برداشت از حواس پرتی راننده در میان رانندگان نوجوان

Abstract Teenage drivers have been shown to have some of the highest crash risks. Crash data provide some insights on factors related to crash likelihood, but rarely capture all issues that can arise from driver distraction. The goal of this study was to assess teenage drivers’ opinions and perceptions of driver distraction. A survey of 1893 Iowa teenagers was conducted to determine and compare the frequency of engagement in distracting activities while driving to their opinions of what they actually consider to be distractions. A cluster analysis was conducted based on their indicated engagement in distracting activities with three groups emerging and classified as INFREQUENT, MODERATE, and FREQUENT engagers. Across all cluster groups, the majority (over 80%) indicated that they considered text messaging to be a distracting task. However, those clustered as FREQUENT engagers still reported a high level of texting while driving even though they considered this task to be distracting. A binary logistic regression model (adjusted for miles driven and license type) showed that FREQUENT and MODERATE engagers were more likely to be involved in a crash when compared to INFREQUENT engagers. The study demonstrates that not all teenagers place themselves at risk. There are subgroups of teenage drivers that often engage in activities they know are distracting, potentially putting themselves in danger. However, this is not the case for all teenage drivers and it is important to target interventions appropriately as well as foster a culture of safety both in schools and at home

1. Introduction Studies show that teenage drivers are overrepresented in vehicular crashes (Beck et al., 2003, Ferguson, 2003, Hartos et al., 2001 and Lee et al., 2011) and have a higher propensity to engage in many distracting tasks (Chisholm, Caird, & Lockhart, 2008). Examination of US crash data shows that distraction has been documented as a contributing factor in approximately 25% of all crashes (Stutts, Reinfurt, Staplin, & Rogeman, 2001) and is therefore, of great interest for policy makers, safety engineers, and researchers. Ranney, Mazzae, Garrott, and Goodman (2000) characterize driver distraction as any activity that takes a driver’s attention away from the task of driving. The effects of performing dual tasks were observed early on by Brown, Tickner, and Simmonds (1969) who found that response time and task accuracy were negatively impacted when concurrently driving and speaking on the telephone. Teenagers are of primary interest because studies have shown that they have fewer visual scans toward the driving area and have more hard braking occurrences when cognitively distracted (Harbluk & Noy, 2002). Teenagers also indicate a greater willingness to engage in all types of multitasking activities (Lerner, 2005). Teenage drivers tend to be overly confident in their driving ability, as well as underestimate the danger of specific driving situations when compared to older drivers (Finn and Bragg, 1986 and Matthews and Moran, 1986). Teenagers tend to have a greater propensity to use technology in their vehicle with many indicating that they have texted while driving (Madden & Lenhart, 2009). However, some risky behavior appears modifiable with appropriate feedback and training. In a study by Ginsburg et al. (2008), teens with authoritative or involved parents were less likely to use their cell phones while driving. McGehee, Raby, Carney, Lee, and Reyes (2007) and Simons-Morton, Hartos, Leaf, and Beck (2002) also demonstrated the importance of parental involvement for young drivers. One method of examining the impact of driver distraction is through crash data (Neyens and Boyle, 2007, Neyens and Boyle, 2008 and Sheridan, 2004). This method captures police-reported crashes but provides limited information on driver-distraction related events because it is often difficult for the officer to pinpoint specific behaviors. Further, most drivers will not volunteer incriminating information related to distractions unless there is evidence that the distraction occurred (Neyens and Boyle, 2007 and Wallace, 2003). Hence, the number of distraction-related crashes is typically underreported (Wallace, 2003). Another major limitation is the consistency in reporting across different states in the US. To address this issue, several states are working with police and other official reporting organizations to standardize the reporting of crucial data such as cell phone use (Violanti, 1997) but standardization across jurisdictions has not been formalized to date. Cellular phone-related driver distractions are extensively examined given their potential impacts toward driver safety. As of 2000, 90% of drivers who own cell phones reported using it while driving (Goodman, Tijerina, Bents, & Wierwille, 2000) and the US Census Bureau estimated a 300% increase in cellular subscribers from 34 Million in 1995 to 159 Million in 2003 (Bergman, 2004). Data from the World Bank showed that in 2004, 63% of the US population had mobile cellular subscriptions, and by 2009, the percentage rose to 89% (World Bank, 2010). Distraction studies on in-vehicle devices have largely been examined in controlled environments to minimize the potential on-road risks to drivers as they traverse through safety–critical situations (Donmez et al., 2007 and Greenberg et al., 2003). For example, driving simulator studies showed that engagement in phone conversations resulted in slower reaction times, a twofold increase in rear-end collisions (Strayer & Drews, 2004), and a higher workload and response time to traffic signals (Drews & Strayer, 2009). Hancock, Lesch, and Simmons (2003) showed greater failures in identifying traffic signals when engaged in a cell phone task. The growth in naturalistic studies, where drivers are observed (using instrumented vehicle with video data) in their natural driving environment, has also provided other insights on driver distractions. For example, Stutts et al. (2005) was able to observe that distractions was a common component of everyday driving and that distractions were frequently associated with decreased driving performance. The advancement in technology has enabled mobile devices to be used for more than phone conversations. These smart devices now include the capability to send and receive emails and text messages, access the Internet, and play music. There are also distractions unrelated to technology that can be detrimental for teenage drivers such as talking to passengers, eating, and drinking (Regan, Lee, & Young, 2009). Simons-Morton et al. (2011) further examined the effects of passengers and showed that risk propensity was lower with adult passengers, but higher with risky friends. Studies in controlled settings (e.g., simulators, on-road closed courses) allow researchers to observe changes in driver performance based on specific distracting tasks. As noted earlier, cell phones and text messages are key areas of examination. However, there are other secondary tasks that cannot be examined in a controlled setting but are important to capture because of safety implications. Surveys can provide a means of capturing self-identified behavior (Mann, Vingilis, Leigh, Anglin, & Blefgen, 1986) that may not be observed otherwise in crash data, controlled settings, or even in naturalistic studies. Surveys also provide insights into stated preferences that may not necessarily match actual performance and can therefore, reveal subpopulations of drivers that differ based on their attitudes toward different distracting activities. The goal of the study was to examine driver distraction among teenagers including what tasks they consider to be distracting as compared to their level of engagement in these same distracting tasks. To address this goal, a survey targeted toward teenage drivers was designed, distributed and analyzed as part of the current study.

Results 4.1. Descriptive statistics In the state of Iowa, teenagers who are 14 years old or older may obtain a school instruction permit that allows them to drive to and from school on a direct route. Intermediate licenses may be granted at age 16. Full licenses are granted to teenagers that have held their intermediate license for 1 year without obtaining any moving violations. As shown in Table 1, most students surveyed held some type of license (approximately 94%, n = 1730). The 106 respondents that did not have a driver’s license were not included in the cluster and regression analysis on driving. Table 1. Distribution by license type. Gender License type Count Age Mean SD Males Instruction permit 168 15.91 0.86 Intermediate license 307 16.80 0.67 Full license 252 18.01 0.56 None 44 16.83 0.98 Females Instruction permit 210 15.76 0.72 Intermediate license 391 16.81 0.63 Full license 202 17.92 0.54 None 62 16.25 0.98 Total 1636a a There were a total of 1836 valid surveys; of which 52 did not indicate their gender, and 148 did not indicate license type. Table options 4.2. Cluster groups The cluster analysis revealed three distinct clusters that were labeled based on their engagement in distracting tasks: FREQUENT (or habitual) engagers, MODERATE engagers, and INFREQUENT engagers. A total of 1603 drivers were placed into clusters; 127 of the 1730 eligible respondents were excluded from cluster membership because they were missing one or more responses within the 13 questions that were included in the clustering. The INFREQUENT engagers (n = 790, 49%) consisted of drivers who reported engaging in fewer distracting activities while driving when compared to the other two clusters. For example, INFREQUENT engagers would tune the radio (median = 4) and adjust climate controls (median = 4) while driving, but not as often as the FREQUENT (median: 7 and 6) and MODERATE engagers (median: 6 and 6) (see Table 2). This group very rarely text messaged (median = 1), dialed (median = 2), or talked (median = 2) on the phone while driving ( Fig. 1). Table 2. Distraction engagement by cluster group. Activity Cluster Response (%) 1 2 3 4 5 6 7 Never Sometimes Frequently Doing homework Infrequent 94.3 3.7 1.1 0.8 0.0 0.0 0.1 Frequent 72.0 9.0 4.0 2.5 3.4 2.2 6.8 Moderate 82.5 10.4 3.7 1.4 1.6 0.2 0.2 Looking for items in wallet/purse/backpack Infrequent 43.9 24.3 18.6 9.5 2.4 0.8 0.5 Moderate 7.9 12.6 25.5 25.1 14.7 11.4 2.9 Frequent 5.3 4.3 10.2 18.6 20.2 22.0 19.3 Reading Infrequent 75.9 16.8 5.2 1.5 0.5 0.0 0.0 Moderate 54.0 21.2 13.0 7.7 3.1 0.8 0.2 Frequent 46.3 15.2 12.4 12.7 6.2 3.1 4.0 Daydreaming Infrequent 39.5 25.9 18.5 9.6 4.1 0.9 1.5 Moderate 4.5 7.5 13.0 17.7 22.0 22.2 13.0 Frequent 12.4 8.4 9.3 14.0 17.1 16.8 22.0 Using a device brought in vehicle Infrequent 52.8 14.7 11.4 9.9 4.3 3.5 3.4 Moderate 27.9 19.6 9.0 12.8 9.6 11.2 10.0 Frequent 8.7 6.2 6.5 13.4 14.6 20.5 30.1 Changing CDs/cassettes Infrequent 34.6 17.5 14.7 15.1 8.6 6.8 2.8 Moderate 18.5 8.1 10.8 14.9 14.9 18.9 13.8 Frequent 3.4 1.2 3.4 11.2 13.0 24.5 43.2 Eating/drinking Infrequent 17.1 19.2 19.1 24.6 11.9 6.1 2.0 Moderate 2.9 8.1 13.2 23.2 22.8 21.2 8.6 Frequent 0.3 0.6 3.7 11.8 14.9 31.4 37.3 Thinking of a complex problem Infrequent 20.6 20.0 19.9 21.6 10.1 6.1 1.6 Moderate 1.0 3.9 7.5 15.7 22.6 28.9 20.4 Frequent 5.6 4.3 5.9 15.5 14.9 23.6 30.1 Tuning radio Infrequent 13.8 9.5 12.8 19.5 14.1 16.3 14.1 Moderate 3.1 4.3 5.3 10.0 13.8 24.6 38.9 Frequent 0.3 0.9 1.2 7.1 11.5 27.3 51.6 Changing heat/AC Infrequent 10.3 10.9 16.3 25.6 16.2 15.2 5.6 Moderate 1.2 3.5 7.3 11.4 18.5 32.8 25.3 Frequent 0.9 1.6 5.0 9.0 14.6 26.7 42.2 Note 1: Highest proportion of response for each cluster is shown in bold. Table options Cell phone use by cluster groups. Fig. 1. Cell phone use by cluster groups. Figure options MODERATE engagers (n = 491, 31%) performed more distracting activities (i.e., engaged in more non-driving activities) while driving when compared to the INFREQUENT engagers. This includes more cell phone use ( Fig. 1), daydreaming, eating or drinking, tuning the radio, and adjusting the climate controls ( Table 2). The FREQUENT engagers (n = 322, 20%) performed all non-driving activities more often when compared to both the MODERATE and INFREQUENT engagers. In addition, FREQUENT engagers also regularly engaged in dialing a cell phone (median = 6), text messaging (median = 6), changing CDs or cassettes (median = 6), using a device (iPod, laptop, etc.) brought into the vehicle (median = 6), looking for items in their wallet/purse/backpack (median = 5), and thinking about something difficult (complex problem, relationship, argument, etc.; median = 6). INFREQUENT engagers had the lowest percentage of crashes (27%) and FREQUENT engagers had the highest (43%). The difference among clusters and number of crashes was significant at α = 0.05 (χ2(2, 1543) = 27287, p < 0.0001). As shown in Table 3, FREQUENT engagers reported the highest proportion of friends as passengers (over 80%) when compared to the other two clusters. INFREQUENT engagers reported the highest proportion of adults as passengers when compared to FREQUENT and MODERATE engagers (over 37%). Table 3. Descriptive statistics by cluster group. Variable Cluster Total Infrequent Moderate Frequent Sample size n (% of total sample) 790 (49.3%) 491 (30.6%) 322 (20.1%) 1603 (100%) Crashes n (% of total sample) 214 (27.1%) 175 (35.6%) 138 (42.9%) 527 (30.3%) Age Mean (SD) 15.3 (0.85) 17.5 (0.55) 16.4 (0.64) 16.4 (1.05) Gender (n and% of cluster) Male 423 (53.5%) 182 (37.1%) 135 (41.9%) 740 (46.2%) Female 343 (43.4%) 296 (60.3%) 177 (55.0%) 816 (50.9%) NAa 24 (3%) 13 (2.6%) 10 (3.1%) 47 (2.9%) Miles driven yesterday Mean (SD) 13.6 (18.6) 20.0 (21.0) 7.1 (30.7) 18.3 (22.9) Passengers (n and % of cluster) Siblings age 6 or younger 32 (4.1%) 26 (5.3%) 20 (6.2%) 78 (4.9%) Siblings age 7 or older 297 (37.6%) 215 (43.8%) 130 (40.4%) 642 (40.0%) Adults (over age 21) 295 (37.3%) 128 (26.1%) 62 (19.3%) 485 (30.3%) Friends 407 (51.5%) 347 (50.3%) 258 (80.1%) 1012 (56.9%) If drive friends, mean number of friends (SD) 1.8 (1.0) 2.0 (0.8) 2.2 (1.1) 2.0 (0.97) a Some drivers did not indicate whether they were male/female. Table options 4.3. Engagement in distracting activities Overall, 85% of survey participants considered text messaging distracting. Within the FREQUENT engagers, over 80% reported that texting was distracting (Table 4), but most still chose to engage in this activity while driving (Fig. 1). INFREQUENT engagers rarely engaged in texting while driving and most (89%) considered this activity to be distracting. Table 4. Distractions while driving: perceptions from the teenage cluster groups. Do you think the following is a distraction while driving? Cluster group (%) Non-drivers Infrequent Moderate Frequent Texting 88.8 87.9 80.6 79.6 Doing homework 82.5 81.9 74.3 68.9 Looking for items in wallet/purse/backpack 80.6 77.0 72.7 68.9 Reading 80.0 76.7 73.3 68.9 Dialing a cellphone 78.1 70.8 66.0 70.9 Talking on cellphone 70.8 61.1 59.4 71.8 Daydreaming 68.8 56.4 56.8 60.2 Using a device brought in vehicle 68.1 60.3 54.9 55.3 Changing CDs/cassettes 57.0 52.9 56.5 44.7 Eating/drinking 56.5 48.4 51.7 46.6 Thinking of a complex problem 53.5 46.1 44.1 43.7 Tuning radio 43.4 37.7 44.4 37.9 Changing heat/AC 35.7 30.2 37.5 30.1 Note: The top three ranked for each group are shown in bold. Table options The next three activities that teenage drivers considered distracting (doing homework, looking for an item in wallet/purse/backpack, and reading) were rarely performed while driving by any cluster group. Many survey respondents considered talking (64.4%) and dialing (72.2%) to be distracting while driving but there were differences across the three clusters (Talking: χ2(2) = 200.05, p < 0.0001, Dialing: χ2(2) = 211.08, p < 0.0001) and this can be observed in Fig. 1. All cluster groups reported more frequent engagement in cell phone activities than in any other activities. Among the cell phone activities, conversing on a cell phone while driving was highest (mean = 3.8 out of 7 [median = 4]) followed by dialing (mean = 3.5, median = 3) and texting (mean = 3.2, median = 3). Other activities examined include being distracted while looking for an item in a purse/wallet/backpack (mean = 3.2) and using another electronic device brought into the vehicle (mean = 3.2). Slightly over half of the drivers surveyed consider changing CDs/cassettes, eating and drinking to be distracting activities and approximately the same portion indicated that they have often engaged in these activities. Driving while thinking about a complex problem, tuning the radio, and adjusting the climate controls were ranked least distracting (48.5%, 41.1% and 33.8% respectively) and all drivers tend to engage in them frequently regardless of the cluster group. 4.4. Driving patterns across clusters The clusters were developed based on the level of engagement in various distracting activities, but other patterns also emerged when examined across other survey responses. For example, the FREQUENT engagers were also more likely to engage in distracting activities at night, with passengers, and while driving to/from work and school (median scores ranging from 4 to 6) (see Table 5). INFREQUENT engagers rarely engaged in any of the distracting activities with median scores ranging from 2 to 3. Table 5. Scenarios where drivers have performed distracting activities while driving (median scores). Driving scenario Cluster group Infrequent Moderate Frequent At night 3 4 6 While driving to/from school 3 5 6 While driving to/from work 2 4 6 In rain, snow, or other bad weather 2 3 4 On highways 2 4 5 With passengers in your vehicle 3 5 6 Responses ranged from 1: Never to 4: Sometimes to 7: Frequently. Table options The FREQUENT engagers was the only group to indicate that they had, on occasion, nearly hit the car directly in front of them (median = 3). With respect to how often the drivers observe their parents/guardians doing any distracting activities, the median response were 5 for FREQUENT and MODERATE engager cluster groups, and 4 for the INFREQUENT engagers. 4.5. Predicting crash likelihood A binary logistic regression model was used to examine the likelihood of teenage drivers being involved in a crash. Cluster group and license type were included as explanatory variables and provide some indication for risk taking while driving and level of driving experience. More frequent drivers may also be more frequent engagers in distracting activities. Hence, the miles that the teenager drove the week prior to the survey being completed was included in the model to account for this potential confounder. This variable was not significant (p > 0.05) and had an odds ratio of 1.0. The model (Table 6) showed that those teenagers that were most likely to engage in distracting activities (FREQUENT engagers) were also 1.45 times more likely to be involved in a crash than the INFREQUENT engager cluster group. As expected, drivers with Full and Intermediate Licensure (and who have been driving for longer periods) were more likely to be involved in crashes than those with Permits only (OR = 2.16 and 1.45 for Full and Intermediate, respectively). Teenagers that typically had friends, as their passengers were 1.37 times more likely to be involved in crashes than those teenagers that did not typically have friends in their car. Table 6. Likelihood of crash involvement. Variable Estimate Std. error Odds ratio 95% CIa z Value Pr(>|z|) (Intercept) −1.914 0.168 −11.38 <0.0001 Frequent engager 0.374 0.163 1.450 (1.06, 2.00) 2.29 0.0218 Moderate engager 0.254 0.141 1.290 (0.98, 1.70) 1.80 0.0724 Intermediate license 0.747 0.194 2.110 (1.45, 3.11) 3.86 0.0001 Full license 1.160 0.202 3.190 (2.16, 4.77) 5.73 <0.0001 Miles driven (past week) 0.001 0.001 1.000 (1.00, 1.00) 1.83 0.0672 Usually have friends in car 0.318 0.144 1.370 (1.04, 1.82) 2.21 0.0272 Log-likelihood at zero 1673.0 Log-likelihood at convergence 1578.7 Number of observationsb 1316 a CI: Confidence Interval. b Information on miles driven was missing for 287 drivers. Table options 4.6. Comparison to non-drivers Consistent with others studies (Ginsburg et al., 2008, McGehee et al., 2007 and Simons-Morton et al., 2002), this study centered on teenage drivers and not other age groups. However, as noted earlier, information was collected from teenagers that did not drive as well. There were 106 teenagers who indicated that they did not possess a license of any type. These teenagers had a mean age of 16.47 years (SD = 1.03) with 44 (or 41.5%) being male and 61 (57.5%) being female (one respondent did not identify the gender). A comparison between teenagers who do and do not drive was conducted to identify whether differences exist between these groups in what was considered to be distractions while driving (Table 4). Consistent with the three cluster groups of drivers, texting was considered distracting by most teenagers in the non-driving group and changing climate controls (AC and heat) was considered least distracting. One interesting observation was that talking on a cell phone was considered more distracting by more non-drivers when compared to the percentages reported by all three cluster groups of drivers.