Respondents: 2500 motorists were randomly sampled. Of this number 504 responded with a completed questionnaire. A further 118 responded via the E-questionnaire. The tables below indicate the demographic details of the respondents.

Table i. Sex of Respondents by Age

Table ii. Sex of Respondents by Insurance Group of main Vehicle Driven

Table iii. Age of Respondents by Insurance Group of main Vehicle Driven

Table iv. Sex of Respondents by Type of Accident Reported.

1.     One female had been involved in two head-on accidents.

2.     Two males had been involved in two rear-end accidents, and a further two had been involved in three rear-end accidents.

3.     One male had been involved in two intersection accidents.

4.     One female had been involved in two intersection accidents.

5.     Three males had been involved in two single-vehicle accidents.

6.     One female had been involved in three accidents involving either a pedestrian and/or cyclist.

Table v. Sex of Respondents by Number of Reported Accidents.

Table vi. Age of Respondents by Type of Accident Reported (respondents may have been involved in more than one accident).

Single Accidents: There are two routes to being involved in one accident within three years. In addition to the primary endogenous variable (one accident), three other endogenous predictors were also found with the use of the above linear regression technique. The first (exposure) was dependent upon another exogenous factor, cognitive anxiety. Age, the third exogenous factor, was a direct link dependent on no other factors. In turn, cognitive anxiety was predicted by seven exogenous factors, which included somatic anxiety, sex, cognitive failure, risky driving behavior, vehicle maintenance, urgency and sensation seeking.

Evaluating the Model:      

* With respective correlation coefficients

  Examining the values for each of these paths, it can be seen that the strongest path to being involved in one accident within three years is the direct path from age.

  Multiple Accidents: There are three separate routes to having more than one accident within the space of three years.

The first of these, safety risk estimation (i.e. shirking personal safety and vehicle legality) is predicted by the type of vehicle used, risky driving behavior (reckless behavior, e.g. speeding, racing against other vehicles), and somatic anxiety.

Moving across to the other predictor, we see high sensation seeking being implicated in the equation. In turn, we also see high sensation seeking being predicted by sex, age, and insurance group of vehicle. In addition, we also see that sensation seeking is predicted by hostility (a TABP trait), cognitive anxiety, and risky driving behavior.

Finally, the third exogenous variable that predicts multiple accidents is vehicle maintenance. In turn, this was predicted by type of vehicle and sex.

It is also observable that there is a relationship between safety risk estimation and vehicle maintenance.  

Evaluating the Model:      

* With respective correlation coefficients

Examining the values for each path, it can be seen that the strongest path to being involved in more than one accident within three years is the indirect path from motorist etiquette via safety risk estimation.

Accident Characteristics: 

Head On: This type of accident is solely predicted by the Type A trait of urgency. In turn, urgency is predicted by a combination of factors, two of these (4 and 5) are derived from the CFQ factor analysis, the others are; high life imbalance and competitiveness (two other Type A traits), elevated cognitive anxiety, low safety risk estimation, and high sensation seeking.

Evaluating the Model:      

* With respective correlation coefficients

Examining the values for each of these paths in can be seen that the strongest path to Head On Accident is the indirect path from the TABP trait Life Imbalance via the TABP trait Sense of Urgency.

Rear End: This type of accident is solely predicted by high exposure. In turn, high exposure is predicted by low cognitive anxiety and high risky driving behavior.

Evaluating the Model:

* With respective correlation coefficients

Examining the values for each of these paths in can be seen that the strongest path to Rear End Accident is the indirect path from Cognitive Anxiety via Exposure to the Road Environment.

Single-Vehicle Accident: This type of accident is predicted by high risky driving behavior, high sensation seeking, and a high level of competitiveness (a Type A trait).

Evaluating the Model:

*With respective correlation coefficients.

Examining the values for each of these paths in can be seen that the strongest path to Single Vehicle Accident is the direct path from Safety Risk Estimation.

Accident Involving Pedestrian or Cyclist: This type of accident is predicted by low exposure, low vehicle maintenance, low safety risk estimation.

Evaluating the Model:

* With respective correlation coefficients.

Examining the values for each of these paths in can be seen that the strongest path to Accident Involving either Pedestrian or Cyclist is the direct path from Vehicle Maintenance.

Discussion: 

Comparison of Single and Multiple RTA Path Diagrams

Single RTAs: By referring to the tables of sex of respondents by type of accident reported and number of reported accidents, it can be observed that over 76% of recorded RTAs, within the last three years, were isolated incidents. Initial observation of the path diagram for being involved in a single accident within the last three years shows that there are a number of indirect routes to being involved in an accident in addition to one direct route. The indirect routes are comprised of two endogenous variables (in addition to the primary dependent variable; one accident). These are exposure to the road environment followed by cognitive anxiety. In turn, eight exogenous factors independently complete the ‘chain’; they are as follows, somatic anxiety, sex of road user, cognitive failure, risky driving behavior, vehicle maintenance, sense of urgency and sensation seeking.

By taking a closer examination of the path diagram, the relationships between the factors are revealed. Exposure, the strongest predictor of being involved in a single accident within the last three years, is positively related to the dependent variable. From this we can assume that as exposure increases so to does the likelihood of being involved in an accident. To elucidate, the exposure factor is comprised of a number of variables, these include, weather conditions that the road user is prepared to drive in, how many times a day they drive, the highways they typically drive on, and their average miles per year. For arguments sake, if a road user was typically prepared to drive in all weather conditions, was likely to drive more than twice a day, predominantly used urban roads, and on average drove more than 25,000 miles per year, he would be classed as a high exposure road user. On the other hand, a road user who is only prepared to drive after dark (and avoids heavy rain, fog and snow and ice), uses their vehicle less than seven times a week, prefers motorways and dual carriageways, and has an annual mileage between 100-5000 miles, would be classed as a low exposure road user. Research has also shown that very high exposure is likely to lead to fatigue, which in turn may cause the motorist to fall asleep ‘at the wheel’ (e.g. Brown, 1993).

Howard & Joint (1994) report that the business motorist is most likely to fall asleep at the wheel, especially when on a long distance assignment, an a long journey, or commuting across country. Howard & Joint (1994) went on to argue that the number of motorists affected by this problem is likely to be on the rise. In the mid 1980’s, a study of motorway motorists showed fatigue to be a factor in 11 per cent of accidents. The research also showed that around 5 per cent of motorway motorists had lost consciousness, whilst driving, before their accident. The further people travel, the more likely they were to feel sleepy, according to the survey. Of those who had driven less than 100 miles that day 3 per cent felt sleepy at the time of their accident. This figure rose to 14 per cent of people who travel more than 200 miles. More recently, BR Automotive, the auction and car-leasing group, found that approximately 50 per cent of their sample of 1,250 business motorists had fallen asleep at the wheel on one or more occasions. On the day that they were questioned 39 per cent planned to spend six to eight hours behind the wheel on business, whereas 43 per cent planned to spend between three and five hours driving. Motorists were once forced to stop if they wanted to make a telephone call, giving them a chance to ‘stretch their legs’, take some fresh air and visit the toilets. Today, with the advent of cellular phone technology, motorists can technically drive and use the phone at the same time.

Furthermore, common sense dictates that driving in potentially hazardous conditions, such as fog, snow and ice, dramatically increase risk due to poor visibility, drastically reduced breaking distances and probability of vehicle skidding.

To further investigate the chain of single accident causation, exposure was entered into the linear regression analysis as the dependent variable. This yielded one predictor, cognitive anxiety. This, in turn, goes some way to identifying the role over-confidence plays in single accident causation. To explain further, cognitive anxiety is a component of trait anxiety, or A-trait as it is sometimes referred. This component is typified by the person affected displaying a tendency to perceive certain situations as threatening thereby displaying negative thought patterns such as worry, feelings of self doubt, and nervousness. This may play a crucial role in terms of accident causation, as the diagram shows, individuals with high cognitive anxiety are less likely to have the high exposure to the road environment that in turn could lead to the involvement in an accident. On the other hand, road users with low cognitive anxiety are more likely to have greater self-confidence in their driving ability and are therefore less reluctant to expose themselves to greater mileage and frequency of driving, thus increasing the likelihood of involvement in an accident. To explain further, situations that cause fear or anxiety, depend on people’s perceptions and emotional reactivity. Because of individual differences in temperament or experience, some people may find a situation, such as driving, stressful and others may not. One of the most important variables that determine whether an aversive stimulus will cause a stress reaction is the degree to which the situation can be controlled. Situations that permit some control are less likely to produce signs of stress than those in which other people, or machines, control the situation (Gatchel, Baum & Krants, 1989). On the other hand, cognitive anxiety may be related to affective responsibility in the form of altruistic and moral behavior. A high level of cognitive anxiety might be associated with being fearful of causing injury or damage to other road-users, and care about others' feelings. While a low level of cognitive anxiety might be associated with the disregard, or minimal thought, for the feelings and rights of other road-users, and the denial of culpability after poor or dangerous driving. In either case, the anxious individual finds the easiest method to reduce, or rather control, the anxiety is to minimize the time spent on the road.

Alternatively, cognitive anxiety could be a reflection of the individuals’ self-efficacy. To explain, self-efficacy is a perceived ability to perform a behavior and an outcome expectation. In the driving situation this could manifest itself as an optimistic self -belief in one's driving ability, and a conclusion that their driving skill will bring about a desired outcome, e.g. getting to a destination faster and more efficiently.

Most of the research into trait anxiety and self-efficacy has found a significant relationship between the two. However, this research has mostly focused on areas such as academic test anxiety (e.g. Wang & Liu, 2000), and occupational stress (e.g. Kumari & Singh, 2000). Whereas, the majority of the work on self-efficacy and driving has concentrated on driving whilst under the influence of alcohol, (e.g. Wells-Parker et al., 2000). Nonetheless, past research has shown, when considering the similarity between self-efficacy and the absence of cognitive anxiety, similar findings to those implied by the path diagram. For example, Farrow & Brissing (1990) found that males perceived greater driving skill in risky situations and used their motor vehicle to enhance self-efficacy more than females. In addition, they found that male road users scored higher on anger/hostility and sensation seeking scales. Clearly, further research into other driving related issues and self-efficacy is required.

Further examination of the exogenous factors finds somatic anxiety having a causal relationship with cognitive anxiety. Past research has suggested that somatic anxiety will affect performance if the extent of the somatic response is so large that the individual becomes excessively concerned and distracted with their perceived physiological state, (e.g. Martens et al., 1990). Nevertheless, the results of the analysis are still surprising insofar as the model indicates the physical aspect of anxiety, i.e. the palpitations, sweating and feelings of nausea precede the psychological. This is similar to Zajonc et al’s (1989) theory that proposed a radical explanation of emotional expression. Zajonc et al. (1989) suggested that facial muscles act to tighten facial blood vessels and regulate cerebral blood flow, which in turn influences subjective feelings through a register in the brain of subjective temperature. In other words, the subjective experience of emotions follows facial expression rather than precedes it. Physiologically, the internal carotid crosses the cavernous sinus, which, according to Zajonc et al. (1989), may respond to facial muscle tightening and relaxing. In this view, raising the temperature of brain blood leads to unpleasant feelings, and lowering it leads to pleasant ones (Adelman & Zajonc, 1989). Something similar may be occurring in the path diagram. In this case, it is likely that the adverse physical symptoms of anxiety are eliciting facial muscle tightening which in turn raises the temperature of the brain; this then leads to negative affect in the form of cognitive anxiety.

Sex was the next variable to be identified by the regression analysis In this respect, it can be observed that females are more likely to display cognitive anxiety. This also indicates that females are less likely to have higher exposure to the road environment that could ultimately lead to the occurrence of RTA. This is certainly in line with Parry’s (1968) findings that showed females, in general, displaying more anxiety than males when driving. Parry (1968) went on to add that this was likely due to females having a less aggressive nature both in and outside the motoring situation. Fundamentally, Parry’s (1968) assumption was that males were the predominant sex on the road and their very presence made the road an aggressive place. As a result, females entering this aggressive environment were more likely to exhibit anxiety. In an attempt to diminish this anxiety, female motorists may dramatically reduce the amount of time they spend driving, i.e. exposure minimized.

The next exogenous factor that predicts cognitive anxiety is cognitive failure (or slips and lapses as it is sometimes referred). Slips occur when the understanding of the situation is correct and the correct intention is formulated, but the wrong action is accidentally triggered, i.e. slips represent the occurrence of an alternative and incorrect action. On the other hand, lapses represent the failure to carry out an action. As such, they are tied directly to failures of memory. In this case, the path diagram shows a positive relationship between cognitive anxiety and cognitive failure. A likely explanation for this stems from road users self awareness of their own cognitive limitations. In other words, individuals who are prone to cognitive failure are likely to be conscious of the fact. Furthermore, they are more likely to become anxious when they encounter a situation that could be potentially dangerous (such as driving) if they made a slip-up or had a lapse of attention or memory. This could lead to an increase in stress and anxiety when driving, and in turn (as the path diagram demonstrates) low exposure through concern over driving. It should also be noted that cognitive failure correlates highly with somatic anxiety the second component of trait anxiety. This was to be expected as the somatic component of anxiety often accompanies the cognitive component if a highly stressful situation is encountered.

The next factor that forms part of the route to having one accident is risky driving behavior (e.g. jumping traffic queues, driving too close to the vehicle in front). In this case, the relationship between risky driving behavior and cognitive anxiety is negative. This means that road users who display a high degree of risky driving behavior are more likely to have low cognitive anxiety, i.e. they are confident (possibly over-confident) in their abilities as a motorist. Not surprisingly, there is also a highly significant correlation (and covariance) between this factor and the final endogenous factor; sensation seeking. This bolsters established research that has discovered similar results. For example, As Arnett, Offer & Fine, (1997) found, sensation seeking correlated highly with risky driving behavior when paired with aggressiveness. Indeed, as risky driving behavior and sensation seeking are so significantly correlated (they have identical regression coefficients with cognitive anxiety) one can assume that they are closely interrelated. In this respect, it can also be assumed that, similar to sensation seeking, risky driving behavior is also a mutually reinforcing activity, i.e. if the consequence of an action, like driving over the speed limit, causes positive affect within the perpetrator it is likely to be repeated, (e.g. Lecoeur et al., 1997). In other words, if an episode of risky driving elicits a 'high' within the motorist it is conceivable that the motorist will attempt to repeat the episode in order to achieve the same, or similar, feeling of elation. In addition, the behavior is also likely to be repeated as success raises self-efficacy and failure lowers it, i.e. if the episode of risky driving has a successful outcome (no RTA) the road user is more likely to have his belief of being a good motorist boosted. However, it should be noted that the fundamental difference between the two factors is likely to be state and trait. Sensation seeking is likely to be more general in nature, and extend to other areas of life. On the other hand, risky driving behavior is, as the phrase suggests, specific to driving vehicles. It should also be noted that there was also a significant relationship between high risky driving behavior and male road users. This goes some way in supporting the established research and government data that illustrate the over representation of male road users in RTA statistics. Subsequently, another significant relationship was recorded between high sensation seeking and male road users.

It was also discovered that road users that never service their vehicle were more likely to exhibit high cognitive anxiety. Whereas, road users that frequently service their vehicle were less likely to display cognitive anxiety. This propensity is likely to occur, as the motorist is more confident to undertake his vehicles servicing. However, the reason why this preoccupation with the safety and maintenance of the vehicle eventually leads to high exposure and RTA is unclear. Perhaps periodically maintaining ones vehicle is indicative of an enthusiastic motorist, this would go some way to explain the connection between low cognitive anxiety and high exposure. A significant correlation was also noted between vehicle maintenance and sex of road user, with males being more likely to service their vehicles on a regular basis and females being more likely to neglect their brakes, oil/ coolant levels and tyre pressures. Behavior such as this can generally be observed in young male road-users, i.e. they can often be found working on their vehicles in an attempt to improve performance and reliability.

The next exogenous factor to predict cognitive anxiety is the TABP trait sense of urgency. In the path diagram, it can be observed that a high sense of urgency is related to high cognitive anxiety. Again, this is faithful to what is already known about individuals who exhibit this trait. It is long established that Type A individuals seem to be in a constant struggle against the clock. Often, they quickly become impatient with delays and unproductive time, schedule commitments too tightly, and try to do more than one thing at a time. This can be highly stressful and is likely to lead to a heightened level of cognitive anxiety. As Perry (1986) found, sense of urgency was a significant factor that leads to a deficit in driving performance. A significant relationship was also noted between urgency and risky driving behavior. This is to be expected, as road users in a hurry maybe more likely to take extra chances when driving, e.g. speeding, overtaking on the inside lane, making calls on their cellular phone, etc.

Finally, the remaining exogenous variable is a direct path from age to one accident within the last three years. Predictably, the path diagram indicates that the younger the road user, the more likely it is that they will be involved in an RTA. By examining Table , it can be observed that road users aged between 17-25 have had the most singularly occurring RTAs within the last three years. Indeed, the table indicates that of the sample of 17-25 year olds, over 43% had been involved in one accident. The age group 26-34 followed closely behind with over 40% of the sampled road users reporting that they had also been involved in one accident within three years. This trend in decreasing percentiles continued until the 53 plus age group, at which stage, there was an upturn in the RTAs. This adds further support to government statistics, and established research, that states young motorists are more at risk of being involved in RTAs compared to older motorists.

Examining the correlations with the other endogenous factors and variables, it can be observed that the younger age groups have the strongest relationship with sensation seeking and risky driving behavior. In each case, the younger the road user the more likely it is they will engage in sensation seeking and risky driving behavior. Again, this supports established research (e.g. Trimpop, Kerr & Kirkaldy, 1998; Arnett, Offer & Fine, 1997).

Multiple RTAs: By referring to the tables of sex of respondents by type of accident reported and number of reported accidents, it can be observed that nearly 24% of recorded RTAs, within the last three years, occurred on two or more occasions. By examining the path diagram, being involved in multiple RTAs (two or more) within the last three years, it can be observed that there are two possible endogenous factors that the exogenous factors predict en route to multiple accidents. The first of these, safety risk estimation, was originally conceived through factor analysis of the aberrant driving behavior items on the questionnaire. This factor comprises of the items that are related to the way motorists take risks that are more likely to jeopardize their own safety, and driving license, rather than the safety of others. To elucidate, if a motorist fails to wear a seat belt (one of the factor variables) it will significantly increase personal injury in the event of an RTA, however the failure to perform this action will have no bearing on the other motorist’s safety. In addition, driving without MOT and/or insurance (the second variable in the factor) is likely to lead to a fine and penalty points in the event of a stop check by police. Whereas, driving without an MOT is likely to lead to an invalid insurance policy in the event of an accident. The annual MOT is also important in the identification of vehicle safety faults.

The first exogenous factor to predict safety risk estimation is motorist etiquette. This factor was extracted from the aberrant driving behavior items, and is characterized by behaviors such as, failing to use indicators, and failing to dip headlights for an incoming vehicle at night. Essentially, these behaviors do not necessarily have a direct relationship with sensation seeking. However, they are more closely linked to a disregard for other motorists. In this case, low motorist etiquette predicts low safety risk estimation, i.e. a road user who has little consideration for other road users would also likely disregard their own personal safety and their vehicle’s legality. In addition, it can also be observed that motorist etiquette correlates highly with risky driving behavior, which was also extracted from the same original aberrant driving behavior items on the questionnaire. In turn, this chain of behaviors would lead to the occurrence of two or more RTAs within the last three years. 

Risky driving behavior, the next exogenous factor to predict safety risk estimation, revealed behaviors, such as speeding, jumping traffic queues, and racing against other vehicles, that were related to safety risk estimation. In other words, a road user who frequently engages in risky driving is more likely to disregard their own personal safety and their vehicle’s legality. This then goes on to predict multiple accidents.

An alternative route, still utilizing risky driving behavior, has the exogenous factor exerting a strong predictive influence over sensation seeking. As mentioned earlier in the text, this relationship is predictable as the two factors have been shown in previous research to correlate highly (e.g. Arnett, Offer & Fine, 1997).

By referring to the tables for sex and age of respondents by type of accident reported, males can be observed as being over-represented in recorded multiple RTAs, i.e. over 69% of road users that had reported involvement in two or more RTAs within the last three years were male. Examining the path diagram, males are found to be more likely to exhibit sensation-seeking behaviors that go on to predict multiple RTAs. From this it can be assumed that the mediating factor in male multiple RTAs is sensation seeking behaviors. This discovery offers further support to earlier research, such as Arnett’s (1991), that found sensation seeking induced driving recklessness common in young adult males. Whereas, Zhong et al. (1998) found young males more likely to engage in risk taking behaviors such as driving too fast for the conditions, exceeding the speed limit, and driving while impaired by either alcohol or drugs.

Surprisingly, the results indicate that the 26-34 age group, followed by the 35-43 age group, have had the greater incidence of multiple RTAs within the last three years. The age group least represented in multiple RTAs was the 53 plus group, followed by 44-52 year olds. This finding offers further support to Arnett (1991), who found driving related sensation seeking to strongly persist in male motorists aged up to twenty-seven.

Likewise, the path diagram also shows a strong relationship between sex and insurance group, insofar as male sensation seekers preferring vehicles that have higher insurance groups. To reiterate, as a rule, the higher the insurance group the greater the performance and desirability of the vehicle. From this it can be assumed that sensation seekers favor vehicles with greater performance in order to reach a feeling of elation upon demand more easily. For example, they can accelerate quicker, and as their vehicles are more likely to have superior road holding, they can approach bends in the road faster.

It was also noted that two TABP character traits predicted involvement in sensation seeking activities that would lead to the occurrence of multiple RTAs. Both hostility and urgency had the strongest relationship between any two exogenous factors in the path diagram, thereby demonstrating their common association. Indeed, research has shown that there is a relationship between hostility and sensation seeking. For example, McMillan et al. (1992) found road users with multiple offences for driving under the influence of alcohol were significantly higher than first offenders in hostility and sensation seeking. Multiple offenders also had significantly more accidents, and traffic tickets than first offenders. In addition, Whiteside & Lynam (2001) found an association, via impulsiveness, between urgency and sensation seeking.

The last exogenous factor to predict sensation seeking, and ultimately multiple RTAs, is cognitive anxiety. Similar to the path diagram for being involved in a single accident within the last three years, we find individuals with high cognitive anxiety being less likely to have high sensation seeking behavior that would eventually lead to the involvement in multiple accidents. On the other hand, road users with low cognitive anxiety are more likely to have greater self-confidence in their driving ability and are therefore more likely to display higher sensation seeking behavior and, in turn, increased likelihood of involvement in two or more RTAs.

A Comparison of RTA Characteristics

  Previous research that has investigated the characteristics of RTAs, e.g. head-on, rear-end, single vehicle accident, etc., have focused on basic variables such as, sex, age, and occupant role in the crashed vehicle(s), i.e. motorist or passenger (e.g. Tavris et al., 2001). Such research has consistently found diversity between males and females, and the various age ranges, with respect to the RTA characteristics. However, other psychologically orientated factors have largely been ignored. In view of this, a further set of regression analyses, and path diagrams, was produced in order to identify the roles personality, cognition, and behavior plays in specific accident causation.

Head-on Accident: The first type of RTA to be distinguished was Head-on accident (where two vehicles traveling in opposite directions on the same road collide). By referring to the path diagram, it can be observed that this revealed one endogenous factor that predicted a further seven exogenous factors. This data is comparable to Perry's (1986) study that found TABP traits significantly related to the incidence of RTAs and violations. Perry (1986) also indicated that, according to his findings, the sense of urgency factor was the main contributor to the deficit in driving performance that led to RTAs and violations.  However, Parry's (1986) sample size was low, and as a result, the findings were generalized to include all types of RTA characteristics. However, based upon the sample of respondents, the results of this study have indicated that true TABP road users, i.e. individuals that score highly in all four TABP traits, are only more likely to be involved in a Head-on accident. Whereas, individuals that exhibit one or two of the traits are more likely to be involved in a single RTA or multiple RTA within three years respectively.

Sense of urgency is the sole mediating factor in the path to Head-on accidents. Therefore, it can be assumed that it is the strongest Head-on RTA determining Type A trait from which all the other factors are channeled. It can also be assumed that this causality occurs as urgency brings with it increased speed and a predilection to take more risks in order to get a destination in less time. This is exacerbated by significant life-imbalance, i.e. the Type A individual is, as a rule, highly work orientated, and therefore is more likely to have a 'no holds barred' approach to achieving deadlines, and being seen as efficient. In addition, the presence of the Type A competitive achievement orientation trait adds to the probability that the road user will resort to increased speed and risk taking to get ahead of the other motorists. Whilst hostility compounds the likelihood that the Type A road user is unlikely to 'back down' from an altercation, or is likely to perceive another road user's act as having hostile intent. Urgency is further compounded by increased cognitive anxiety and cognitive failure. This suggests that, as pace of life increases so to does fear of failure. In turn, pre-occupation with getting to one's destination on time causes a competition of cognitive resources. For example, over concern about being late diverts attention away from the task, i.e. driving. This is likely to lead to an increase in the occurrence of slips, lapses and mistakes. All too often, these result in a RTA (but in this case the characteristic is Head-on).

Earlier in the text the fundamental difference between sensation seeking and risky driving behavior were discussed. The conclusions drawn indicated that the underlying difference between the two was trait and state respectively. By referring to the path diagram for Head-on accidents, sensation seeking is shown to have an indirect effect, via urgency, in the chain of accident causation. From this it can be assumed that road users that display high sensation seeking in general are more likely to be exhibit urgency when driving. This can be corroborated by recent research that has shown both urgency and sensation as being two distinct, but related, facets of personality associated with impulsive behavior (Whiteside & Lynam, 2001).

The final factor to predict urgency is safety risk estimation. However, contrary to the positive relationship found with multiple accident involvement, high sense of urgency, in this case, is likely to be predicted by low safety risk estimation. One likely reason for this may stem from the typical Type A's predilection for fastidiousness. In the driving situation, this would manifest itself as a pre-occupation with   seatbelt fastening, and servicing their vehicles at regular intervals.

Rear-end Accidents: A succession of regression analyses revealed that the path diagram for rear-end accidents (Diagram 4a & 4b) was identical to the path diagram for involvement in one RTA within three years (Diagram 1a & 1b), save for a marginally higher beta coefficient between exposure and the dependent variable. From this it can be assumed that, based upon the road users sampled, the majority of single accidents that occurred, within the three years prior to the commencement of the study, have had rear-end (where two vehicles traveling in the same direction on the same road collide)  characteristics. Indeed, by referring to the tables for age and sex of respondents by type of accident reported it can be observed that the 42% of the recorded RTAs were rear-end in origin.

Traditionally, the culpability of rear-end accident causation has been seen as clearly the fault of the following vehicle. Of course there are some exceptions such as a 'domino effect' series of rear-end collisions, where one vehicle is struck from the rear with such force that it sets in motion a chain of collisions. However, as a rule, rear-end accidents are seen as the product of close following vehicles, or tail-gating as it is sometimes referred, and their significantly reduced stopping distances.

Single-vehicle Accidents: Referring to the path diagrams for single-vehicle accidents (Diagrams 5a & 5b), it can be observed that there are three distinct, and indirect routes, to having an RTA where one vehicle loses control and usually runs of the road.

The first of these is, similar to the path diagram for involvement in multiple RTAs (Diagrams 2a & 2b), motorist etiquette and risky driving behavior mediated by safety risk estimation. A third exogenous variable (main vehicle used) was extracted in the regression analysis. However, the variable, which implicated the sampled motor cyclists as being low in safety risk estimation, was discounted from the path diagram as items within the factor, safety risk estimation, bore little relevance to motor cycle ownership.

The second route to involvement in a single-vehicle accident relies, similar to the path diagram for head-on accident, upon the four TABP traits. However, contrary to head-on accidents, the principle personality trait that mediates all the others is competitive achievement orientation. This likely to manifest itself as a predisposition towards spontaneous acceleration against other road users or frequently 'exploring' the limitations of their motor vehicle, either assumption is bolstered by the presence of a fourth exogenous variable that indicates the typical Type As' penchant for high performance vehicles. Whilst urgency points towards goal directed behavior, such as getting to an appointment on time, competitive behavior suggests that the Type A road user is influenced by recreation or entertainment. From this it can be concluded that the Type A road user is more likely to be involved in a head-on accident when driving for business, or necessity, and a single-vehicle accident when driving for pleasure.

The third indirect route to single-vehicle accidents distinguishes sensation seeking as the mediating endogenous factor. Again, similar to the path diagram for multiple accidents (Diagram 2a &2b), it can be observed that the same combination of factors and variables predict the behavior. The only difference being, a slightly elevated beta coefficient between sensation seeking and the dependent variable.

Accidents Involving Pedestrian or Cyclist: There are three indirect paths that predict accidents involving pedestrians and/or cyclists (i.e. an RTA in which a vehicle collides with either a pedestrian or a cyclist). The first path is mediated by exposure, to reiterate, a factor that combines annual mileage, conditions that may be encountered when driving, frequency of driving, and the most frequent type of highway used. Exposure to the road environment has already been implicated as a major predictive factor when a road user is involved in only one accident within three years and, almost identically, in rear-end accidents. However, contrary to the previous path diagrams, the relationship with pedestrian/cyclist RTAs is in the opposite direction. Here it can be observed, that low exposure to the road environment leads to the involvement in an RTA. Similar to previous path diagrams, high cognitive anxiety is seen to have a negative relationship with exposure, i.e. the higher the anxiety the lower the exposure. It can be concluded from these relationships that the underlying mechanism is a combination of parallel processing and lack of experience.

Over-caution, a pre-occupation with safety and an unwillingness to be exposed to any unnecessary risks, under normal circumstances, these would be regarded as commendable characteristics. However, when these behaviors manifest to such a high degree, as in driving related cognitive anxiety, they can have a deleterious effect upon task performance. Indeed, this effect often leads to a competition of cognitive resources and so, as a result, oncoming features in the road environment, and the execution of driving maneuvers are processed in a parallel fashion. In other words, stimuli requiring attention, such as speed, distance to the vehicle in front, position in the road, when indication and turning is required, etc. would be dealt with individually in a time consuming manner. This delay in perception can be hazardous as other road users may fail to be considered.

This is often reflected in the susceptibility of novice road users to accidents. To elucidate, every new motorist goes through a period of habituation to the driving environment. While gaining experience, they process information and operate their vehicle in a parallel manner. For example, they may cut short a conversation with a passenger while they undertake an irregular manoeuvre, as they cannot attend to thee two sets of stimuli. The more experienced the motorist, the more automatic and serial the information processing and operation, (James, 1997).

A typical example of parallel processing sees the road user attempting to attend to a number of stimuli simultaneously. In the first instance, the motorist is concentrating intently on the road ahead unconsciously they have decelerated. Switching their attention to check their speedometer, they realize that they are moving below the speed limit, and so they adjust their speed accordingly, they then switch back to monitoring the road ahead. This will continue for the duration of the journey. However, if there are other distracting stimuli present, e.g. talkative passenger, mobile phone, heavy traffic, etc. this will add to the competition of cognitive resources. Nonetheless, for a time, the road users attention is diverted, and this is when an RTA can occur. This likelihood is compounded by the presence of the factor that denotes low exposure. To explain, notwithstanding the negative aspects, i.e. hazardous driving conditions and fatigue, high exposure is also indicative of experience. In general, the greater the mileage covered the greater the likelihood of being able to interpret potential hazards in the road. Indeed, research has shown that inexperienced road users frequently fixate on a position approximately eight to ten metres in front of their vehicle. Whereas, experienced road users constantly vary their scanning pattern of the road ahead (e.g. Green & Senders, 1999). In the case of accidents involving pedestrians and/or cyclists, the inexperienced, and anxious, road user may fail to notice a pedestrian or cyclist venture into the path of their vehicle.

Similarly, those with their attention focused on other matters not related to the road will have little thought for other road-users. Again, this stems from the same competition of cognitive resources. For example, a parent with a car full of noisy children on the 'school run', or a business man on his mobile phone, will be less mindful of their driving and other road-users, inadvertently they may park or maneuver precariously and be subject to an RTA.

The next indirect path reveals a relationship between pedestrian/cyclist RTAs and vehicle maintenance. Here it can be observed that low vehicle maintenance is more likely to predict the occurrence of an RTA. A realistic assumption considering incorrect tyre pressures and poorly maintained brakes will lead to a dramatic reduction in stopping distances. This is further compounded by three other predictors, insurance group, cognitive failure, and sex of motorist. Firstly, vehicles of higher insurance groups are more likely to be maintained less than vehicles from lower insurance groups. This is possibly due to their complicated nature, or the fact that they are still under manufacturers warranty. Secondly, motorists high in cognitive failure are shown to be more likely to maintain their vehicles on a regular basis. This may be due to them not remembering the last time they serviced their vehicle, so they service it again to be sure. Thirdly, the diagram indicates that females are more likely to have their vehicles maintained on a regular basis in comparison to males.

The final indirect path to pedestrian/cyclist RTAs sees low safety risk estimation as being mediating factor by which risky driving behavior and motorist etiquette predict accident involvement. Again, this relationship is similar to the indirect   path observable for single-vehicle accidents.

Re-examining TABP and RTAs

At best, the results of this study only offer partial support to the previous research that has claimed a direct causal relationship between TABP and RTA risk, (e.g. Perry, 1986; Magnavita et al, 1997). Only two characteristics (hostility and urgency), normally associated with TABP, were found to be associated with sensation seeking behavior that went on to predict being involved in multiple RTAs within three years. Moreover, only one TABP characteristic (urgency) was found to be involved in the prediction of the occurrence of one RTA within three years via mediating endogenous factors of cognitive anxiety and exposure to the road environment.

However, when analyzing RTAs by characteristics, i.e. type of crash, it was discovered that two indirect paths to both head-on and single-vehicle accidents involved all four of the personality traits. It can be inferred from this that TABP's relationship with accident involvement is wholly dependent upon the characteristics of the RTA that the motorist is involved in.

Earlier in the introduction section, it was discussed in some detail why some individuals engage in maneuvers that are likely to lead to an accident. By examining the path diagrams for head-on accidents (Diagram 3a & 3b), and single-vehicle accidents (Diagram 5a &5b), it can be observed that motorists who exhibit Type A characteristics are more likely to display these maneuvers. In order to understand what is occurring between these personality traits and life threatening RTAs, it will be necessary to examine in detail the underlying mechanism, albeit theoretically.         

One possible explanation comes from the early frustration aggression hypothesis, (Dollard et al., 1939). To summarize, intended partly to translate some of Freud's psychoanalytic concepts into learning theory terms, Dollard et al. (1939) published a controversial work entitled Frustration and Aggression. In their own words they saw aggression as, "...always a consequence of frustration and, contrariwise, ...the existence of frustration always leading to some form of aggression", (Dollard et al., 1939). In other words, while in agreement with Freud that aggression was an innate response; Dollard et al. (1939) believed that frustrating situations, that impedes or prevents some form of ongoing goal-directed behavior, would act as the 'catalyst' for aggressive behavior. A number of subsequent studies into the displacement of aggression offered some support for this view. For example, research by Barker et al. (1941) observed that children who were deliberately frustrated by being denied access to attractive toys behaved aggressively towards toys with which they were allowed to play.

Nevertheless, it soon became apparent that the frustration-aggression hypothesis, in its original form, was an overstatement. In answer to this, Miller (1941) subtly altered the hypothesis by arguing that frustration was an instigator of aggression, but more importantly, situational factors, such as learned inhibition and fear of retaliation may prevent actual aggressive behavior from occurring. This meant that although frustration may make aggression more likely, it was far from being the sufficient cause of aggression. Indeed, it soon became widely accepted that frustration could produce a variety of responses, notwithstanding aggression, which included regression, depression, and lethargy, (Seligman, 1975). In addition, frustration has also been shown in different situations to produce different responses within different individuals. This was further supported by Bandura's (1973) claim that frustration might be a source of arousal, but frustration-induced arousal (like other types of arousal) could have a variety of outcomes, of which aggression was only one, and whether it actually occurred was more the result of learned patterns of behavior triggered by environmental cues. Nevertheless, Kulik & Brown (1979) found that frustration was more likely to produce aggression if the frustrating stimulus had not been anticipated, and also if individuals had reason to believe that the person(s) responsible for the frustration did so deliberately, and without good cause, thereupon showing the importance of cognitive factors as cues for aggressive behavior.

From a similar line of argument, Berkowitz (1962) proposed a number of modifications to the original frustration-aggression hypothesis. His major argument was that frustration produced anger rather than aggression. Moreover, what was important about frustration was the fact that it was psychologically painful, and anything which is psychologically, or physically painful can lead to an aggressive 'backlash'. For anger or psychological pain to be converted into open hostility, certain cues are needed; these are environmental stimuli associated either with aggressive behavior or with the frustrating object or person. In support of this assumption, Marks (1990) found, in a study of moods and emotions on motorist behavior, that negative effects such as irritability and anger can find behavioral expression in hostile or aggressive acts. In turn, this would lead to deterioration in driving performance.

It has only been lately that the value of the hypothesis, as a possible explanation of road user aggression, has been recognized. Notably, Shinar (1998) has seen increasing traffic congestion and delays as a cause of the rise in aggressive driving. According to Shinar (1998), this increase stems from traffic conditions that hinder road users from reaching their goals, e.g. mobility and pleasure. NEED MORE FROM SHINAR

Putting these points in order, we may end up with a model similar to the one below, (Figure iv.).

Figure iv. The Chain of Active Accident Causation Using Berkowitz's Revised Frustration-Aggression Hypothesis (1962)

 

 

 

 

Time

Looking at the above example, (Figure iv.) frustration is likely to occur within the TABP motorist as their sense of urgency, a key component of the personality trait, will be impeded by the slow moving traffic. The effect of this will be amplified by their other key TABP component, hostility. As Kulik & Brown (1979) stated, anticipation at this stage may play an important role, i.e. if the slow traffic was anticipated then it would be less likely that frustration be transformed into anger. Alternatively, there may be an initial evaluation of who is responsible for the slow traffic. If a specific individual could be targeted, e.g. a car towing a caravan, or a learner motorist, then it would be more likely that anger followed frustration. At this point, if the stimulus remains unchanged, or if there is a revaluation of responsibility for the delay, anger may give way to open aggression and a risky driving maneuver. In support of this assumption, the New York Times July 18 (1997) reported the Chief Federal Highway Safety Officer for the United States, as declaring at a Congressional hearing on road safety, that several factors were involved in every car crash. However, rage was present in two-thirds of the 41,907 traffic related deaths, and in a third of the nonfatal crashes, during 1996, which resulted in three million injuries. "The more serious the crash, the more likely that aggressive driving was involved," he argued. However, Seligman (1975) noted that a motorist not prone to aggressive expression may exhibit more passive behaviors when confronted by the slow moving traffic, e.g. depression, lethargy.

  A number of studies have also suggested that the mere physical presence of weapons is enough to increase the occurrence of aggressive behavior. As Berkowitz (1968) stated:  "Guns not only permit violence, they can stimulate it as well. The finger pulls the trigger, but the trigger may also be pulling the finger." This may relate, to some extent, to the assumption that, when unabated, motorist based frustration leads to anger, aggression and the execution of risky maneuvers. Throughout the history of the motor vehicle the motorist has been encouraged by manufacturers to draw an analogy between their vehicles and an assortment of arms, e.g. Scimitar, Magnum, Rapier, Fireblade, etc., or an unrestrained and dangerous animal, e.g. Puma, Cobra, Mustang, Barracuda, etc. However, the association may extend beyond weapons, or wild and menacing creatures at the motorist's disposal.

By their very nature, most motor vehicles are steel 'jackets' encapsulating their occupant. It has been suggested that they could unconsciously could be regarded by the motorist as a form of 'suit of armour', thus lowering even further the inhibitions that 'hold back' the aggression (Marsh & Collett, 1986).

Strengths and Limitations of Path Analysis for RTAs

By using path analysis, it is possible to identify and focus on the factor in each route that would be easiest to modify with a view to reducing RTAs. In the case of being involved in a single accident, it is possible to declare that reducing exposure should decrease the likelihood of being involved in an RTA. Alternatively, if it is impractical to reduce exposure (e.g. the road user drives for a living) then we can attempt to lower the motorist's self-efficacy, making him comes to terms with his actual skill as a motorist rather than his perceived skill. As stated earlier in the text, research has consistently found that road users tend to over-rate their driving skill higher than it actually is (e.g. Brown & Groeger, 1998; Delhomme, 1991).

In each path diagram, additional arrows can be seen pointing into each endogenous node (factors taken as dependent variables) to signify unexplained variance. This is the variation within the factor due to other factors not included in the analysis. However, the variance explained by the factors in both multiple and single RTA path diagrams are still somewhat low. Therefore, it must be stressed that other factors, not included in the analyses, also contribute to RTAs. Ultimately, these factors would be difficult, if not impossible, to measure in a questionnaire based upon psychological correlates. For example, a study by Abdel & Radwan (2000) demonstrated the importance of heavy traffic volume, narrow lane width, larger number of lanes, urban roadway sections, narrow shoulder width and reduced median width in the causation of RTAs, indicating the importance of civil engineering factors. Whereas, Rumar et al. (1989) found headlights, a crucial factor in vehicle detection on the road, and despite relatively small numbers occurring in accident statistics, vehicle/equipment related failure should also be considered (e.g. Treat et al., 1977; Joshua & Garber, 1992). The assumption that there are other random factors at work, outside the scope of the questionnaire, is further encouraged by the failure of the multiple regression analysis to yield any significant results for intersection and sideswipe accident characteristics.

This dichotomy is also compounded by the likelihood that a proportion of the road users who completed the questionnaire were passive recipients, as opposed to active participants, in the causation of the RTAs recorded. A previous study by the author (McNally, 1999), using a smaller sample size and number of factors, found diversity between the path diagrams for passive and active accident involvement. However, this course of analysis was not pursued in this study due to a number of methodological problems. For example, it was discovered that attempting to distinguish between passive and active RTAs leads to inaccuracies within the data set. To explain, some road users may be inclined to report their involvement as being passive when, in truth, it should be active. The reason for this may stem from a distorted subjective perception of their culpability. For example, it may be found that, due to their inaction, they caused a vehicle to swerve and collide with another. In the first instance, it appears they were not to blame as their vehicle remained unscathed. However, under careful scrutiny, it is observable that they were the cause of the accident. Nevertheless, because their vehicle has no damage they believe it was the other party’s fault. Under these circumstances, it is easier to classify all RTAs equally and address the dichotomy later in the discussion.

Strengths and Limitations of the Internet Questionnaire

In the last few years, it has become possible to conduct meaningful behavioral research via the Internet. In May 1999, the American Psychological Society list of Psychological Research on the Net reported that the number of Internet based experiments and surveys had nearly doubled in less than twelve months. Of the experiments listed in the APS Web site, there were 24 in social psychology, 13 in cognitive, 8 in sensation/perception, 5 in health psychology, 4 in developmental, 3 in clinical, 3 in personality and industrial-organizational, 2 in biological, 2 in emotions, and one in general psychology. Although this list does not include all experiments, it gives a proportional estimate that indicates the growth of research conducted via the Web.

Surprisingly little research has gone into using the Internet as a method of psychological questionnaire delivery. The few studies that have investigated Internet based surveys have found that the Internet-mediated version had similar psychometric properties to its conventional equivalent and compared favorably as a measure of self-monitoring (Buchanan & smith, 1999). Further results suggest that findings from Web-based questionnaire research were comparable with results obtained using standard procedures (Davies, 1999). Whereas, Birnbaum (2000) argued that some comparisons show that Web data are of higher quality than lab data. However, surveys administered over the Internet have been plagued by low response rates, and at times have provoked anger from targeted individuals towards researchers sending offensive unsolicited e-mail, (e.g. O'Neil & Penrod, 2001; Cho & LaRose, 1999).

The results found a significant difference between male and female respondents in both questionnaires. To elucidate, significantly more females responded in the hand delivered questionnaire than males. In addition, a highly significant number of males responded to the e-questionnaire compared to females. Furthermore, it was also found that there were more respondents to the traditional method of questionnaire delivery despite the two methods running concurrently.

This goes some way to suggest that the Internet is good for some areas of social science research but not for others. The researcher suggests that the limited number of respondents to the Internet questionnaire was due to the passive nature of the internet, i.e. the researcher has to wait for the respondent to visit their site rather than actively ‘going out’ and sampling them. Information is everywhere on the Internet, existing in large quantities and continuously being created and revised. The constant distraction of this competing material can make it difficult to get people to come to your sight, let alone take the time to fill in a detailed questionnaire.

In addition, some websites are more likely to be visited by males rather than females (and vice versa). There may be an equal ratio of male and female motorists in the UK, but the research has indicated that females are less likely to pursue driving related issues as an interest on the World Wide Web. For fear of disappointment, careful consideration about target populations would benefit future Internet based surveys.

Implications for Road Safety Measures

In-vehicle Hardware: Many solutions have bee put forward in an attempt to curb the rising number of incidents of aggressive and risky driving. As a result, a number of researchers have speculated that some of these could take the shape of in-vehicle hardware. Part of the rationale for this stems from research that has shown that when people are aware of being observed they tend to modify their behavior. This phenomenon could potentially be used to encourage individuals to behave more safely when driving if there were means of providing feedback to the motorists about their behavior on the road. So-called 'vehicle data recorders' offer such a means of providing behavioral feedback by confronting motorists with their recorded driving actions. In one such experiment to test the feasibility of their use, Wouters & Bos (2000) conducted a field trial with the objective of investigating whether this feedback mechanism would reduce the number of road traffic accidents in everyday driving conditions. In the study, seven experimental vehicle fleets were involved, which varied widely in terms of the kind of transport sector concerned, the type of vehicles used, and the traffic circumstances in which the vehicles are operated. In total, accident and exposure data were collected for 840 vehicles of which 270 equipped with a recorder. During an observation period, which represented over three thousand vehicle years, the vehicles were involved in over eighteen hundred RTAs. Analysis of the effects of the use of data recorders in these fleets resulted in an average estimated accident reduction of some 20% thus demonstrating the usefulness of the device.

The Department of Transport has also shown interest in a apparatus invented a few years ago to complement the European Community's drive programme. DRACO, or Driving Accident Co-ordinating Observer, is a tool similar to the in flight recorder, or 'black-box', of an airplane. It is said to monitor, and record in the event of an accident, a vehicle's speed, acceleration, braking and illumination. Project researchers maintain that the device would significantly alter the behavior of road users.

Similarly, an in-car computer from the Netherlands has been developed that relays corrective audio-visual messages to the motorist taken from sophisticated electronic sensors situated throughout the vehicle. The computer, which is hoped will be made standard in all cars, also has an innovative collision avoidance system that activates a mechanism which increases counter-force on the accelerator pedal in the eventuality of venturing too near to the car in front. Nevertheless, lack of collusion between governments, motor vehicle manufacturers, and more importantly pressure groups, realistically consigns such devices to the realms of science fiction, at least for the foreseeable future.

Other methods of mechanical intervention include an advanced brake warning system (ABWS), which is a mechanism that activates the brake lights in response to a rapid disengagement of the gas pedal, before the motorist's foot reaches the brake pedal. Previous studies showed that such rapid releases of the accelerator pedal are typically followed by brake activation, and the ABWS can prevent a high percentage of rear-end collisions in which an attentive following motorist maintains a respectable distance from the vehicle ahead. In a study by Shinar (2000), the RTA involvement of 764 government cars and light trucks was tracked over an average period of 35 months. The vehicles were matched in pairs, one of each pair with the ABWS and one without it. Data analyses focused on collisions in which the government vehicles were rear-ended. Overall, ABWS-equipped vehicles were not significantly less involved in rear-end collisions. However, an examination of the struck vehicles showed that the ABWS-equipped vehicles were involved in fewer rear-end collisions per kilometre driven than were the vehicles without the ABWS.

However, another innovative idea that assists the motorist, more passively than the last two, and has seen widespread acceptance and use, is the Traffic Master. This is an navigational aid that enables the motorist to avoid congested areas that could cause undue stress, thus minimizing the risk of becoming an aggressor or victim of road rage.

At the time of writing, the American Automobile Association has recently announced that it intends to introduce a new universal hand-signal motorists can use to say "sorry", in the hope that it will curb unnecessary aggression on the road. As to this date however, the AAA are still unsure as to what this new signal will be, and have invited the public to come forward with any appropriate ideas.

  Educating the Road User: A number of researchers have found that they have been unable to contribute a comprehensive solution against RTAs by passively attempting to educate road users against risky driving behavior. For example, Brown & Groeger (1988) stated that, "...estimates of objective risk will ...be of limited usefulness to the individual road user". They went on to argue, "[that] they may contribute to strategic decisions, such as choice of transport mode, ...even tactical decisions [such as] ...driving to work on a route that minimizes turning across oncoming traffic. However, their effects on motorist's decision making in the long term appear to be relatively ineffective". In addition, it was demonstrated by Anderson (1978) that, "...factually based traffic safety materials did not reduce road users' involvement in accidents and offences during the six months after exposure to those materials, even when it was targeted specifically at age and sex subgroups of road users". In fact, it has been suggested that because the objective estimates of risk happen to be associated with past events (RTA statistics are mostly national, and tend to be over twelve months old), and not present or future actions, a negating and counterproductive effect may be occurring on subjective perception. This may be perpetuating the belief that accidents happen to other people, and not the individual to which the education is aimed at, (e.g. Parry, 1968; Anderson, 1978). Research has demonstrated that most road users are inclined to have higher estimations of their own driving ability (e.g. Finn & Bragg, 1986; Svenson et al, 1985), fostering a belief that they are more skilful, and safer, than their peers, (Brown & Groeger, 1988). As a result it has been suggested that, "...objective risk estimates will tend to be viewed as somewhat irrelevant by most road users, when assessing their own behavior, except as confirmation of their own superiority", (Brown & Groeger, 1988). In view of this, the Italian government has adopted a less than subtle way of attempting to distribute a message of road safety to the masses. This allows the Italian television news media to embrace a 'no holds barred' approach to reporting fatalities on the road, no doubt hoping to shock motorists into reducing their speed and driving more sensibly. Indeed, on the television news, it is possible to view the aftermaths of RTAs with dead victims lying on the highways, well before the viewing watershed. Whereas, the press often print current accident and fatality statistics followed by reported comments designed to instill fear or, in the very least, caution in the motorist. For example, at the height of the annual summer 'exodus' to the Italian holiday resorts, a local tabloid reported that the head of Italian police had instructed road police to be "ruthless" with unruly motorists, (Metro: Roma, Anno 1, Venerdi 11, Agosto 2000, page 16).

  Taking education a step further, suggestions for behavior modification have been proposed. For example, the Response-Cost Technique provided by Friedman & Rosenman (1974), suggested that, if they wished, road users exhibiting the characteristics of TABP could effectively overcome some of their competitive and urgent traits through actions similar to the one outlined below:

Whenever you catch yourself speeding up your car to get through a yellow light at an intersection, penalize yourself immediately by turning to the right at the next corner. Circle the block and approach the same corner and signal light again. After such penalization, you may find yourself racing a yellow light a second, but probably not a third time.

Friedman & Rosenman, (1974), p.264

A further example cited, is related to the motorist taking his vehicle's keys up long before the vehicle is in sight, i.e. the road user is prepared to drive well in advance of approaching the vehicle. In this case, the Response-Cost Technique would involve placing the keys back into the pocket each time they are impulsively taken out in advance of physically needing them. Friedman & Rosenman (1974) stated, that for this exercise to be effective the motorist must keep the keys in his pocket until they actually require to open the car door. Furthermore, it is reasoned that this exercise will impede the motorist from reinforcing anticipatory and hurried behavior typical of TABP.

Similar cognitive behavior modification techniques have proven on a number of occasions to be successful in overcoming the hostility component of TABP. For example, Navaco (1976) illustrated the positive effects of self-instruction coupled with relaxation training on an assortment of clients diagnosed as exhibiting 'anger control problems'. Self-instruction involved the methodical and regular practice of self-statements designed to deal with subjective anger scenarios that the client originally had difficulties coping with without getting angry. For example, "...as long as I keep my cool, I am in control of the situation, …my anger is a signal of what I need to do, …time for problem solving".   

Nevertheless, it would be naive to suppose that any behavior modification would be an easy accomplishment. Individuals that display TABP will, in most cases, oppose and withstand considerable efforts at altering their established mode of behavior. They do not want to have an accident, but at the same time, they do not want to change their behavior.  Indeed, as a rule, those Type As' that actually acknowledge that their personality is borderline aberrant will still believe that their traits aid them in coping with their daily routine. Moreover, competitiveness, pre-occupation with work, and a sense of urgency, are character traits that are often rewarded in business, and so are more likely to be regarded as essential and impractical to give up.

  Another example of further education for road users is advanced driving courses. However, suggesting that motorists undertake such courses to encourage better road skills when there is not enough incentive is unrealistic, i.e. insurance companies do not offer a significant enough reduction in premiums to merit the motorists' extra expenditure (both time and financial).

  Perhaps the answer to the dilemma is a revision of the way we deal with hostility on the road. As a rule, actions that constitute threats (e.g. police stop and cautions) are unlikely to elicit shifts in behavior with road users high in TABP or sensation seeking behavior. As stated earlier in the text, TABP and Sensation Seeking is so frequently rewarded that it is very resistant to change, (Suinn, 1977). For example, rather than 'dispensing' fines and endorsements that create bitter, and not better road users, a system could be devised that effectively tackles the 'cause of the disease' and not the 'symptoms'. Such systems could take the shape of compulsory attendance of behavior modification courses for those involved in repeat traffic violations or RTAs.      

For example, a pilot study, utilizing re-training and expert advice, in Wisconsin during the late 1960's involving 233 motorists with a history of high accident risk, found a dramatic decrease of approximately 96% of reported accidents in the following years (World Road News, 1966a). In addition, courses such as this could be coupled with continuous longitudinal testing and provisional entitlement to drive for repeat offenders. Such testing could also give us an insight into the stability of sensation seeking, and TABP and its sub-components over long periods.

The Cross-cultural Perspective: Initially, this study was to be conducted in two parts. The first part, outlined in this paper, was successfully directed at motorists within the United Kingdom. Whilst the second part would have explored whether or not the set of path diagrams gleaned from the UK data needed to be modified within other European cultures, namely Finland and Italy.

Europe these days is very much a contentious issue. For example, we have frequent debate over a single currency, trade, law, and numerous other topics. In this melee there is little scope for debate over our similarities, i.e. are we fundamentally different in every day matters such as driving behavior.

A great deal of research has supported the proposition that driver behavior does indeed differ between European states.

For example, Marsh & Collett (1986) reported a study by Forgas (1986) in which the author drove a Volkswagen through four European countries sporting an Australian registration plate. The Author discovered that the French, Spanish and Italian motorists were quicker to beep their horns at the VW when it failed to move at a clear junction. In comparison, it was found that the Germans were the least likely to use their horns. In the study, the Italians were the quickest to use their horns.

Moreover, official government statistics (International Road Traffic & Accident Database (OECD), Issue: May 2000) has shown stark differences in RTA fatalities and injuries within the European states. For example, In 1998 there were 25 motorcycle related deaths and 232 car related deaths in Finland. In the United Kingdom, there were 509 motorcycle related deaths and 1789 car related deaths. However in Italy, there were 1193 motorcycle related deaths and 3522 car related deaths. It appears, at first glance, the further south one travels in Europe, the greater the risk of RTA. However, if the statistics are considered per 100,000 population, we see an altogether different picture, i.e. Italy still has the highest fatality rate with 11.0 killed per 100,000 population. However, Finland becomes second highest with 7.8 killed per 100,000 population in comparison with the UK's lowest of 6.0 killed per 100,000 population. On the other hand, RTA injury accidents per 100,000 populations show the UK as being the highest at 415 injured per 100,000 populations, followed by Italy and Finland with 355 and 134 per 100,000 populations respectively. Despite these statistics, the researchers await data that will indicate reported RTAs were there has just been damage to vehicle(s) (regardless of fatality or injury).

The maintenance of fixed individual distance varies greatly in different human societies. Distances that may be considered as invasions of personal space depend very much on the individual’s culture and can easily lead to misunderstandings. In Southern Europe, people interact at close range, e.g. kissing when greeting and saying goodbye, and frequently touching during conversation. Indeed, this propensity has been found to extend from social to business situations (Gesteland, 1996). However, this type of behavior is found to be very disconcerting in the UK, and contact is usually kept to a minimum, (Hall, 1966). This propensity may well extend into the driving environment with increased likelihood of RTAs in Northern European countries through altercations brought about through ‘tailgating’. However, in Southern European countries, such as Italy, close driving is seen by many as the norm, and therefore less likely to elicit hostility between motorists. Nevertheless, this close following behavior may increase accident risk through reduced stopping distances.

Cross-cultural research has consistently found a causal link between TABP and RTA risk, e.g. the differences in the expression of TABP traits between PSV motorists in North America and India, (Evans, Palsane & Carrere, 1987); and between motorists in Finland and Australia, (Lajunen et al., 1998). This is likely to be partly due to cross-cultural differences in personality. For example, TABP variance between sub-populations has been reported in some cases to be as high as 55%, (e.g. Cohen, 1974; Howard, Cunningham & Rechnitzer, 1976). Nevertheless, the majority of the studies have a history of poor ecological validity, i.e. small sample sizes; lack of diversity between the occupations and sexes of the participants; and the possible psychometric variance arising from translating tests, such as Bortner’s (1966) TABP scale, into other languages, e.g. Magnavita et al., (1997). Based upon these propositions, it is anticipated that the research be continued, at a later date, to incorporate the other diverse European states.

  Conclusion

  It has been estimated that by the year 2010 there will be significantly more vehicles on the road in the UK. The subsequent road congestion from the increase in traffic volume and extended urbanization, brought about by the growth in population, will amplify exposure to accident risk even further. In addition, this is likely to be compounded by the stresses of driving longer distances too and from work, and having to deal on a daily basis with more vehicles and equally stressed motorists. Furthermore, the future is likely to bring other, yet to be identified, hazards that will broaden the factors that already influence RTA risk. For example, the unprecedented rise in the use of in car cellular phones has made it necessary to conduct more research into their usage whilst operating a motor vehicle. One such study, conducted in Canada, compared each motorist's phone use immediately preceding the collision to the same periods on several days before the RTA day. Phone billing records were used to analyze call patterns. Overall, 24% of the motorists had used a cellular phone within ten minutes before their RTAs (Redelmeier & Tibshirani, 1997).

Although to date there is no indication of mediation by TABP, the issue raises some concern when the deficit in reaction time performance of Type A's is considered, (Abrahams & Birren, 1973; Price & Clark, 1978; Glass, 1977). Moreover, as some personality types have been shown to be more likely to attempt more than one task at the same time its prudent to suggest that cellular phone usage while driving may lead to a growth in the establishment of non-response to road-traffic conditions, and therefore a decrement in crucial response time. Still more notably, it has been demonstrated that in car cellular phone use is between two and three times more distracting for road users aged fifty and over compared to younger road users, (McKnight & McKnight, 1997). These results could be more consequential if applied to TABP individuals of this age group.

The AA has long advised motorists not to drive for more than eight hours in a day, and to take a break of at least 20 minutes’ duration every three hours. Older motorists were advised to only drive for two hours between breaks. The AA Defensive Driving checklist, issued to all AA members, advises them to take breaks every two hours. However, this advice does not tackle the specific problem of the business motorist. The AA recommends eight hours driving a day is the maximum that can safely be undertaken, but this does not take into account time spent working. In the interest of road safety, the AA has called on all employers to urge their employees to take regular breaks. Furthermore, the AA have argued that fleet managers should be able to detect those employees, who have to travel long distances, that are prone to fatigue.

In conclusion, the data suggests that it may be worthwhile for companies and organizations to screen motorists for excessive vulnerability to the factors that have showed a likelihood to lead to RTA risk (e.g. over exposure to the road environment, and excessive sensation seeking behaviors), and to consider taking countermeasures. Moreover, the distinct age-related involvement patterns confirm the need to tailor prevention and safety programs to specific age groups.

Finally, subsequent research should be encouraged to investigate personality, behavior, and cognition within road users with a view to adding to the knowledge that has already been acquired.

And please forgive the 'injection of humor here and there.