Using psychology to explain young drivers traffic accidents

The over representation of young drivers in road traffic accidents is a serious social problem. In Britain, road crashes are the single biggest killer of 17 – 24 year olds even though just one in eight licence holders fall within that age bracket (Break, 2006). Additionally, 17 year old drivers have 50% more crashes per year than do 25 year olds, who in turn have 35% more crashes per year than do 50 year olds (Maycock, Lockwood & Lester 1991). Young driver overrepresentation in traffic accidents is also a world wide problem, with young drivers accounting for 27% of all driver fatalities across all OECD countries whilst only representing 10% of the population (JTRC, 2006). It is therefore important to understand why younger drivers are overly involved in traffic accidents and psychological research can help address this issue by assessing the age related deficits (in both skill and social factors such as attitudes) that are associated with accident causation (Elander, West, & French, 1993). In deed, many studies have examined age-related deficits in driving skill and their association with traffic accident involvement or proxies for accident risk (aberrant driving behaviours). This research therefore provides insight into why young drivers are overrepresented in traffic accidents. This essay will critically examine the research in the area of traffic psychology. The following potentially explanatory factors for young driver accident overrepresentation will be considered, hazard perception, awareness of risk, passenger influence, personality and attitudinal social cognition variables.

Hazard perception

Higher order perceptual and cognitive skills are necessary to safely interact on the road with an awareness of ones surroundings (Deery, 1999), these skills evolve from experience. When young novice drivers detect hazards they do so less quickly and do not analyse the whole traffic environment as efficiently as older drivers (Deery, 1999). High crash involvement was hypothesised by Pradhan, Hammel, DeRamus, Pollatsek, Noyce & Fisher (2005) to be largely due to young driver’s inability to acquire and asses’ information in hazardous situations. 24, 16 – 17 year olds, 19 – 29 year olds and 60 – 75 year olds were recruited. A driving simulator was used with 16 risky driving situations, eye movements were recorded with a head mounted eye tracker. The situations involved exposure to warning signs, traffic control devices, pedestrians, and other vehicles that could cause potential hazards. Significant age related differences were found in driver eye scanning behaviour. The 16-17 age group spotted on average 35.1% of the potential risk shown on the simulator where as the 19 -29 age group spotted 50.3% and 66.2% of hazards was spotted by the 60 – 70 year olds. The results indicate that younger drivers have poorer hazard perception skills than older drivers; this contributes to crash involvement as these skills gained by experience enable drivers to safely interact on the road with an awareness of ones surroundings (Deery, 1999). Pradham et al’s (2005) study included only those above 60 in the older driver category. It is noted in research that drivers above 60, although they have vast experience, their hazard perception ability is likely to decline due to decreases in cognitive and visual function (Horswill, Marrington, McCullough, Wood, Pachana, McWilliam & Raikos, 2008).

Laboratory based studies of hazard perception involve recording eye movement but the pattern of eye movement in the simulated setting may not reflect the movements that would occur in the real world. Inside and outside the motor vehicle there are distractions that individuals focus their attention on such as passengers and sights which draw drivers’ attention away from potential hazards. If such real world distractions were taken into account the results produced may show that hazards are detected to a lesser extent in the real driving environment than in laboratory studies. Opposite to this the simulator may produce lower hazard detection scores as the amount of cognitive resources used in the study may decrease (ie, less concentration) as there are no serious consequences of driver error (Young, Regan and Hammer, 2003). Test track studies (Lee, Klauer, Olsen, Simsons-Morton, Dingus, Ramsey and Ouimet (2009) have assessed age-related differences in hazard perception skills outside of the laboratory approximating real driving conditions (Young et al, 2003). The test track in this instance was a two way highway with hidden hazards such as stop signs, pedestrians and lane closures. This branch of research provides evidence that older drivers showed significantly better hazard detection observing more hazards and showing clear acknowledgment of the existence of potential hazards understanding risky situations that could be fatal. Thus despite the limitations of laboratory based simulator research, one can still be confident of the validity of the results as they apply to a real driving track setting.

While the research reviewed above implies that young drivers have poorer hazard perception skills than older drivers, contradictory findings are provided by Horswill et al (2008). These researchers developed a video based hazard perception test involving what they describe as ‘genuine traffic conflict’. The conflict involved situations where traffic accidents occurred or near accidents were viewed with their vehicle and pedestrians, cyclists or stationary vehicles causing hazards. Horswill et al (2008) compared hazard perception of 16 young drivers with a mean age of 19 and 17 older drivers with a mean age of 41.94. Participants viewed a video sequence with 33 potential hazards and were instructed to press a response button as quickly as possible when they perceived a hazard. There was no significant difference in the number of hazards detected by both age groups. However, the mean hazard perception reaction time of older drivers was significantly faster than the younger drivers with a large effect size. The risk of traffic accident may be due to young drivers’ reaction times to hazards in the traffic environment. Studies have reported a lack of concentration in young drivers which affects reaction times to hazards, this has been reported in situations on the road which are familiar or seem simple in which they do not proceed with caution or high concentration as risk is deemed to be low (Makishita and Matsunaga, 2008). The null findings of young driver’s hazard perception in comparison to older drivers may lie in the method of this study. Participants pressed a response button in reaction to a hazard, in this situation participants may have paid more attention to the video thus hazards more readily as they were not involved in vehicle handling alike the participants in simulator or track studies where participants operated as he/she would a vehicle on the open road.

Overall, many lab based studies show younger drivers have poorer hazard perception skills than do older drivers, studies outside the lab also conform this. There is good reason to explain the null finding of Horswill et al’s (2008) study which under different circumstances may have produced a significant difference in the hazard perception of younger and older drivers.

Awareness of risk

Awareness of risk produces another possible explanation for why young drivers have a higher accident rate than older drivers. If drivers are not aware of risk they do not proceed through the driving environment with caution. When risk is perceived drivers make an assessment of their ability to act in situation (Groeger & Brown, 1989), the manor in which drivers perceive and take account of their abilities, assessing their own skill, impact on the type of behaviour they undertake. Young drivers are said to be miscalibrated, believing their ‘road craft’ skills are better developed than they actually are (Deery, 1999). Lower perceived level of risk and miscalibrated confidence creates unsafe driving practices. Finn & Brag (1986) carried out an investigation with 18- 24 year old and 30 – 50 year old males in their estimations of accidents. Participants carried out a questionnaire about accident involvement, in the second condition they were exposed to 10 still photographs of different driving situations and were asked to indicate how much risk would be involved for them, drivers of their own age and their peers. The third condition assessed the same factors yet in a video tape sequence of 15 scenarios. The younger and older participants both estimated young drivers as more likely to be involved in an accident. However, young drivers estimated their individual chances of being in an accident significantly lower than there age group. In no situation did the younger drivers see themselves as at more risk of an accident than their peer group or older drivers. The younger group reported enhanced confidence in estimations of their own skill having less risk of an accident in situations of tailgating, driving at night, driving in snow covered roads and driving after consuming six beers in one hour. In situations requiring skill young drivers regard themselves as highly able due to miscalibration. In unexpected circumstances (i.e. when pedestrian suddenly appear) they view the situation as dangerous. Older drivers through experience and learned exposure know that unforeseen circumstances occur and have the knowledge and skill to deal with the unexpected situation (Finn and Bragg 1986).

Self report studies produce the possibility of bias in accounts of driving ability as participants, especially those of a younger age, may wish to come across in a highly skilful light as it is documented that the personal identity associated with driving portrays self-confidence and power (Stradling & Meadows, 2006). Females were excluded from the above study limiting the sample of the driving population. The sample population provides a further constraint in that participants were recruited from a newspaper advertisement. This recruitment procedure can lead to a more homogenous sample, participants of similar personality types may volunteer suggesting results may distorted as personality types can impact driver behaviour as will be reviewed in section 3. However, other self report studies conducting similar research with different participant samples provide the same results. Mathews and Moran’s (1986) young driver participants were undergraduate students and older drivers were of the universities faculties and staff. The sample is representative of a particular population but one which is different from Finn and Bragg’s (1986) yet both provide the same results.

The findings suggest that awareness of risk and self perceived driving ability is interrelated. Young drivers have a lower threshold to perceive risk. In driving situations when a hazard is detected feelings of risk are generated, young drivers produce lower expectations of the harmful outcomes of risky situations and miscalibrate their skill creating an over- involvement of young drivers in traffic accidents.

Passenger Influence

Passenger effects provide a potential explanation of why young drivers are overrepresented in traffic accidents. For young drivers it has been found that passenger influence has an effect on the likelihood of traffic crash involvement. Preusser, Ferguson and Williams (1998) assessed official Australian traffic accident fatality data and found that younger drivers are more likely than older to be accompanied by one or more passenger at the time of fatal accidents. Young drivers were also shown to be more often involved in ‘at fault’ accidents when accompanied by passengers than when travelling alone. On the other hand the presence of passengers did not affect the extent to which drivers in their mid twenties and above were involved in ‘at fault’ accidents. This study provides clear evidence that younger drivers are more at risk when travelling with other passengers. Although this study was conducted in Australia research in Great Britain also shows that young drivers are particularly susceptible to the negative consequences of carrying passengers.

Baxter, Manstead, Stradling, Reason, Campbell and Parker (1990) provide observational research on the influence of passengers on young driver’s risky driving practices in Great Britain. They define risky driving as changing lanes without indicating, driving above the speed limit, driving on the crown of the road unnecessarily and close following. 244 cars of solo drivers or drivers with one passenger were followed for an average of 4 minutes. The sex of the driver and passenger was noted along with age estimations by the researchers as young (below 30) or old (above 30). There was a significant main effect found for speeding and passenger type, young male drivers were observed to speed more often when accompanied by another young male passenger (M=2.44). This effect was also found for young female passengers but with a smaller effect size (M=1.83). For young drivers changing lanes without indicating was related to the presence of young male passengers. It is clear from this observation that young passengers impact young driver’s behaviour. However, the 4 minute observation period is not enough to make general conclusions of passenger effects. Prevailing traffic conditions may have made the driver act in a way in which was out of his/her control in reaction to other road users. Also, the drivers may not have had the opportunity to commit each of the aberrant behaviours of ‘risky driving’ during the time period. Despite this, the study extends Preusser et al’s (1998) findings in showing that the type of passenger that travels with young drivers is important in determining unsafe behaviour.

Self reported questionnaires and focus group discussions further back up the observational evidence. Mitropoulos and Regan (2002) carried out a telephone survey with 872 participants of 16 to 55 and above. The questions determined the effect of the passengers on driver’s age and passenger’s age. Focus groups discussions were held with 28 participants discussing influences on young driver behaviour. Passengers were found to influence driver behaviour implicitly through their physical presence. Peers made young drivers feel they had to adopt risky driving to portray a certain image or attitude (Stradling and Meadows, 2006) and young male’s reported they would engage in risky driving to show off even in situation where they had not been asked. Passengers also determined risky driving by explicitly suggesting an antisocial behaviour to take part in. However, there is no explicit definition of ‘risky driving’; it would be of value to investigate further the impact drivers have on different risky practices adopted.

Speeding causes traffic accidents, a factor which research suggests is highly affected by passenger presence as observation has shown, this is also clear in Arnett, Offer and Fine’s (1997) self report study. Participants included 64 students of 17 – 18 who kept a log of their driving over 10 days including passenger details and if they exceeded speed limits. There was a significant effect found for the type of passenger and speeding. Post hoc shelf tests revealed that young drivers reported speeding significantly more when accompanied in the car by a friend (M=1.52), than by a parent (M=1.18). The log book data is self report and may incur bias in that individuals do not wish to reveal unsafe driving habits. However, anonymity of responses was insured with no consequence for reporting unlawful driving. As the study was conducted with only young drivers of 17 -18 the results do not act as comparative with the rest of the young driver population (up to 24) or the older driver population. Young drivers were reported to speed only slightly less when alone (M=1.41) than with peers. Further research is needed to establish why individuals feel the need to speed even in situations where there is no influence from passengers. The answer to this may lie in the Social Cognition theories (Scott-Parker et al, 2009) to be discussed in section 5. Despite limitations the study provides evidence that young drivers adopt more risky driving practices when accompanied by passengers of their own age group. On the basis of the evidence reviewed from a number of different types of study (analysis of accident database, observation, and self-report data) passenger effects aid our understanding of why young drivers are involved in more traffic accidents.

The effects of passengers on young driver behaviour can be further explained by Ackers Social Learning Theory (SLT) (Fleiter, Lennon and Watson, 2009). The fact that young drivers are more crash prone than older drivers may be due to adolescence being a period characterised by developmental changes of a physiological, cognitive, behavioural and a social nature. During the development from childhood to adolescents to young adults, individuals undertake cognitive and physical changes alike their peers. It is at that point in which individuals begin to view their peers as more important or influential to them than their parents/guardian. Peers generally have the same attitudes and morals about behaviours (Fleiter et al, 2009) and forming friendships offers acceptance and independence from the family unit (Cook and Dayley, 1997). With a reliance on peers in helping form attitudes and behaviours, peer pressure is evident. SLT takes account of how youths are predicted to be more likely to indulge in antisocial behaviour when in the presence of peers who are accepting of or promote behaviours through encouragement (Akers, Krohn, Lanza-Kaduce & Radosevich, 1979). Behaviour for example speeding is strengthened by positive reinforcement in rewards such as peer approval and avoidance of punishment in the form of negative reinforcement such as peer disproval.

SLT main concept is rewards and punishment that enhance conformity to behaviour which young people may know is wrong and would be disapproved of by parents or the against the law. Scott-Parker, Watson and King (2009) conducted research with 161, 17 – 24 years old drivers (70% students). The study used a questionnaire involving self reported risky driving behaviour, attitudes and the social learning experiences of young drivers. Accidents and offences were also recorded. Items such as, “My parents wouldn’t like the way I drive with friends in the car” assessed conformity of young drivers with peers. Also, anticipated rewards questions such as, “My mates and I talk for ages about the really cool risky things I do in the car”, and punishments, “My mates make fun of me when I don’t show off”. SLT accounted for 42% of variance in risky driving behaviour. Anticipated rewards (β=.23) was the strongest predictor of young drivers adopting risky driving, anticipated punishment had a lesser effect yet still significant (β =-.20). This study is reliable in that it is based in sound theory exploring the psychological influences of peer influence on young drivers yet as noted above the majority of participants were university students likely to be of a higher socioeconomic class thus not representative of the general population.*** A point to consider is that the participant pool holds a larger proportion of 17 – 18 year olds (58) than those of 19-20 (47), 21-22(36) and 23-24 (20). There is a possibility that this may skew results as research shows those at the lower end of the ‘younger driver’ group are more accident prone than those higher in the 20 age group (Preusser, 1998). However, this self report study is further validated by measuring reported crashes and offences and making comparative evaluations.

Young drivers are consciously affected by the presence of passengers who produce positive or negative rewards in relation to driving behaviour. Reinforcement is generally document in producing negative driving behaviours contributing to young drivers disproportionate representation in traffic accidents but this can also be applied to safe driving. Older passengers may disprove of risky driving and provide a salient norm to the driver or explicitly express their attitudes, for example they may gave a young driver into trouble for speeding (Baxter et al, 1990). In contrast, young passengers’ salient norm is that risky driving is ‘good’ and they express the view that driving at high speeds and taking risks is acceptable (Harrington and McBride, 1970). Fleiter, Lennon & Watson (2009) carried out focus group discussions with 67 males and females of 17 – 77 years of age. To guide discussion a semi-structured interview format of open ended questions was used, questions were designed to cover components of the SLT. Questions covered the effects of the presence or absence of others in the car and punishments and reinforces associated with speeding. The presence of passengers led all drivers to drive more slowly including friends, children, and parent passengers. Taken these results into consideration passengers act as a defence for risky driving not an additive to the overrepresentation of traffic accidents. However, these results may be misleading as in discussions ‘young men’ consistently showed a willingness to increase driving speeds with friends in the car. Due to the small number of young males (7) compared to young females (18) the results were deemed non significant. In the eventuality that the number of males and females were equal a significant negative passenger effect may have been produced showing a significant relationship with passengers and increased speeding for young male drivers with passengers present.

SLT helps explain why young drivers adopt deviant driving practices when in the presence of passengers. Another psychological theory that helps explain this is Tajfel and Turner’s (1979) Social Identity theory (SIT) (Scott-Parker et al, 2009). The SIT considers how social identity evolves from group membership. Young people aim to achieve and maintain a positive social identity boosting ‘self esteem’ by acting in ways favourable to the in group of individual’s choice (Brown, 2000). SIT is alike the SLT in that peer attitudes shape and influence behaviour as social identity is influenced from situations which results in group membership or exclusion. Young people thus young drivers act in ways which is believed to be desirable by their peers (Brown, 200), modifying behaviour to fit with desirable social groups so to escape social disproval. In short, the difference between the theories is that in the SIT the young driver does not look to reward or punishment but it is a sense of belongingness to a certain peer group which is particularly important, the gains perceived is in identity to the in-group and acceptance.

Scott-Parker et al (2009) who’s study has been reviewed considering the SLT also asses the SIT as an explanation of the passenger effects and the importance and influence of ‘group identity’. The administered questionnaire looked to how young people do things they do not want to so to fit in with their social group. A strong positive correlation was found between risky driving and group identity providing support in the SIT as a theoretic explanation of why young drivers produce risky driving in the presence of passengers. As described previous, Scott-Parker et al (2009) use a limited participant pool yet Hatfeild and Fernandes (2009) support the theory comparing the younger and older driver population using individuals from varied economic status showing that the SIT is a valid theory to explain why young drivers are more involved in traffic accidents. 89 younger drivers (16 – 24) and 110 older (35+) drivers were compared in terms of risky behaviour under social influence from self report questionnaire results. Participants were recruited outside motor registries from locations of different socioeconomic status. Younger drivers reported stronger motives for speeding and drink driving in relation to social influence; this was shown to be highly significant for young males.

Studies of passenger influence on driver behaviour are based on a wide range of data. Self report data, observations and official traffic statistics. The majority of reported results explain that young drivers are influenced by peers who induce risky driving practices causing an overrepresentation in traffic accidents.

Personality

Certain personality characteristics cause individuals to be more spontaneous or more prone to adopt risky behaviour. Connecting personality traits with traffic psychology may provide a further explanation as to why young drivers are involved in a disproportionate number of traffic accidents. Older drivers tend to associate driving with efficient and economical transport. For younger drivers it is a transition to adulthood giving young people autonomy, control and high status (Scott-Parker et al, 2009). Young drivers are said to believe that driving gives them the chance to express themselves in a way they want others to perceive them (Stradling and Meadows, 2006), it is a way of displaying their personality to society. Psychologists in the early days took the belief that traffic accidents were caused by a minority of individuals who possess certain personality characteristics, this stemmed from Famer and Chambers (1939) theory of Accident Proneness (Ulleberg and Rundmo, 2003). This view has been discredited yet personality traits are still associated with accident involvement, more so in predicting risk taking behaviours of individuals (Elander et al 1993 in Ullberg & Rundmo, 2003) than their actual involvement in accidents. Nonetheless, the two factors are correlated as performing risky driving practices evokes danger and the more likely an accident is to occur.

It is important to establish the personality traits that effect driving before assessing if young drivers are high in such traits causing their overrepresentation in traffic accidents. Sensation seeking is a personality trait which differs on individual degrees depending on the amount one desires novel and intense situations and experiences (Dahlen, Martin, Ragan & Kuhlman, 2005). Sensation seekers are individuals willing to take risk, both physical and social which often compromise the law (Srivastava, 2005). Aggression is a trait associated with risk taking as by acting in fast paced risky situations individuals are able to expel feelings of anger and frustration. Individuals high in normlessness take the belief that socially unapproved behaviours are acceptable and required when trying to achieve certain goals, for example speeding to get to a destination on time (Ulleberg and Rundmo, 2003). Ulleberg & Rundmo (2003) used a questionnaire survey with 1932 participants from 16-23 in which sensation seeking, aggression and normlessness significantly correlated with risk perception, attitudes towards traffic safety and risky driving behaviour. Those high in sensation seeking and normlessness perceived risk of accidents occurring as low, had negative attitudes towards traffic safety and reported engaging in risky driving practices. Aggressive individuals reported an understanding of high risk situations yet undertook more risky behaviour portraying a negative attitude towards traffic safety. Personality essentially influences driving behaviour.

It is important to asses the relationship of personality traits and young drivers, Arnett et al (1997) reported how sensation seeking and aggression are both significantly related to speeding, drink driving, racing and passing in a no-passing zone. 17 – 18 year olds and 41 – 59 year olds carried out questionnaires concerning driving behaviour. Sensation seeking was measured by the Arnett Inventory of Sensation Seeking and aggression by the Californian Psychological Inventory. The younger age group correlated significantly higher on both on both personality traits which were significantly related to the reckless driving variables. It is evident that personality traits such as sensation seeking and aggression create stronger tendencies for individual to participate in reckless driving. Young people are shown to be higher in these traits providing an explanation for why they are they are involved in more traffic accidents. A criticism is in that personality traits are considered to be stable across time. If individuals are high in a trait at a young age then this would carry into adulthood. In effect if you adopted risky driving as a young driver due to personality characteristics then you would do so as an older driver. Nevertheless, factors such as social leaning are naturally believed to influence such traits and when maturing the need for high risk and excitement gaining situations decrease (Bandura, 1977 in Ulleberg & Rundmo, 2003) as with reckless driving.

Sensation seeking is a factor in the overrepresentation of young drivers in traffic accidents as the trait is characterised by poor impulse control associated with risky driving where individuals stand a greater change of being involved in an accident due to their impulsivity (Loo ,1978, in Clement & Jonah, 1984). Clement and Jonah (1984) reported no relationship with sensation seeking and accident involvement of 285 young drivers yet it significantly correlated with speeding for all young drivers and not wearing a seat belt for young female drivers. Speeding as noted previous is a factor in a vast majority of traffic accident and not wearing a seatbelt produced more fatalities in accidents. Sensation seeking has been studied widely in traffic psychology, for example in the Hatfield & Fernandes (2009) study described in section 3 risky driving is said to stem from personality. Younger drivers reported stronger motives for drink driving in relation to sensation seeking. Sensation seeking was found to be a motive for young female drivers and speeding but not for males. However, young male results were near significance at p=0.55 suggesting the effect may be significant with a larger sample size.

Speeding as suggested above did not reach significance for young males. Conceptualising sensation seeking created the opportunity to describe it in the driving context rather than referring to it in a more broad term of personality (Scott-Parker et al, 2009) where effects of the trait may be more relaible. Rimmo and Aberg (1999) carried out a questionnaire with 700 Swedish drivers, 18 – 27 years of age. The participants were randomly recruited from the official register of Swedish car owners providing a participant pool reflective of the population. Factors of sensation seeking, ‘thrill and adventure seeking’ were found to be most associated with speeding. Scott -Parker et al’s (2009) study as described in sections 1 and 2 also defined sensation seeking tendencies to ‘thrill seeking’ resulting in a strong, positive correlation with risky driving.

All the reviewed personality studies use questionnaires to obtain results. Drivers own reports of behaviour, especially those such as drink driving opens up the possibility of bias. Reports maybe intentionally false so to give what participants believe are socially desirable answers concealing risky behaviours (Elander et al 1993). Nevertheless, self report in this instance holds the advantage that risky driving (ie, drink driving and speeding) is not always caught and documented by the authorities. Despite some limitations personality helps explain why young drivers are disproportionately involved in more traffic accidents, especially considering the trait sensation seeking.

Social Cognition

Social cognition variables for example attitudes towards committing risky driving, might usefully explain young drivers overrepresentation in traffic crashes. It is young drivers’ propensity to have positive attitudes towards the commission of risky driving behaviours that explains why they perform those behaviours, subsequently increases risk of a road traffic accident. There are a number of models in social cognition that aim to provide accounts of how attitudes and other socio-cognitive variables influence behaviour. The Theory of Planed behaviour (TPB) is one of these models.