به حداقل رساندن خطرات و حواس پرتی برای رانندگان جوان و مسافران خود: ارزیابی از یک برنامه آموزش راننده مسافر

Abstract The purpose of the present study was to evaluate a pilot program designed to teach communication skills to young drivers and passengers. Sixty-two young males recruited as 31 pairs of friends, all aged between 18 and 21 years and holding a probationary drivers licence, were randomly assigned to a training or no-training condition. A training program was developed based upon elements of existing team training programs. Driver and passenger pairs operated a driving simulator through scenarios designed to measure aspects of safe driving behaviour and hazard response. Communications between driver and passenger were also measured. All participants were administered the Driver Behaviour Questionnaire before and approximately 2 months after simulator testing. Compared to the untrained group the trained participants exhibited a larger following distance, reduced speed significantly when faced with an unexpected hazard on the road, and exhibited more safe communications. Although current passenger restrictions are warranted, the present results reveal an alternative view of adolescent passengers: rather than being a negative influence on drivers, adolescent passengers can potentially be trained to become a positive influence.

1. Introduction Young drivers are over-represented in road crash statistics, and road crashes are the major cause of death for people aged 15–24 in OECD countries (Organisation for Economic Co-operation and Development, 2006). One important risk factor for young drivers is the presence of passengers, although evidence indicates that the presence of certain types of passengers, in particular, older passengers and female passengers, enhances driving behaviour (Engstrom et al., 2008, Geyer and Ragland, 2005, Rice et al., 2003 and Vollrath et al., 2002). There is considerable evidence however that young drivers are more likely to crash when passengers are present (e.g., Chen et al., 2000, Doherty et al., 1998, Geyer and Ragland, 2005 and Preusser et al., 1998). For drivers under 25 years in Australia, carrying two or more peer age passengers is associated with an increased risk of crash, whereas no such association is observed for drivers who were 25 years and older (Lam, Norton, Woodward, Connor, & Ameratunga, 2003). 1.1. Passenger influences on young drivers In the literature noted above the influence of young passengers in distracting young drivers is often acknowledged. Potential distractions may be visual (when drivers shift their eyes from the road to passengers), auditory (when passengers adjust the radio), or cognitive (when drivers and passengers converse). Compounding this vulnerability to distractions are the findings that, compared to experienced drivers, young drivers utilise less efficient hazard detection strategies and processing abilities, exhibit an impaired capacity to attend to passengers and other distractions, as well as demonstrated limited awareness of when to moderate passenger interaction in line with the demands and context of the driving situation (Crundall et al., 2005, Deery, 1999 and Underwood, 2007). Another major factor contributing to the negative effect of passenger carriage on young drivers is risk taking behaviour (Williams, Ferguson, & McCartt, 2007). Young drivers, particularly male drivers, have been observed to take greater driving risks, such as speeding and close following, when accompanied by young male passengers; in contrast, the presence of female and adult passengers has been observed to curb the risky behaviour of young drivers (McKenna et al., 1998 and Simons-Morton et al., 2005). Risk taking behaviour has also been linked to higher offence and accident rates in young male drivers compared to their female counterparts (Laapotti, Keskinen, Hatakka, & Katila, 2001). Young male drivers report engaging in riskier behaviour when driving with young male passengers compared to females or adults (Regan and Mitsopoulos, 2001, Ulleberg, 2004 and Williams et al., 2007), and young male passengers report that they may explicitly encourage risk taking in their male friends (Regan & Mitsopoulos, 2001). 1.2. Managing the risks associated with passengers Several approaches to the management of these passenger risks are discussed in the literature (see Williams et al. (2007) for a review). Many countries, including the US, New Zealand, and Australia, have introduced legal restrictions on the number and age of passengers that a young driver is allowed to carry. In the state of Victoria in Australia, for example, drivers who are within 1 year of having obtained their probationary licence will be permitted to carry, at most, one passenger aged 16–21 years (Arrive Alive Action Plan, 2008). Such restrictions can be complemented with education and training programs. For example, passenger and peer influences were targeted in a traffic safety campaign in Norway, intended to generate positive traffic safety attitudes and awareness of risky driving practices. The campaign included a specific strategy termed ‘Speak Out!’ designed to empower young passengers to declare unsafe driving and to refuse lifts with unsafe drivers. This educational component of the campaign was accompanied by an increase in police enforcement, which targeted speeding and drunk driving in young drivers. Elvik (2000) found evidence of a reduction in passenger injuries; however, because the campaign consisted of both education and police enforcement, the resulting improvements in safety cannot be attributed to education alone. More recently in the US, a 10-week training program, involving interactive lessons, small-group discussions, and role plays, entitled ‘You Hold the Keys’, was implemented among teens in high school to increase safe driver and passenger behaviours (King, Vidourek, Love, Wegley, & Alles-White, 2008). One of the unique features of the program was the student exposure to the experiences of crash victims who had been victimised by unsafe drivers. In the immediate follow-up, students reported being more likely to engage in safe behaviours, such as wearing seatbelts, and were more likely to avoid drink driving and travelling with another drunk driver. Many of the reported improvements were maintained even 6 months after the first evaluation (King et al., 2008). This study extends existing research by providing a training program to promote safe behaviours between driver–passenger pairs, thereby providing a means to support moderation of behaviour, and by assessing the impact of this initiative through both self-reported behaviours and behavioural observation in a driving simulator. The content of the program encompassed elements of previous programs that encouraged teenage passengers to voice their disagreements with drivers when they were driving dangerously (Elvik, 2000 and King et al., 2008). Young male drivers and passengers were chosen given that previous research confirms the negative influence of young male passengers on young drivers, the reluctance of young males to inform the driver when they feel unsafe (Ulleberg, 2004), and higher crash risk. It was hypothesised that the training program, delivered to young male pairs, would improve driver–passenger communications, reduce risky driving (car following and speed), and would enhance hazard detection and response.

3. Results Post-treatment questionnaires were returned by 23 of the trained participants and 22 of the untrained participants. There were no significant differences in the pre-treatment DBQ scores for those participants who returned post-treatment questionnaires and those who did not. All data presented are for the complete dataset. 3.1. Driver Behaviour Questionnaire (DBQ) The specific interactions of interest within the 4-way ANOVA were not significant. Whether participants received training did not influence the DBQ scores pre and post simulator testing (F(1, 41) = 1.04, p > 0.05), and DBQ scores were not influenced by the participants’ role as a driver or passenger (F(1, 41) < 1). As illustrated in Table 3, the mean scores were higher across the board for the trained participants, confirmed by a main effect of Condition (F(1, 41) = 13.86, p < 0.01). Table 3. Mean scores on the DBQ by item type (SD in parentheses). Item type Pre-test DBQ Post-test DBQ Training Control Training Control Lapses 1.56 (0.41) 1.08 (0.38) 1.55 (0.24) 1.32 (0.44) Mistakes 1.46 (0.35) 0.81 (0.15) 1.40 (0.24) 1.07 (0.35) Violations 2.35 (0.91) 1.45 (0.76) 2.12 (0.64) 1.66 (0.52) Table options There was some improvement in self-reported behaviour for the trained participants. For violations, a higher proportion of the trained participants (n = 13) showed improvements, compared to the untrained participants (n = 22; χ2(1) = 4.04, p < 0.05). A similar result was found for mistakes: a higher proportion of the trained participants (n = 11) showed improvements, compared to the untrained participants (n = 4; χ2(1) = 4.64, p < 0.05). For lapses, however, there were no significant differences in the proportion of trained (n = 8) and untrained (n = 7) participants who showed improvements. Acknowledging that the scores were higher for the trained group, these data are presented to highlight that some benefit post-training behaviour was observed for the trained participants. 3.2. Driving performance Mean driving performance measures for the car following task are presented in Table 4. The mean safe headway adopted by drivers in the trained group was significantly larger than for the control group. Measures of mean and maximum speed were virtually identical for both trained and untrained groups. Measures of vehicle control (SD Speed, SD Headway, SD Lateral Placement) did not differ between groups as expected. Table 4. Measures of driving performance and statistical tests for the car following task (SD in parentheses). Performance measure Condition Statistic Training Control Mean speed (km/h) 56.1 (5.67) 57.4 (5.72) t(29) = −0.66 SD speed (km/h) 24.0 (5.50) 23.8 (3.52) t(29) = 0.13 Maximum speed (km/h) 81.1 (4.56) 81.2 (4.42) t(29) = −0.10 Mean headway (m) 112.9 (36.83) 87.3 (23.25) t(29) = 2.29 ⁎ SD headway (m) 17.7 (8.93) 13.3 (5.56) t(29) = 1.62 SD lateral placement (m) 0.361 (0.104) 0.345 (0.091) t(29) = 0.46 ⁎ p < 0.05. Table options In stage 2 of the drive the participants’ response to the presence of an on-call emergency vehicle was assessed. As illustrated in Fig. 1, participant pairs who had received the training reduced speed more rapidly upon detection of the oncoming emergency vehicle. This was confirmed by analysis with training condition as the between-groups variable, and distance to intersection as a within-groups variable. Speed was analysed across 10-m distance profiles from 170 m prior to the point at which the emergency vehicle crossed the path of own cab (resulting in 18 profile segments). There was a significant Condition by Distance interaction (F(17, 476) = 3.01, p < 0.001), and post hoc t-tests confirmed that speed in the trained group was significant lower than for the controls immediately prior to the emergency vehicle turning across (at distance 20 m, t(28) = 2.16, p < 0.05). Mean speed (±SD) approaching the intersection for the hazard detection event. Fig. 1. Mean speed (±SD) approaching the intersection for the hazard detection event. Figure options 3.3. Communication The communications from seven pairs of passengers were not available for analysis due to technical malfunction. Valid data were derived from 12 trained pairs and 12 untrained pairs. First, the audio transcripts were coded in terms of the percentage of time that pairs spent conversing in the vehicle, for both scenarios. On average, the trained group (M = 49%) spent less time conversing than did the untrained group (M = 69%, t(25) = 2.22, p < 0.05). The violations that were either encouraged or discouraged included speeding, running red light, and deliberate braking. The distractions that were commented on included actual and potential hazards, such as pedestrians and other road users. Given two scenarios (car following and emergency vehicle) and two potential behaviours (violation and distraction), there were up to four opportunities to code a passenger as articulating safe comments. Each passenger received a score between zero and four for their safe comments; for example, they scored a full four points only if they discouraged both violation and distraction behaviour in both of the scenarios. Similarly, there were up to four opportunities to code a passenger as expressing unsafe comments; thus, each passenger received a score between zero and four for their unsafe comments. Each group comprised 12 passengers and thus could achieve in total, up to a score of 48 for safe comments and 48 for unsafe comments. The trained passengers scored 33 for safe comments and 12 for unsafe comments, whereas the untrained passengers scored 24 safe comments and 30 for unsafe comments. The differences between the two groups for both safe and unsafe comments were compared by two-sample proportion tests (χ2). Compared to the untrained passengers, the trained passengers emitted significantly fewer unsafe comments (χ2(1) = 14.09, p < 0.01), however, there was no significant difference for safe comments (χ2(1) = 3.52, p > 0.05).