626.60 (20.9,
283 (1.0, 2.8)
258 (0.9, 2.4)
5968.3 (21.5,
5658.31 (20.4,
without looking at passenger
41.4)
41.0)
41.8)
Passenger in adjacent seat – talking
390 (0.7, 2.3)
784.09 (1.4, 5.4)
191 (0.7, 2.5)
199 (0.7, 2.2)
361.37 (1.3, 5.7)
422.71 (1.5, 5.1)
while looking at passenger
Passenger in adjacent seat – any
2 (0.0, 0.1)
12.17 (0.0, 0.5)
2(0.0, 0.1)
0
12.17 (0.0, 0.7)
0
interaction but talking
Passenger in rear seat – talking without
19 (0.0, 0.5)
208.99 (0.4, 4.4)
13 (0.0, 0.6)
6(0.0, 0.2)
161.22 (0.6, 5.9)
47.77 (0.2, 1.8)
looking at passenger
Passenger in rear seat – talking while
14 (0.0, 0.5)
16.79 (0.0, 0.5)
11 (0.0, 0.7)
3(0.0, 0.1)
10.58 (0.0, 0.6)
6.21 (0.0, 0.3)
looking at passenger
Object dropped by driver
1 (0.0, 0.0)
1.15 (0.0, 0.0)
0
1(0.0, 0.1)
0
1.15 (0.0, 0.1)
Reaching for object, other
54 (0.1, 0.4)
110.80 (0.2, 1.0)
24 (0.1, 0.3)
30 (0.1, 0.4)
57.44 (0.2, 1.1)
53.36 (0.2, 0.9)
Object in vehicle, other
69 (0.1, 0.4)
939.03 (1.7, 7.7)
32 (0.1, 0.4)
37 (0.1, 0.4)
506.82 (1.8, 8.4)
432.22 (1.6, 6.9)
Cell phone, holding
140 (0.3, 0.8)
1844.75 (3.3, 11.2)
69 (0.2, 0.8)
71 (0.3, 0.7)
880.09 (3.2, 11.6)
964.66 (3.5, 10.8)
Cell phone, talking/listening, hand-held
112 (0.2, 0.5)
8854.57 (15.9, 38.2)
55 (0.2, 0.5)
57 (0.2, 0.4)
4086.2 (14.7,
4768.34 (17.2,
36.9)
39.4)
Cell Phone, texting/browsing/dial.,
340 (0.6, 1.1)
6668.98 (12.0, 24.8)
171 (0.6, 1.1)
169 (0.6, 1.1)
3374.5 (12.1,
3294.53 (11.9,
hand held
25.4)
24.1)
Cell Phone, locating/reaching/
192 (0.3, 0.7)
549.17 (1.0, 2.5)
99 (0.4, 0.7)
93 (0.3, 0.7)
257.59 (0.9, 2.2)
291.58 (1.0, 2.8)
answering
Cell phone, other
8 (0.0, 0.1)
53.17 (0.1, 1.0)
2(0.0, 0.1)
6(0.0, 0.2)
18.67 (0.1, 1.0)
34.49 (0.1, 1.0)
Adjusting/monitoring climate control
48 (0.1, 0.3)
148.74 (0.3, 1.5)
27 (0.1, 0.3)
21 (0.1, 0.3)
99.36 (0.4, 1.9)
49.38 (0.2, 0.8)
Adjusting/monitoring radio
238 (0.4, 0.8)
1545.12 (2.8, 11.1)
102 (0.4, 0.7)
136 (0.5, 0.9)
740.21 (2.7, 12.9)
804.90 (2.9, 9.0)
Adjusting/monitoring other devices
126 (0.2, 0.5)
549.70 (1.0, 4.4)
61 (0.2, 0.5)
65 (0.2, 0.5)
347.17 (1.2, 5.9)
202.53 (0.7, 2.0)
integral to vehicle
External distraction
699 (1.3, 1.7)
1716.81 (3.1, 5.9)
355 (1.3, 1.7)
344 (1.2, 1.7)
988.86 (3.6, 7.3)
727.94 (2.6, 4.0)
Reaching for food-related or drink
30 (0.1, 0.3)
70.43 (0.1, 0.8)
13 (0.0, 0.2)
17 (0.1, 0.3)
31.65 (0.1, 0.8)
38.78 (0.1, 0.8)
related item
Eating/drinking
38 (0.1, 0.3)
728.39 (1.3, 8.6)
18 (0.1, 0.3)
20 (0.1, 0.3)
332.36 (1.2, 8.6)
396.03 (1.4, 8.6)
Reaching for cigar/cigarette
1 (0.0, 0.0)
9.94 (0.0, 0.4)
0
1(0.0, 0.1)
0
9.94 (0.0, 0.6)
Lighting/smoking/extinguishing
5 (0.0, 0.1)
331.20 (0.6, 7.5)
2(0.0, 0.1)
3(0.0, 0.1)
132.19 (0.5, 6.6)
199.01 (0.7, 8.2)
cigar/cigarette
Combing/brushing/fixing hair
9 (0.0, 0.1)
111.51 (0.2, 2.5)
8(0.0, 0.2)
1(0.0, 0.1)
107.03 (0.4, 3.5)
4.48 (0.0, 0.3)
Applying make-up
1 (0.0, 0.0)
25.71 (0.0, 1.1)
1(0.0, 0.1)
0
25.71 (0.1, 1.5)
0
Biting nails/cuticles
9 (0.0, 0.2)
99.34 (0.2, 2.1)
5(0.0, 0.2)
4(0.0, 0.1)
68.61 (0.2, 2.8)
30.73 (0.1, 1.1)
Removing/adjusting clothing
13 (0.0, 0.2)
83.38 (0.1, 2.0)
11 (0.0, 0.3)
2(0.0, 0.1)
76.74 (0.3, 2.9)
6.64 (0.0, 0.3)
Removing/adjusting jewelry
6 (0.0, 0.1)
154.77 (0.3, 5.4)
4(0.0, 0.1)
2(0.0, 0.1)
26.18 (0.1, 1.2)
128.58 (0.5, 7.5)
Removing/inserting/adjusting contact
5 (0.0, 0.1)
17.79 (0.0, 0.4)
3(0.0, 0.1)
2(0.0, 0.1)
12.25 (0.0, 0.5)
5.54 (0.0, 0.3)
lenses or glasses
Other personal hygiene
18 (0.0, 0.3)
171.39 (0.3, 2.8)
8(0.0, 0.2)
10 (0.0, 0.3)
59.93 (0.2, 1.4)
111.46 (0.4, 3.7)
Other non-specific internal eye glance
85 (0.2, 0.4)
111.40 (0.2, 0.7)
43 (0.2, 0.5)
42 (0.2, 0.4)
53.77 (0.2, 0.7)
57.64 (0.2, 0.7)
Other known secondary task
27 (0.0, 0.3)
351.36 (0.6, 4.2)
11 (0.0, 0.3)
16 (0.1, 0.3)
94.05 (0.3, 2.9)
257.32 (0.9, 5.2) | | | | | |----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--|--|--|--| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |

Table B2 Total number and duration (M, SD) of coded driving task demands (related to two-minute segments).

Situational/Environmental variables		Total number	Total duration in min	Total number, SCEs only	Total number, baselines only	Total duration in min, SCEs only	Total duration in min, basel. only
Traffic density	Level-of-service A1 to B (smooth traffic)	500 (0.9, 0.3)	1013.41 (1.8, 0.7)	244 (0.9, 0.3)	256 (0.9, 0.3)	477.48 (1.7, 0.8)	535.93 (1.9, 0.6)
	Level-of-serv. C to D (moderate traffic)	51 (0.1, 0.3)	104.82 (0.2, 0.6)	31 (0.1, 0.3)	20 (0.1, 0.3)	62.82 (0.2, 0.6)	42.00 (0.2, 0.5)
	Level-of-service E to F (heavy traffic)	7 (0.0, 0.1)	12.37 (0.0, 0.2)	4(0.0, 0.1)	3(0.0, 0.1)	6.57 (0.0, 0.2)	5.80 (0.0, 0.2)

Table B2 (continued)

	Situational/Environmental variables	Total number	Total duration in min	Total number, SCEs only	Total number, baselines only	Total duration in min, SCEs only	Total duration in min, basel. only
Intersection influence	Traffic signal	510 (0.9, 0.9)	252.63 (0.5, 0.5)	239 (0.9, 0.9)	271 (1.0, 0.9)	124.28 (0.4, 0.5)	128.34 (0.5, 0.5)
	Stop sign	258 (0.5, 0.9)	62.88 (0.1, 0.2)	130 (0.5, 0.9)	128 (0.5, 0.9)	32.74 (0.1, 0.2)	30.14 (0.1, 0.2)
	Uncontrolled	160 (0.3, 0.5)	28.77 (0.1, 0.1)	90 (0.3, 0.5)	70 (0.3, 0.6)	16.74 (0.1, 0.1)	12.03 (0.0, 0.1)
	Interchange	47 (0.1, 0.3)	17.58 (0.0, 0.1)	21 (0.1, 0.3)	26 (0.1, 0.3)	7.54 (0.0, 0.1)	10.05 (0.0, 0.1)
	Parking lot	239 (0.4, 0.6)	175.15 (0.3, 0.5)	138 (0.5, 0.6)	101 (0.4, 0.5)	71.78 (0.3, 0.4)	103.37 (0.4, 0.6)
	Other	63 (0.1, 0.4)	14.27 (0.0, 0.1)	35 (0.1, 0.4)	28 (0.1, 0.4)	7.93 (0.0, 0.1)	6.35 (0.0, 0.1)
	No intersection	1078 (1.9,	591.42 (1.1, 0.6)	508 (1.8, 1.2)	570 (2.1, 1.1)	289.88 (1.0, 0.6)	301.54 (1.1, 0.6)
	influence	1.1)
Lighting	Dawn	8 (0.0, 0.1)	16.74 (0.0, 0.2)	4(0.0, 0.1)	4(0.0, 0.1)	8.34 (0.0, 0.2)	8.40 (0.0, 0.3)
	Daylight	437 (0.8, 0.4)	875.55 (1.6, 0.9)	216 (0.8, 0.4)	221 (0.8, 0.4)	418.80 (1.5, 0.9)	456.75 (1.6, 0.9)
	Dusk	10 (0.0, 0.1)	21.05 (0.0, 0.3)	6(0.0, 0.1)	4(0.0, 0.1)	12.65 (0.0, 0.3)	8.40 (0.0, 0.3)
	Darkness lighted	103 (0.2, 0.4)	198.21 (0.4, 0.8)	49 (0.2, 0.4)	54 (0.2, 0.4)	94.55 (0.3, 0.8)	103.66 (0.4, 0.8)
	Darkness not lighted	12 (0.0, 0.1)	16.93 (0.0, 0.2)	7 (0.0, 0.2)	5 (0.0, 0.1)	12.52 (0.0, 0.3)	4.41 (0.0, 0.1)
Weather	No adverse conditions	503 (0.9, 0.3)	1018.40 (1.8, 0.7)	253 (0.9, 0.3)	250 (0.9, 0.3)	495.50 (1.8, 0.7)	522.90 (1.9, 0.6)
	Fog	1 (0.0, 0.0)	2.04 (0.0, 0.1)	1(0.0, 0.1)	0	2.04 (0.0, 0.1)	0
	Mist/light rain	17 (0.0, 0.2)	33.49 (0.1, 0.3)	10 (0.0, 0.2)	7(0.0, 0.2)	18.79 (0.1, 0.4)	14.70 (0.1, 0.3)
	Raining	34 (0.1, 0.2)	70.73 (0.1, 0.5)	14 (0.1, 0.2)	20 (0.1, 0.3)	28.80 (0.1, 0.5)	41.93 (0.2, 0.5)
	Snowing	1 (0.0, 0.0)	1.74 (0.0, 0.1)	1(0.0, 0.1)	0	1.74 (0.0, 0.1)	0
	Rain and Fog	1 (0.0, 0.0)	2.10 (0.0, 0.1)	0	1(0.0, 0.1)	0	2.10 (0.0, 0.1)

Table B3 Total number and categorization of effectively coded driver behaviors.

Driver behavior	Total number (M, SD) Total number categorized as
			Error Violation Unsafe	unclassified	Other
Improper backing, did not see	1 (0.00, 0.04)	1	0	0	0
Improper backing, other	1 (0.00, 0.04)	0	1	0	0
Right-of-way error in relation to other vehicle or person, apparent decision failure	1 (0.00, 0.04)	0	1	0	0
Signal violation, apparently did not see signal	1 (0.00, 0.04)	1	0	0	0
Parking in improper or dangerous location	1 (0.00, 0.04)	0	1	0	0
Apparent unfamiliarity with vehicle	1 (0.00, 0.04)	1	0	0	0
Other sign violation	1 (0.00, 0.04)	0	1	0	0
Other sign (e.g., Yield) violation, apparently did not see sign	1 (0.00, 0.04)	1	0	0	0
Making turn from wrong lane	1 (0.00, 0.04)	0	1	0	0
Following too closely	2 (0.00, 0.06)	0	2	0	0
Driving slowly in relation to other traffic: not below speed limit	2 (0.00, 0.06)	0	2	0	0
Apparent general inexperience driving	2 (0.00, 0.06)	2	0	0	0
Other improper or unsafe passing	2 (0.00, 0.06)	0	1	1	0
Driving slowly: below speed limit	3 (0.00, 0.07)	0	3	0	0
Signal violation, tried to beat signal change	3 (0.01, 0.07)	0	3	0	0
Sudden or improper braking	4 (0.01, 0.08)	0	4	0	0
Avoiding other vehicle	4 (0.01, 0.08)	0	0	4	0
Illegal passing (but on right)	6 (0.01, 0.10)	0	6	0	0
Right-of-way error in relation to other vehicle or person, apparent recognition failure 6 (0.01, 0.10)		6	0	0	0
Improper turn, other	8 (0.01, 0.11)	4	3	0	1
Stop sign violation, intentionally ran stop sign at speed	9 (0.02, 0.14)	0	9	0	0
Wrong side of road, not overtaking	9 (0.02, 0.15)	0	9	0	0
Signal violation, intentionally disregarded signal	11 (0.02, 0.14)	0	11	0	0
Illegal passing on right	11 (0.02, 0.24)	0	11	0	0
Apparent unfamiliarity with roadway	13 (0.02, 0.16)	12	0	0	1
Non-signed crossing violation	14 (0.02, 0.16)	0	14	0	0
Exceeded safe speed but not speed limit	16 (0.03, 0.16)	0	16	0	0
Other	31 (0.06, 0.26)	0	31	0	0
Improper turn, wide right/left turn	43 (0.08, 0.29)	43	0	0	0
Exceeded speed limit	60 (0.11, 0.41)	0	60	0	0
Improper turn, cut corner on right/left turn	107 (0.19, 0.42)	107	0	0	0
Stop sign violation, rolling stop	154 (0.28, 0.66)	0	154	0	0
Failed to signal/Improper signal	164 (0.29, 0.76)	0	164	0	0
Unintentional Lane Deviation	197 (0.35, 0.85)	197	0	0	0

References

Bachoo, S., Bhagwanjee, A., & Govender, K. (2013). The influence of anger, impulsivity, sensation seeking and driver attitudes on risky driving behaviour among post-graduate university students in Durban, South Africa. Accident Analysis & Prevention, 55, 67–76.
Blatt, A., Pierowicz, J., Flanigan, M., Lin, P.-S., Kourtellis, A., Jovanis, P., ... Hoover, M. (2015). Naturalistic driving study: Field data collection (SHRP 2 Report S2- S07-RW-1). Washington, DC: Transportation Research Board of the National Academies.
Bolker, B. M., Brooks, M. E., Clark, C. J., Geange, S. W., Poulsen, J. R., Stevens, M. H. H., & White, J.-S. S. (2009). Generalized linear mixed models: A practical guide for ecology and evolution. Trends in Ecology & Evolution, 24(3), 127–135.
Bond, C. F., & Titus, L. J. (1983). Social facilitation: A meta-analysis of 241 studies. Psychological Bulletin, 94(2), 265–292.
Campbell, K. (2012). The SHRP 2 naturalistic driving study: Addressing driver performance and behavior in traffic safety. Washington, DC: Transportation Research Board of the National Academies.
Carney, C., McGehee, D., Harland, K., Weiss, M., & Raby, M. (2015). Using naturalistic driving data to assess the prevalence of environmental factors and driver behaviors in teen driver crashes. Washington, DC: AAA Foundation for Traffic Safety.
Carsten, O., & Merat, N. (2015). Protective or not? In 4th International conference on driver distraction & inattention. Sydney, Australia.
Conover, W. J., & Iman, R. L. (1981). Rank transformations as a bridge between parametric and nonparametric statistics. The American Statistician, 35(3), 124–129.
Coxe, S., West, S. G., & Aiken, L. S. (2009). The analysis of count data: A gentle introduction to Poisson regression and its alternatives. Journal of Personality Assessment, 91(2), 121–136.
Dahlen, E. R., & White, R. P. (2006). The Big Five factors, sensation seeking, and driving anger in the prediction of unsafe driving. Personality and Individual Differences, 41(5), 903–915.
De Winter, J., & Dodou, D. (2010). The driver behaviour questionnaire as a predictor of accidents: A meta-analysis. Journal of Safety Research, 41(6), 463–470. Deffenbacher, J. L., Lynch, R. S., Oetting, E. R., & Swaim, R. C. (2002). The driving anger expression inventory: A measure of how people express their anger on the road. Behaviour Research and Therapy, 40(6), 717–737.
Dingus, T. A., Guo, F., Lee, S., Antin, J. F., Perez, M., Buchanan-King, M., & Hankey, J. (2016). Driver crash risk factors and prevalence evaluation using naturalistic driving data. Proceedings of the National Academy of Sciences, 113(10), 2636–2641.
Durbin, D. R., McGehee, D. V., Fisher, D., & McCartt, A. (2014). Special considerations in distracted driving with teens. Annals of Advances in Automotive Medicine, 58, 69–83.
Fitch, G. M., Hanowski, R., Burnett, G., & Crundall, D. (2013). Development of a protocol to classify drivers' emotional conversation. In 3rd International conference on driver distraction and inattention. Gothenburg, Sweden.
Fournier, D. A., Skaug, H. J., Ancheta, J., Ianelli, J., Magnusson, A., Maunder, MN., ... Sibert, J. (2012). AD model builder: Using automatic differentiation for statistical inference of highly parameterized complex nonlinear models. Optimization Methods and Software, 27(2), 233–249.
Guo, F., Klauer, S. G., Fang, Y., Hankey, J. M., Antin, J. F., Perez, M. A., ... Dingus, T. A. (2016). The effects of age on crash risk associated with driver distraction. International Journal of Epidemiology, 1–8.
Hickman, J. S., Hanowski, R. J., & Bocanegra, J. (2010). Distraction in commercial trucks and buses: Assessing prevalence and risk in conjunction with crashes and near-crashes (Reprt No. FMCSA-RRR-10-049). Washington, DC: Federal Motor Carrier Safety Administration.
Horrey, W. J., & Wickens, C. D. (2006). Examining the impact of cell phone conversations on driving using meta-analytic techniques. Human Factors: The Journal of the Human Factors and Ergonomics Society, 48(1), 196–205.
Hubbard, A. E., Ahern, J., Fleischer, N. L., van der Laan, M., Lippman, S. A., Jewell, N., ... Satariano, WA. (2010). To GEE or not to GEE. Epidemiology, 21(4), 467–474.
IBM Corp. Released (2011). IBM SPSS statistics for windows, Version 20.0. NY: IBM Corp Armonk.
Jellentrup, N., Metz, B., & Rothe, S. (2011). Can talking on the phone keep the driver awake? Results of a field-study using telephoning as a countermeasure against fatigue while driving. In Proceedings of the 2nd international conference on driver distraction and inattention. Gothenburg, Sweden.
Johansson, Ö., Wanvik, P. O., & Elvik, R. (2009). A new method for assessing the risk of accident associated with darkness. Accident Analysis & Prevention, 41 (4), 809–815.
Klauer, S. G., Dingus, T. A., Neale, V. L., Sudweeks, J. D., & Ramsey, D. J. (2006). The impact of driver inattention on near-crash/crash risk: An analysis using the 100-car naturalistic driving study data (Report No. DOT HS 810 594). Washington, DC: National Highway Traffic Safety Administration.
Klauer, S. G., Guo, F., Simons-Morton, B. G., Ouimet, M. C., Lee, S. E., & Dingus, T. A. (2014). Distracted driving and risk of road crashes among novice and experienced drivers. New England Journal of Medicine, 370(1), 54–59.
Olson, R. L., Hanowski, R. J., Hickman, J. S., & Bocanegra, J. L. (2009). Driver distraction in commercial vehicle operations (Report No. FMCSA-RRR-09-042). Washington, DC: Federal Motor Carrier Safety Administration.
Precht, L., Keinath, A., & Krems, J. F. (2017). Identifying crash risk factors of crashes associated with driver distraction, impairment, or neither using the crash trifecta concept and naturalistic driving data. Manuscript submitted for publication.
Precht, L., Keinath, A., & Krems, J. F. (2017a). Effects of driving anger on driver behavior Results from naturalistic driving data. Transportation Research Part F: Traffic Psychology and Behaviour, 45, 75–92.
Precht, L., Keinath, A., & Krems, J. F. (2017b). Identifying the main factors contributing to driving errors and traffic violations Results from naturalistic driving data. Transportation Research Part F: Traffic Psychology and Behaviour, 49, 49–92.
Qiu, L., & Nixon, W. (2008). Effects of adverse weather on traffic crashes: Systematic review and meta-analysis. Transportation Research Record: Journal of the Transportation Research Board, 2055, 139–146.
R Core Team (2016). R: A language and environment for statistical computing. Vienna: Austria. Retrieved from <https://www.R-project.org/>.
Reason, J. (1990). Human error. Cambridge: Cambridge University Press.
Reason, J., Manstead, A., Stradling, S., Baxter, J., & Campbell, K. (1990). Errors and violations on the roads: A real distinction? Ergonomics, 33(10–11),
Rosenbloom, T., & Perlman, A. (2016). Tendency to commit traffic violations and presence of passengers in the car. Transportation Research Part F: Traffic Psychology and Behaviour, 39, 10–18.
Ross, V., Jongen, E. M. M., Brijs, K., Brijs, T., & Wets, G. (2016). Investigating risky, distracting, and protective peer passenger effects in a dual process framework. Accident; Analysis and Prevention, 93, 217–225.
Rowe, R., Roman, G. D., McKenna, F. P., Barker, E., & Poulter, D. (2015). Measuring errors and violations on the road: A bifactor modeling approach to the driver behavior questionnaire. Accident Analysis & Prevention, 74, 118–125.
Simmons, S. M., Hicks, A., & Caird, J. K. (2016). Safety-critical event risk associated with cell phone tasks as measured in naturalistic driving studies: A systematic review and meta-analysis. Accident Analysis & Prevention, 87, 161–169.
Singh, S. (2015). Critical reasons for crashes investigated in the national motor vehicle crash causation survey (Traffic Safety Facts CrashStats. Report No. DOT HS 812 115). Washington, DC: National Highway Traffic Safety Administration.
Skaug, H., Fournier, D., Bolker, B., Magnusson, A., & Nielsen, A. (2016). Generalized linear mixed models using AD model builder. R package version 0.8.3.3. Staubach, M. (2009). Factors correlated with traffic accidents as a basis for evaluating advanced driver assistance systems. Accident Analysis & Prevention, 41
Strayer, D. L., Cooper, J. M., Turrill, J., Coleman, J., Medeiros-Ward, N., & Biondi, F. (2013). Measuring cognitive distraction in the automobile. Washington, DC: AAA Foundation for Traffic Safety.
Strayer, D. L., & Turrill (2014). Measuring cognitive distraction in the automobile II: Assessing in-vehicle voice-based interactive technologies. Washington, DC: AAA Foundation for Traffic Safety.

Summala, H., Rajalin, S., & Radun, I. (2014). Risky driving and recorded driving offences: A 24-year follow-up study. Accident Analysis & Prevention, 73, 27–33. Tian, R., Li, L., Chen, M., Chen, Y., & Witt, G. J. (2013). Studying the effects of driver distraction and traffic density on the probability of crash and near-crash events in naturalistic driving environment. IEEE Transactions on Intelligent Transportation Systems, 14(3), 1547–1555.

Transportation Research Board of the National Academies of Science (2013). The 2nd strategic highway research program naturalistic driving study dataset. Available from the SHRP2 NDS InSight Data Dissemination web site: <https://insight.shrp2nds.us>.

Victor, T., Dozza, M., Bärgman, J., Boda, C.-N., Engström, J., Flannagan, C., ... Markkula, G. (2015). Analysis of naturalistic driving study data: Safer glances, driver inattention, and crash risk (No. SHRP 2 Report S2–S08A-RW-1). Washington, DC: Transportation Research Board of the National Academies.

Vollrath, M., Meilinger, T., & Krüger, H.-P. (2002). How the presence of passengers influences the risk of a collision with another vehicle. Accident Analysis & Prevention, 34(5), 649–654.

World Health Organization (2017). Save LIVES - A road safety technical package. Geneva. Licence: CC BY-NC-SA 3.0 IGO.

Young, K. L., & Regan, M. A. (2013). Defining the relationship between behavioural adaptation and driver distraction. In C. M. Rudin-Brown & S. L. Jamson (Eds.), Behavioural Adaptation and Road Safety: Theory, Evidence and Action. Boca Raton, Florida: CRC Press.

Young, K. L., Salmon, P. M., & Lenne, M. G. (2013). At the cross-roads: An on-road examination of driving errors at intersections. Accident Analysis & Prevention, 58, 226–234.

Zhang, T., Chan, A. H. S., & Zhang, W. (2015). Dimensions of driving anger and their relationships with aberrant driving. Accident Analysis & Prevention, 81, 124–133.

Glossary

DBQ: Driver Behaviour Questionnaire GLMM: Generalized Linear Mixed Model LMM: Linear Mixed Model

NDS: Naturalistic driving study SCE: Safety critical event

SHRP 2: Second Strategic Highway Research Program

VTTI: Virginia Tech Transportation Institute

prechtidentifyingeffectsdrivingsecondary2017.md 621 KB 文件歷史 原始文件