Printer Friendly

Young Novice Driver Subtypes: Relationship to High-Risk Behavior, Traffic Accident Record, and Simulator Driving Performance.

Two studies were undertaken to obtain empirical support for the existence of driver subtypes in the young novice driver population. In Study 1, 198 participants (55% male) aged 16 to 19 completed an extensive self-report questionnaire. Five novice driver subtypes were identified through a cluster analysis of personality and driving-related measures. Two relatively high-risk or deviant subtypes (Clusters 1 and 5) were identified, characterized by high levels of driving-related aggression, competitive speed, driving to reduce tension, sensation seeking, assaultiveness, and hostility. The individuals in Cluster 5 also reported low levels of emotional adjustment and high levels of depression, resentfulness, and irritability. In Study 2, a subset of participants from each of the subtypes drove several scenarios in a driving simulator. The subtypes differed in their responses to an emergency situation and several potential traffic hazards. They also differed in the proficiency with which they could control their at tention among concurrent tasks in highworkload situations. Most of the significant differences were related to lower levels of driving skill among the two most deviant subtypes (Clusters 1 and 5). The potential applications of this research include the design of training programs and other countermeasures to address the young novice driver crash problem.

INTRODUCTION

The problem of young driver safety is well documented in terms of the size and nature of the problem. It is well established, for example, that young drivers play a disproportionately large role in traffic crashes. In Australia, 16- to 24-year-olds comprise about 20% of the driving population but account for around 35% of fatal and 50% of injury crashes (Macdonald, 1994). The situation in many overseas countries, including the United States (U. S. Department of Education, 1988) and Canada (Transport Canada, 1984), is similar to that in Australia: Young drivers are more likely to be injured or killed than their more experienced counterparts.

The young novice driver problem is often considered to stem from two main factors, age and inexperience. This distinction between age and inexperience corresponds to what several authors (Deery & Love, 1996a, 1996b; Elander, West, & French, 1993) have termed driving style (or behavior) and driving skill (or performance). Driving skill, which is expected to improve with practice or training, is concerned with performance limitations on aspects of the driving task, such as the time taken to respond to traffic hazards. Driving style relates to decision-making aspects of driving -- that is, the manner in which people choose to drive or driving habits that have developed over time. Such choices may include, for instance, driving speed and how close one drives to the car in front.

Research on driving skill indicates that compared with more experienced drivers, novice drivers' performance is inferior in several ways (see Mayhew & Simpson, 1995, for a review). The skills most critical to the crash problem of novice drivers include hazard perception (e.g., detecting, recognizing, and dealing with traffic hazards), attentional control (e.g., attending to the right things, in the right amounts, at the right time), time-sharing (e.g., resource management and attention switching while undertaking two or more concurrent tasks), and calibration (e.g., matching one's performance with the task demands).

There is also some evidence that younger drivers are more likely to adopt a risky driving style than are older drivers. A relationship has been found between youth and leaving shorter distances to the car in front (Evans & Wasielewski, 1983), adopting a faster driving speed (Wasielewski, 1984), accepting narrower gaps when entering traffic (Bottom & Ashworth, 1978), and running yellow lights (Koneci, Ebbesen, & Koneci, 1976).

The concept of a subgroup of "young problem drivers" is often advocated in the literature. In 1993, for example, the National Highway Traffic Safety Administration (NHTSA) prepared a report to the U.S. Congress on young driver safety. It suggested that although there is good reason to develop general countermeasures for young drivers because of the overall risk of that group, all young drivers are not equivalent, and some subgroups merit special consideration because of their increased risk. Part of the research agenda of NHTSA is to determine the effectiveness of educational and other programs that are designed for specific young driver subtypes (U. S. NHTSA, 1993).

Concerning road safety, drunk driving is the area in which population subtypes have been identified most successfully. For example, researchers have identified a number of driving-while-intoxicated (DWI) offender subtypes, defined by specific biographic and personality variables, who could be the target for prevention or intervention programs (Donovan & Marlatt, 1982; Wieczorek & Miller, 1992). The identification of driver subtypes within the population of drivers involved in multiple crashes has also been successful (Donovan, Umlauf, & Salzberg, 1988; Wilson, 1991). However, the extent to which these results generalize to the young novice driver population has not been well established.

The aim of this research was to examine the characteristics of young novice driver subtypes. Two separate but related studies were undertaken. In Study 1, five novice driver subtypes were identified through cluster analysis of personality and driving-related measures. In Study 2, the simulated driving performance of the young novice driver subtypes was examined. Risk perception and attentional control were the main aspects of performance studied, as they have been identified as sources of individual differences in road accidents and are known to be a significant problem for young novice drivers (Elander et al., 1993; Mayhew & Simpson, 1995). Methods for assessing these skills have also been developed in the driving simulator that was used in the study (Deery, Wilson, & Triggs, 1997; Regan, Deery, & Triggs, 1998a, 1998b).

STUDY 1

Method

Participants. We recruited 198 novice drivers (109 males, 89 females) aged 16 to 19 years (M = 17.9 years, SD = 0.76) in approximately equal numbers from three driver licensing offices in metropolitan Melbourne, Australia. The offices were chosen on the basis of demographic data collected by the licensing authority, indicating that the novice drivers undertaking their license test at those offices provide a reasonably representative sample of Melbourne's new driver population.

Questionnaire. Participants completed an extensive self-report questionnaire based on that developed by Donovan and Marlatt (1982). It comprised six sections. The first section included a number of items measuring demographic and background variables. The second section consisted of 105 true-false items concerned with personality functioning. Five subscales of the Buss-Durkee Hostility Inventory were used: assaultiveness, indirect hostility, verbal hostility, irritability, and resentment (Buss & Durkee, 1957). Other measures of personality functioning included assertiveness (Rathus, 1973), depression (Costello & Comrey, 1967), emotional adjustment (Howarth, 1976), locus of control (Rotter, 1966), and sensation seeking (Zuckerman, 1971).

The third section comprised 38 true-false items concerned with driving-related attitudes and behaviors. These included competitive speed, aggression, and perceived responsibility for accidents (Goldstein & Mosel, 1958), and driving inhibition (Donovan & Marlatt, 1982). The fourth section comprised 9 items measuring self-reported driving style (Deery & Love, 1996a). The fifth section assessed attitudes toward the safety of young drivers and the adequacy of novice driver training, the Victoria Police's "booze bus" and speed camera programs, and the Transport Accident Commission's (TAC) road safety television advertisements. (The TAC is the statutory body responsible for covering personal injury claims from motor vehicles in Victoria. Since 1989, it has placed "hard-hitting" advertisements on television that target specific road safety issues, such as young drivers, drunk driving, speed, and fatigue.) The final section assessed involvement in a number of high-risk behaviors, such as drug and alcohol use. Alcoho l use was measured with a modified version of Cahalan, Casin, and Crossley's (1969) quantity-frequency index.

The personality and driving-related measures used in the questionnaire have been well validated in previous research and shown to be associated with high-risk driving practices (Deery & Love, 1996a; Donovan & Marlatt, 1982; McMillen, Pang, Wells-Parker, & Anderson, 1992). The reliability of the measures has also been demonstrated during their development and psychometric assessment (e.g., Costello & Comrey, 1967; Deery & Love, 1996a; Howarth, 1976; Rathus, 1973; Zuckerman, 1971). However, to minimize the length of the questionnaire, shorter and potentially less reliable scales were used here (e.g., only some of the original subscales or items were used from some of the measures of personality functioning).

Procedure. Potential participants were approached after they had undertaken their learner permit or probationary license test and were asked to participate in the study (learner permits are available to Victorians aged 16+ years; new drivers must be at least 18 years old and hold a probationary license for their first three years of driving). Those who agreed completed the questionnaire in a quiet area of the office. The questionnaire took about 30 min to complete (range 20-60 min), and participants were offered $20 for their time.

Results

Cluster analysis was used to derive subtypes within the young novice driver sample studied. Cluster analysis uses "distance" from the mean to establish the similarity of participants' profiles on the measures under study, so that individuals with similar profiles can be grouped together. Scores derived from the driving-related attitudes and behavior measures, general personality traits, and hostility and aggression were analyzed using the squared Euclidean distance measure (the sum of the squared differences over all the measures). Standardized scores were used to avoid the problem of comparing squared Euclidean distances based on different measurement scales (Kaufman & Rousseuw, 1990).

Ward's hierarchical clustering method was used to combine cases into clusters. This technique is particularly good at identifying the number of clusters in the data (Wieczorek & Miller, 1992). With the Statistical Analysis Software (SAS) package, the cubic clustering criterion (CCC) can be produced as part of Ward's cluster analysis to help determine the number of clusters present (Sarle, 1983). Of 30 procedures for determining the number of clusters in a data set, Milligan and Cooper (1985) found the CCC to be among the most valid.

Wieczorek and Miller (1992) noted that because Ward's technique is hierarchical, it is unable to separate clusters created at previous steps and thus it cannot provide a solution with optimal between-cluster heterogeneity. It is also subject to outliers and ordering effects. K-means clustering, on the other hand, produces k number of clusters by minimizing the sum of the squared distances from the cluster means. Hence a Ward's cluster analysis with CCC output was undertaken initially to identify the number of clusters in the data. This number was then forced in the final analysis using k-means clustering.

Cluster profiles. Examination of the CCC from the Ward's cluster analysis suggested a five- or seven-cluster solution. After examining the profile of cluster means, a five-cluster solution was retained and forced in the final k-means analysis, as it provided the most meaningful distribution of cases on the variables used in the analysis.

Table 1 shows the standardized cluster means of the variables used in the k-means analysis. The pattern of means indicates that Cluster was a relatively high-risk or deviant group. The individuals in this cluster reported relatively high levels of driving-related aggression, competitive speed, and driving to reduce tension. They also reported high levels of assertiveness, sensation seeking, assaultiveness, and verbal hostility.

The individuals in Cluster 2 were the most inhibited while driving and reported an external locus of control. They were also depressed, irritable, hostile, and resentful. The individuals in Cluster 3 tended to score moderately on all of the measures. Cluster 4 was the least deviant group, reporting the lowest levels of driving-related aggression, competitive speed, and driving to reduce tension. They also reported being emotionally and behaviorally well adjusted.

Cluster 5 was the highest-risk or most deviant group. It was similar but not identical to Cluster 1. Like the individuals in Cluster 1, those in Cluster 5 reported high levels of driving-related aggression, competitive speed, driving to reduce tension, sensation seeking, and verbal hostility. Conversely, the individuals in Cluster 5 were more depressed, resentful, irritable, hostile (indirect), and emotionally maladjusted than those in Cluster 1.

Cluster validation. Blashfield (1980) recommended that cluster solutions based on personality measures be externally validated with variables not used in the initial cluster analysis. The five clusters of young novice drivers were compared on a variety of demographic, attitude, and behavioral measures external to the cluster analysis (see Table 2). In comparing the clusters on these variables, a measure of the strength of the association, [[eta].sup.2], was calculated. [[eta].sup.2] represents the proportion of variance in the dependent variable that is attributable to the effect of cluster membership, and thus it provides an indication of the practical utility of the dependent variables (Tabachnick & Fidell, 1989). In other words, the more variance explained by a variable, the greater its practical utility in differentiating the young novice driver subtypes.

The attitude and alcohol use measures were analyzed with separate one-way multivariate analyses of variance (MANOVAs). With the use of Wilks' criterion, an effect of cluster membership was found for the attitude, F(24, 580) = 1.6, p [less than] .05, [[eta].sup.2] = .057; and alcohol use, F(12, 447) = 4.8, p [less than] .001, [[eta].sup.2] = .102, measures. Analysis of variance (ANOVA) was used to examine the impact of cluster membership on each dependent variable. Driving style and age were also analyzed with separate ANOVAs.

A significant effect of cluster membership was found for driving style, F(4, 171) = 43.1, p [less than] .001, [[eta].sup.2] = .502; as well as for attitudes toward DWI enforcement, F(4, 171) = 2.4, p [less than] .05, [[eta].sup.2] = .054, the Victoria Police's "booze bus" program, F(4, 171) = 3.3, p [less than] .05, [[eta].sup.2] = .071, speed camera program, F(4, 171) = 4.4, p [less than] .01, [[eta].sup.2] = .093, and the TAC's television advertising campaign, F(4, 171) = 3.1, p [less than] .05, [[eta].sup.2] = .068. Significant effects were also found for each of the alcohol use measures: frequency, F(4, 171) = 3.6, p [less than] .01, [[eta].sup.2] = .077, quantity, F(4, 171) = 13.5, p [less than] .001, [[eta].sup.2] = .240, and total consumption per month, F(4, 171) = 6.9, p [less than] .001, [[eta].sup.2] = .139.

The [[eta].sup.2] values indicate that driving style and alcohol use measures (quantity per drinking occasion and total consumption per month) provide greater practical utility in terms of differentiating the novice driver subtypes than the attitude measures.

Specific group contrasts were undertaken using Tukey tests with alpha set at .05 (see Table 2). Of particular interest was the finding that the two most deviant groups (Clusters 1 and 5) differed from the least deviant group (Cluster 4) on several measures. Those in Clusters 1 and 5 reported a more risky driving style and greater alcohol use (frequency, quantity, and total consumption per month) than did those in Cluster 4. Those in Cluster 1 also reported less favorable attitudes toward Victoria's speed camera program than did those in Cluster 4.

Compared with the three least deviant groups (Clusters 2, 3, and 4), a higher proportion of participants in the two most deviant groups (Clusters 1 and 5) reported having smoked tobacco regularly, [[chi].sup.2] = 12.0, 4 df, p [less than] .05, currently smoking tobacco regularly, [[chi].sup.2] = 12.4, 4 df, p [less than] .05, and using prohibited drugs, [[chi].sup.2] = 13.6, 4 df p [less than] .01. A higher proportion of participants in the two most deviant groups were also male, [[chi].sup.2] 18.4, 4df, p [less than] .01.

Traffic accident record. Logit analyses were undertaken to examine whether or not traffic accident record (involvement and responsibility) varied as a function of cluster membership. A significant association was found between cluster membership and accident involvement, [[chi].sup.2] = 11.5, 4 df, p [less than] .05. Orthogonal contrasts were undertaken and revealed that the proportion of individuals who reported being involved in at least one accident was significantly higher among Clusters 2 and 5 than the other three clusters, which did not differ from each other (there was also no significant difference between Clusters 2 and 5).

Discussion

The aim of this study was to identify subtypes of young novice drivers. Cluster analysis revealed five distinct subtypes of drivers based on differential levels of driving-related attitudes and behaviors, general personality traits, and hostility and aggression. Several demographic, attitude, and behavioral measures were used to externally validate the cluster solution. Significant differences were found for a number of measures, particularly between the two most deviant subtypes (Clusters 1 and 5) and the least deviant subtype (Cluster 4). In this regard, self-reported driving style and some measures of alcohol use provide the greatest practical utility in terms of differentiating the novice driver subtypes.

It may be useful to review the results for the most deviant young novice driver subtype (Cluster 5). This subtype comprised a relatively large proportion of males. The individuals in this subtype were characterized by high levels of driving-related aggression, competitive speed, and driving to reduce tension. They also had low levels of emotional and behavioral adjustment; they reported high levels of sensation seeking, assaultiveness, indirect and verbal hostility, depression, resentfulness, and irritability. Furthermore they reported the riskiest driving style and a poor traffic accident record. They were also likely to take part in other high-risk behaviors, such as drinking large quantities of alcohol regularly, smoking tobacco, and using illicit drugs. These results are consistent with previous research indicating that high-risk driving practices among young drivers are often one part of a network of problem behaviors (Jessor, 1987; Wilson & Jonah, 1988). They are also consistent with research on DWI of fender subtypes, in which young, hostile, and heavy-drinking groups with a particularly poor driving record have been identified (Donovan & Marlatt, 1982; Wieczorek & Miller, 1992).

STUDY 2

Method

Participants. Participants from each of the clusters identified in Study 1 were randomly selected, telephoned, and asked to take part in this study, with the aim of attaining an equal number from each of the five young novice driver subtypes. A total of 54 participants agreed to take part, including 12, 12, 12, 7, and 11 from Clusters 1, 2, 3, 4, and 5, respectively.

Apparatus. A midrange driving simulator was used for the study. The simulator consisted of a full-size passenger car body (Ford Falcon) with cabin controls that operated in the same manner as a normal vehicle. The vehicle was mounted on a partial-motion platform that provided road feel but no lateral acceleration forces. It was surrounded by screens on which images were projected, providing 180[degrees] forward and 60[degrees] rear views. Other automated vehicles were included in the simulation and programmed to behave in specific ways. Two small switches were secured to the steering wheel at the 10 and 2 o'clock positions and were used for specific tasks (described later).

Simulated environments and tasks. Five simulated driving scenarios were used. The first scenario consisted of a 3-km drive along a rural road with other ambient traffic that behaved in a safe and predictable manner. The aim of this drive was to provide participants with practice in the simulator.

The second scenario involved a drive in an arterial road environment. This drive was approximately 6 kin, had a posted speed limit of 70 km/h, and included other ambient traffic that behaved in a safe and predicable manner. The aim of this drive was to obtain measures of general driving style, such as speed choice.

The third scenario was similar to that developed by Regan et al. (1998b) to examine the skill of attentional control during bigh-workload situations. It was developed within an arterial road environment along a 3-km length of road with no oncoming vehicles in the right lane.

In this scenario a driving and a numeric calculation subtask were performed concurrently. The driving subtask required participants to change and maintain their speed according to speed limit signs, which varied from 60 to 80 km/h. The speed limits were unusual in that they differed from those used on Australian roads (e.g., 63 km/h, 71 km/h). Six speed signs appeared during the drive, requiring participants to alter their speed between 8 and 14 km/h.

The numeric calculation subtask required participants to calculate the absolute difference between the values of consecutive speed signs. They were required to respond by pressing one of the two buttons on the steering wheel to indicate whether the value of the current task (the absolute difference between the current and previous speed signs) was the same as or different from that of the previous task (the absolute difference between the previous sign and the one before it). An equal number of same and different responses were required during a drive. Three versions of the drive were developed (with different speed signs) so that participants' overall performance and rate of performance improvement could be examined.

The fourth scenario was designed to examine precautionary behavior around potential traffic hazards. It comprised two drives, each approximately 6 km, along an arterial road with two lanes in each direction. Each drive contained three potentially hazardous situations corresponding to three of VicRoads' definitions for classifying accidents (DCA). (VicRoads is the agency responsible for the management of Victoria's road system. It maintains a database of police-reported traffic accidents.) The DCA system classifies accidents according to their type and nature. The DCAs were chosen because young drivers are overrepresented in crashes of these kinds, at least in part because of a cognitive skill deficit (Catchpole, 1997).

In designing the DCAs in the simulator, it was necessary to decide what is and is not safe or skillful behavior in each situation (cf. Fischer, Glaser, Laurie, & Pollatsek, 1998). One might assume, for example, that safe behavior involves moderating one's driving speed. However, none of the situations actually developed into an emergency, and thus it could be argued that slowing down represents unsafe behavior because it might cause other drivers to react in unpredictable ways. Therefore every effort was made to ensure that there were clear cues to indicate potential danger in the DCAs and that participants had sufficient time to perceive the cues and take some action to minimize the possibility of a crash. Indeed, previous research has revealed that drivers with well developed risk perception skills display a more gradual and controlled anticipatory response to these situations than do drivers with less well-developed risk perception skills (Deery et al., 1997). More experienced drivers, for example, begin to moderate their driving speed earlier, slow down in a smoother and more controlled manner, and reach a slower minimum speed than do less experienced drivers.

The first risk perception drive contained potential hazards corresponding to DCA 110, 111, and 130. For DCA 110, a car approached a four-way intersection from the left, braked late without an indicator, and stopped. There was some ambiguity both as to whether or not the car would actually stop and its intended direction at the intersection. Parked cars forced participants to travel in the left lane. DCA 111 was similar to DCA 110, except that the car approached from the right of a four-way intersection. For DCA 130, a car in front of the participant's vehicle slowed almost to a stop without an indicator or a brake light activated, as though its engine had stalled, and then turned right.

The second risk perception drive contained potential hazards corresponding to DCA 102, 121, and 134. For DCA 102, an adult pedestrian stood on the left curb and a child pedestrian on the center median. This situation was intended to raise the possibility that one of the pedestrians might cross the road in front of the participant's vehicle. For DCA 121, a vehicle approached a T-intersection from the opposite direction and slowed rapidly without an indicator activated, raising the possibility of its turning across the participant's path. For DCA 134, a car was parked in the left lane. When the participant was 50 m or so from the parked car, it began to accelerate slowly. However, a truck was also parked in the left lane 100 m ahead, raising the possibility that the other car might merge into the participant's lane to maneuver around the truck.

Participants were asked to press one of the buttons on the steering wheel as quickly as possible to indicate that they had detected any situation that they considered to be a potential hazard during the second risk perception drive. This task was included because the time taken to detect hazards has been identified as a risk factor for crash involvement and is one area in which novice drivers perform poorly (Quimby, Maycock, Carter, Dixon, & Wall, 1986). Participants were not asked to perform this task during the first risk perception drive, to avoid alerting them to the fact that it contained potential hazards.

The final scenario consisted of a drive along a 6-km arterial road. It included ambient traffic throughout, as well as one emergency in which another vehicle reversed from a driveway partly into the participant's lane. This situation required participants to perform some evasive steering, braking, or both to avoid a collision.

Procedure

Upon arrival at the laboratory, participants were given a brief overview of the operation of the simulator. They then undertook the practice drive before data collection commenced. Data collection included the drive designed to assess driving style, the three attentional control drives, the two risk perception drives, and the emergency drive. Participants were instructed to drive as they normally do in the real world.

Results

Driving style. The aim of the first drive was to assess participants' driving style. Measures of driving style included average and standard deviation (SD) of speed, which were calculated for the entire drive, excluding the first and last 500 m when participants were likely to be accelerating and decelerating, respectively. An ANOVA revealed that there were no statistically significant differences between the young novice driver subtypes for both average, F(4, 49) = 2.2, [[eta].sup.2] = .150, and standard deviation, F(4, 49) = 1.0, [[eta].sup.2] = .075, of speed.

Attentional control. For the three attentional control drives, four data collection areas were available, corresponding to the last four speed signs. Several performance measures were taken and a number of summary scores derived. These scores, which represent the mean across the four data collection areas, included SD of speed, time to reach desired speed, maximum acceleration, and the time taken to initiate a speed change. These measures were analyzed with separate Cluster x Drive ANOVAs, with drive as a repeated measure. The main effect of drive was significant for maximum acceleration, F(2, 92) = 4.2, p [less than] .05, [[eta].sup.2] = .084, time to reach desired speed, F(2, 86) = 16.1, p [less than] .01, [[eta].sup.2] = .273, and SD of speed, F(2, 90) = 3.9, p [less than] .05, [[eta].sup.2] = .080. Participants displayed a smoother and more efficient performance with practice, irrespective of cluster membership (see Table 3).

For the numeric calculation subtask, a significant main effect of clusters was found for the percentage of correct answers, F(4, 44) = 3.3, p [less than] .05, [[eta].sup.2] = .230. Scheff[acute{e}] tests revealed that those in Cluster 5 were less likely to be correct than those in Clusters 1, 2, and 3, regardless of which of the three attentional control drives they were undertaking.

Risk perception. For the risk perception and emergency drives, speed and lateral position were recorded every 10 m. To compare the performance of the subtypes around the DCAs, data were examined from 100-200 m before to 50 m after the potential point of collision or hazard. In particular, participants' speed and lateral position around each DCA were analyzed with separate Cluster x Distance ANOVAs, with distance from the potential point of collision as a within-subjects effect. When a significant Cluster x Distance interaction was found, simple effects tests were undertaken and subsequent group comparisons were made using Scheffes procedure with alpha set at .05.

For both risk perception drives, the main effect of distance was significant for all DCAs when both speed and lateral position were analyzed. These results indicate that regardless of cluster membership, participants altered their lateral position and speed throughout the DCAs. For some DCAs, however, a significant Distance x Cluster interaction was found, indicating that the driver subtypes differentially responded to the potential hazards over time. The significant interactions are considered in the following section.

DCA 111 (other vehicle approached an intersection from the right, as though it may not stop). The interaction was significant for speed, F(80, 900) = 1.5, p [less than] .01, [[eta].sup.2] = .117. The clusters were, on average, traveling at similar speeds 150 m before the potential point of collision. However, participants in Cluster 1 (high levels of driving-related aggression, competitive speed, sensation seeking, etc.) were traveling significantly faster than those in Cluster 5 (similar to Cluster 1, but also emotionally maladjusted and depressed, resentful, and irritable) near the potential point of collision (see Figure 1).

DCA 110 (other vehicle approached an intersection from the left, as though it may not stop). The interaction was significant for speed, F(100, 1125) = l.9, p [less than] .00l, [[eta].sup.2] = .l43. Simple effects tests of distance within cluster revealed that Clusters 1 (high levels of driving-related aggression, competitive speed, sensation seeking, etc.) and 4 (least deviant group) were the only groups that did not significantly alter their speed throughout the DCA (see Figure 2).

DCA 102 (adult pedestrian on left curb, child pedestrian on center median). The interaction was significant for lateral position, F(60, 660) = 2.0, p [less than] = .001, [[eta].sup.2] = .156. Simple effects tests of distance within clusters revealed that Cluster 1 (high levels of driving-related aggression, competitive speed, sensation seeking, etc.) was the only group not to significantly alter its lateral position throughout the DCA (see Figure 3).

DCA 134 (a parked vehicle in the left lane moves off as participants approach; the other vehicle's lane is blocked by a parked truck ahead). The interaction was significant for lateral position, F(100, 1100) = 1.6, p [less than] .01, [[eta].sup.2] = .130. Participants in Cluster 5 (high levels of driving-related aggression, competitive speed, sensation seeking, depression, irritability) were, on average, traveling farther to the left than those in Cluster 3 (moderate on all personality and driving-related measures) near the parked truck (see Figure 4).

Participants were instructed to press one of the buttons on the steering wheel when they detected a potential hazard in the second risk perception drive. ANOVA failed to reveal any statistically significant effects in terms of the response time to detect the potential hazards.

Emergency situation. An ANOVA revealed a significant interaction for speed, F(60, 615) = 1.6, p [less than] .01 [[eta].sup.2] = .137. Participants in Cluster 5 (high levels of driving-related aggression, competitive speed, sensation seeking, depression, resentment, and irritability), on average, were traveling significantly faster than those in Cluster 4 (the least deviant group) near the potential point of collision (see Figure 5).

The [[eta].sup.2] values indicate that performance on the attentional control task (time to reach desired speed and the proportion of correct responses) provides greater practical utility in terms of differentiating the novice driver subtypes than does performance during the risk perception and emergency drives.

Discussion. The results of this study suggest that the novice driver subtypes, identified in Study 1 on the basis of differential levels of driving-related attitudes and behaviors, general personality traits, and hostility and aggression, also differed in terms of their level of driving skill. Differences were found in the way that the novice driver subtypes responded to an emergency situation and to several potential traffic hazards in the simulator. Differences were also evident in the proficiency with which they could control their attention among concurrent tasks while driving. In fact, performance on the attentional control task (time to reach desired speed and the proportion of correct responses) provided the greatest practical utility in terms of differentiating the driving skill of the young novice driver subtypes.

Most of the significant differences in performance were related to lower levels of driving skill among the two highest-risk subtypes (Clusters 1 and 5). Participants in Cluster 1 displayed a relative lack of caution near some potential hazards. For example, they were traveling the fastest throughout DCA 111 and did not alter their driving speed throughout DCA 110. This relative lack of caution around some potential hazards may enhance this group's risk in similar real-world situations.

Those in Cluster 5 also appear to have difficulties with particular areas of driving skill. Unlike Cluster 1, the individuals in Cluster 5 showed caution around all DCAs in terms of moderating their driving speed. However, they gave themselves a relatively small lateral space near the parked truck in DCA 134. This DCA was particularly notable for the need to anticipate or predict a sequence of gradually unfolding events, suggesting that individuals in Cluster 5 have some difficulties with this area of driving skill.

The individuals in Cluster 5 also displayed a relative lack of caution in terms of driving speed in the emergency situation. Those in Cluster 5 were also the least proficient in the numeric calculation subtask during the attentional control drives, suggesting that they have difficulty controlling their attention between competing tasks in a driving situation. This is likely to cause problems in high-workload situations in the real world (e.g., driving in heavy traffic), in which the driver needs to control his or her attention between competing tasks, such as maintaining a safe headway while scanning for and responding to traffic hazards.

In Study 1, the individuals in Cluster 2 reported being involved in more traffic accidents than did those in Clusters 1, 3, and 4. The individuals in Cluster 2 did not, however, report particularly high levels of other deviant behaviors, such as drug and alcohol use. Similarly, the individuals in Cluster 2 displayed relatively high levels of caution around most of the potential hazards, especially in terms of moderating their driving speed. Hence the reason for the relatively high accident rate of Cluster 2 is not entirely clear. One explanation is that the individuals in Cluster 2 are characterized by high levels of resentment, depression, and irritability, and that these characteristics manifest themselves in aspects of their driving practices not examined here. Additional research is needed to address this issue.

GENERAL DISCUSSION

The results of this research indicate that young novice drivers are not a homogeneous group. Five distinct subtypes of young novice drivers were identified that differed in their driving-related attitudes and behaviors, general personality traits, hostility, and aggression. They also differed in terms of their self-reported driving style, traffic accident record, participation in high-risk behaviors, and level of attentional control and risk perception skill. These results are consistent with the concept of a group or groups of young problem drivers in the young novice driver population (Macdonald, 1994; NHTSA, 1993).

It was noted earlier that a recent NHTSA report to the U.S. Congress suggested that all young drivers are not equivalent and that some subgroups merit special consideration because of their increased risk (NHTSA, 1993). Part of NHTSA's research agenda is to determine the effectiveness of specific countermeasures for different novice driver subtypes. The results of the current research may have implications for the design of such countermeasures. Several training approaches, for example, have recently shown a great deal of promise and might be appropriate for novice drivers generally and certain subtypes particularly (e.g., Clusters 1 and 5). Regan et al. (1998a) found that PC-based mediated instruction in risk perception skills, such as scanning the road environment, keeping ahead of the situation, and choosing the safest option, combined with the opportunity to practice and obtain feedback on those skills, enhanced the precautionary behavior of novice drivers near traffic hazards in a driving simulator. McK enna and Crick (1997) showed that the risk perception skills of novice drivers were enhanced with training that included predicting unfolding events in video sequences of traffic hazards.

The novice drivers in Cluster 5 might also benefit in particular from training to enhance the skill of attentional control. Regan et al. (1998b) showed that this skill can be enhanced with variable priority training (VPT). With VPT, participants perform two or more tasks concurrently. Across training trials, they are instructed to systematically vary the relative amount of attention that they give to each task. Regan et al. (1998b) found that 26 trials of VPT in a driving simulator enhanced both novice drivers' attentional-control skill and their ability to detect and respond to traffic hazards.

It should be emphasized that the training methods described previously have been used only in laboratory settings, and thus real-world evaluations of their effectiveness are needed. If they are shown to be effective, it might be possible to offer all novice drivers a training product, such as a CD-ROM (e.g., Blank & McCord, 1988; Regan et al., 1998b), that integrates VPT, mediated instruction in risk perception, and other appropriate training content. The product might be offered, for example, as part of the licensing process. When necessary, particular novice driver subtypes could be offered additional training that is tailored to their specific skill or motivational needs. This might enhance the benefit provided by a general driver training strategy that uses an integrated training product, such as a CD-ROM, for all novice drivers.

Requiring all novice drivers to undertake an assessment for the purpose of receiving tailored training would probably not be cost-beneficial. It would also be difficult to justify: One cannot punish people for their characteristics or for something that they have not yet done, such as cause harm on the roads. In other words, undertaking a tailored training program would probably need to occur either on a voluntary basis, as a punitive measure for those drivers with a poor driving record, or both. In terms of voluntary participation, programs could be developed that offer incentives for undertaking specific training. These could be consistent with previously attempted approaches in Australia and elsewhere, such as reduced insurance premiums for those drivers who undertake the training (e.g., Malfetti, 1993).

Malfetti (1993) suggested that automobile insurance companies represent a powerful source for improving the safety of young novice drivers. He proposed a pilot program be undertaken and evaluated consisting of four main components: (a) those young drivers participating would be required to pass a rigorous driver education course, such as a tailored driving skill course; (b) the insurance company and the young driver would sign a contract stipulating what the young driver would do in terms of safety, such as not willfully commit any moving violations and not drive while drunk; (c) the company would charge the young driver the regular beginner's insurance premium (however, after a specified period, the company would refund the difference between the beginner's premium and that charged to an older driver, assuming that the young driver had met the terms of the contract); and (d) if the young driver violated the contract, he or she would forfeit the refund and return to the company's regular insurance rates for beginner drivers.

Discussions we have had with one major insurance company in Australia indicate that the company is considering a trial of this type of program, including the identification of novice driver subtypes. Such a program would seem to offer several benefits to young drivers and act as a motivator for safety. For example, it clearly illustrates the rewards and penalties for both safe and unsafe driving. It also makes insurance policies more personal, which is likely to foster greater commitment in terms of safety. Indeed the psychological literature (e.g., Rotter, 1966) suggests that when people perceive that they have control over events and outcomes in their lives, they are likely to behave favorably toward them.

Several potential limitations of the research questionnaire used in Study 1 should be noted (Donovan & Marlatt, 1982). As mentioned earlier, by including a range of measures, shorter and potentially less reliable scales were used. A self-reported traffic accident record was used, in part, to externally validate the cluster solution. To avoid possible inaccuracies with these data, longitudinal research could be undertaken in which participants' official driving records are followed up some years later. Exposure data could also be collected to adjust for any differences in the extent to which the driver subtypes have been exposed to potential crashes.

It was noted earlier that the pattern of results found for some driver subtypes was not entirely consistent across the studies. For example, those in Cluster 2 reported a relatively high level of traffic accident involvement in Study 1, yet, contrary to expectation, they also displayed relatively high levels of caution near most of the potential traffic hazards in Study 2. A related issue is that the potential unreliable measurement of personality and attitude variables means that any effort to identify high-risk driver subtypes for practical purposes, such as training, could lead to a number of nonproblem drivers being incorrectly identified (a high false alarm rate). Additional research is needed to further validate the novice driver subtypes identified here and to understand more fully the differences among them. It may be, for example, that the subtypes differ in terms of aspects of driving style that contribute to the young novice drivers' crash problem but that were not assessed here, such as the dista nce left to the car in front (Evans & Wasielewski, 1983), gap acceptance when entering traffic (Bottom & Ashworth, 1978), and running yellow lights (Koneci et al., 1976).

CONCLUSION

The results of this research indicate that young novice drivers are not a homogeneous group. Five distinct subtypes of young novice drivers were identified based on differential levels of general personality and specific driving-related attitude and behavior measures. Subsequent analyses revealed that the subtypes differed in terms of several demographic, attitude, and behavioral variables, including self-reported driving style and traffic accident involvement. They also differed in terms of their level of risk perception and attentional control skill. These results would seem to have implications for driver training programs and other countermeasures that are designed to address the young novice driver crash problem.

ACKNOWLEDGMENTS

This research was supported by funding from the NRMA Ltd. We would like to thank the managers of VicRoads' Registration and Licensing offices for allowing data collection to be undertaken from their offices; Naomi Kowadlo, Tiffany Westphal-Wedding, Naomi Wilson, and Nebojsa Tomasevic for their assistance with data collection; Elizabeth Wells-Parker for providing part of the research questionnaire; and Sanjeev Narayan for his advice on analysis.

Hamish A. Decry is a research fellow at the Monash University Accident Research Centre. He received a Ph.D. in psychology from La Trobe University in Melbourne, Australia in 1996.

Brian N. Fildes holds the RACV Chair in Road Safety at the Monash University Accident Research Centre. He obtained his Ph.D. in psychology from Monash University in Melbourne, Australia in 1987.

REFERENCES

Blank, D., & McCord, R. (1988). Design and presentation of a CD-ROM driving program (Driver-ZED). In Proceedings of the Human Factors and Ergonomics Society 42nd Annual Meeting (pp. 1401-1402). Santa Monica, CA: Human Factors and Ergonomics Society.

Blashfield, R. K. (1980). Propositions regarding the use of cluster analysis in clinical research. Journal of Consulting and Clinical Psychology, 48, 456-459.

Bottom, C. G., & Ashworth, R. (1978). Factors affecting the variability of driver gap acceptance behavior. Ergonomics, 21, 721-734.

Buss, A. H., & Durkee, A. (1957). An inventory for assessing different kinds of hostility. Journal of Consulting and Clinical Psychology, 21, 343-349.

Cahalan, D., Casin, I. H., & Crossley, H. M. (1969). American drinking practices. A national study of drinking behavior and attitudes (Monograph No. 6). New Brunswick, NJ: Rutgers Center of Alcohol Studies.

Catchpole, J. (1997). An investigation of factors contributing to the overrepresentation of young drivers in traffic accidents. Unpublished master's thesis, La Trobe University, Melbourne, Australia.

Costello, C. G., & Comrey, A. L. (1967). Scales for measuring depression and anxiety. Journal of Psychology, 66, 303-313.

Deery, H. A., & Love, A. W. (1996a). The driving expectancy questionnaire: Development, psychometric assessment and predictive utility among young drunk-drivers. Journal of Studies on Alcohol, 57, 193-202.

Deery, H. A., & Love, A. W. (1996b). The effect of a moderate dose of alcohol on the traffic hazard perception profile of young drivers. Addiction, 91, 815-827.

Deery, H. A., Wilson, N., & Triggs, T. J. (1997). Young drivers' perception of risk in traffic hazards: Implications for simulator-based training. Australian Journal of Psychology, 49 (Suppl. 1), 8.

Donovan, D. M., & Marlatt, G. A. (1982). Personality subtypes among driving-while-intoxicated offenders: Relationship to drinking behavior and driving risk. Journal of Consulting and Clinical Psychology, 50, 241-249.

Donovan, D. M., Umlauf, R. L., & Salzberg, P. M. (1988). Derivation of personality subtypes among high-risk drivers. Alcohol, Drugs end Driving, 4, 233-234.

Elander, J., West, R., & French, D. (1993). Behavioral correlates of individual differences in road-traffic crash risk: An examination of methods and findings. Psychological Bulletin 113, 279-294.

Evans, L., & Wasielewski, p. (1983). Risky driving related to driver and vehicle characteristics. Accident Analysis and Prevention, 15, 121-136.

Fischer, D. L., Glaser, R., Laurie, N. E., & Pollatsek. A. (1998). Evaluation of PC-based younger driver training program: Use of a driving simulator. In Proceedings of the Human Factors and Ergonomics Society 42nd Annual Meeting (pp. 1403-1407). Santa Monica, CA: Human Factors and Ergonomics Society.

Goldstein, L. G., & Mosel, J. N. (1958). A factor study of drivers' attitudes, with further study on driver aggression. Highway Research Board Bulletin, 172, 9-29.

Howarth, E. A. (1976). A psychometric investigation of Eysenck's Personality Inventory. Journal of Personality Assessment, 40, 173-185.

Jessor, R. (1987). Risky driving and adolescent problem behavior: Theoretical and empirical linkages. In T. Benjamin (Ed.), International Congress and Symposium Series: No. 116. Young drivers impaired by alcohol and other drugs (pp. 97-110). New York: Royal Society of Medicine Services.

Kaufman, L., & Rousseuw, P.J. (1990). Finding groups in data: An introduction to cluster analysis. New York: Wiley.

Koneci, C., Ebbesen, E. B., & Koneci, D. K. (1976). Decision processes and risk taking in traffic: Driver response to the onset of yellow light. Journal of Applied Psychology, 6, 359-367.

Macdonald, W. (1994). Young driver research program: A review of information on young driver crashes (Report CR 128). Canberra, Australia: Federal Office of Road Safety.

Malfetti, J. L. (1993). A study of young driver insurance records: Opportunities for automobile insurers. Washington, DC: American Automobile Association Foundation for Traffic Safety.

Mayhew, D. R., & Simpson, H. M. (1995). The role of driving experience: implications for training and licensing of new drivers. Toronto, Canada: Insurance Bureau of Canada.

McKenna, F. P., & Crick, J. (1997). Hazard perception in drivers: A methodology for testing and training (No. 297). Crowthorn, England: Transport Research Laboratory.

McMillen, D. L., Pang, M. G., Wells-Parker, E., & Anderson, B. J. (1992). Alcohol, personality, and high risk driving: A comparison of young, drinking driver groups. Addictive Behaviors, 17, 525-532.

Milligan, G. W., & Cooper, M. C. (1985). An examination of procedures for determining the number of clusters in a data set. Psychometrika, 50, 159-179.

Quimby, A. R., Maycock, G., Carter, I. D., Dixon, R., & Wall, J. G. (1986). Perceptual abilities of accident involved drivers (Report 27). Crowthome, England: Transport Research Laboratory.

Rathus, S. A. (1973). A 30-item schedule for assessing assertive behaviour. Behaviour Therapy, 4, 398-406.

Regan, M., Deery, H. A., & Triggs, T. J. (1998a). A technique for enhancing risk perception in novice car drivers. In Proceedings of the Road Safety Research, Policing, Education Conference (pp. 51-56). Wellington, New Zealand: Department of Transport.

Regan, M., Deery, H. A., & Trigga, T. J. (1998b). Training for attentional control in novice car drivers: A simulator study. In Proceedings of the Human Factors and Ergonomics Society 42nd Annual Meeting (pp. 1408-1413). Santa Monica, CA: Human Factors and Ergonomics Society.

Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 80, 1.

Sarle, W. S. (1983). Cubic clustering criterion (Tech. Report A-108). Cary, NC: SAS Institute.

Tabachnick, R. G., & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). New York: HarperCollins.

Transport Canada. (1984). Canadian motor vehicle traffic accident statistics (Report TP3322). Ottawa, Canada: Road Safety Directorate.

U.S. Department of Education. (1988). Youth indicators 1988: Trends in the well-being of American youth (DE Publication No. 065-000-00347-3). Washington, DC: U.S. Government Printing Office.

U.S. National Highway Traffic Safety Administration. (1993). Addressing the safety issues related to younger and older drivers: A report to Congress, January 19, 1993, On the research agenda of the National Highway Traffic Safety Administration (DOT-HS-808-161). Washington, DC: Author.

Wasielewski, P. (1984). Speed ass measure of driver risk: Observed speed versus driver and vehicle characteristics. Accident Analysis and Prevention, 16, 89-103.

Wieczorek, W. F., & Miller, B. A. (1992). Preliminary typology designed for treatment matching of driving-while-intoxicated offenders. Journal of Consulting and Clinical Psychology, 60, 757-765.

Wilson, R. J. (1991). Subtypes of DWIs and high-risk drivers: Implications for differential intervention. Alcohol, Drugs and Driving, 7, 1-12.

Wilson, R. J., & Jonah, B. A. (1988). The application of problem behavior theory to the understanding of risky driving. Alcohol, Drugs and Driving, 4, 173-191.

Zuckerman, M. (1971). Dimensions of sensation seeking. Journal of Consulting and Clinical Psychology. 36, 45-52.
                  Mean Standardized Score on the Measures
               Defining the Young Novice Driver Clusters [a]
                        Novice Driver Cluster
                                  1              2        3        4
Variable                      (n = 27)        (n = 19) (n = 50) (n = 59)
Driving-related
   Aggression                   0.47           -0.01    -0.31    -0.56
   Competitive speed            0.93           -0.17    -0.24    -0.65
   Externality                  0.30            0.64    -0.32     0.04
   Internality                  0.07           -0.08    -0.11     0.17
   Inhibition                  -0.32            0.45     0.27     0.16
   Tension reduction            0.28           -0.03     0.09    -0.32
Personality
   Assertiveness                0.76           -0.45    -0.50     0.38
   Depression                  -0.54            1.39     0.06    -0.51
   Emotional adjustment        -0.66            1.15    -0.01    -0.45
   Internality                 -0.05            0.25     0.38    -0.55
   Externality                  0.15            0.16     0.51    -0.58
   Sensation seeking            0.67           -0.12     0.15    -0.67
Hostility & aggression
   Assaultiveness               0.90            0.36    -0.09    -0.32
   Indirect hostility           0.22            0.58    -0.24    -0.38
   Verbal hostility             0.87            0.39    -0.28    -0.57
   Irritability                 0.18            0.90    -0.18    -0.65
   Resentment                  -0.12            1.07    -0.04    -0.66
                           5
Variable                (n = 21)
Driving-related
   Aggression             1.72
   Competitive speed      1.35
   Externality           -0.32
   Internality           -0.24
   Inhibition            -0.09
   Tension reduction      0.33
Personality
   Assertiveness         -0.46
   Depression             0.72
   Emotional adjustment   1.08
   Internality            0.47
   Externality            0.05
   Sensation seeking      0.79
Hostility & aggression
   Assaultiveness         0.33
   Indirect hostility     0.85
   Verbal hostility       0.78
   Irritability           1.22
   Resentment             1.16
(a.)For each measure, higher scores reflect greater levels of
the trait or behavior, except for emotional adjustment, for
which higher scores reflect lower levels of adjustment.
                Descriptive Statistics (Means with Standard
                 Deviations in Brackets, or Proportion of
               Participants) for the Driving Style, Driving
                Record Attitudes, and Demographic Measures
                                   Novice Driver Cluster
                                           1                     2
Variable                                (n = 27)              (n = 19)
Driving style [a]                       2.5 (0.6) [d,e,f,g]  1.8 (0.3) [c,g]
Driving record
   Involved in an accident (%)            3.7                  15.8
   Responsible for an accident (%)        3.7                  10.5
   Moving violations                    0.2 (0.6)            0.0 (0.0)
Attitudes [b]
   Booze buses                          4.4 (0.7)            4.2 (1.4) [f]
   Drunk driving                        4.3 (0.8)            4.1 (1.2) [f]
   Speed cameras                        3.7 (1.4) [e,f]      4.1 (1.4)
   TAC advertisements                   4.4 (0.7)            4.1 (1.4) [f,g]
   Driver training                      4.0 (1.0)            3.7 (1.2)
   Young driver safety                  4.2 (0.9)            4.2 (1.2)
Alcohol use
   Drinking occasions per month         5.6 (6.4) [f]        3.1 (5.1)
   Drinks per occasion                  6.6 (2.9) [d,e,f]    2.8 (3.2) [c]
   Total quantity per month            43.6 (62.7) [e,f]    19.4 (36.4)
Other drug use
   Use prohibited drugs                   37%                  21%
   Ever smoked tobacco regularly          52%                  37%
   Smoke tobacco regularly now            41%                  32%
Demographic variables
   Age                                 17.7 (0.9)           17.9 (0.8)
   Sex (% male)                        82%                  47%
                                        3                 4
Variable                             (n = 50)          (n = 59)
Driving style [a]                   1.9 (0.4) [c,f,g]  1.6 (0.4) [c,e,g]
Driving record
   Involved in an accident (%)      2.0                  3.4
   Responsible for an accident (%)  2.0                  1.7
   Moving violations                0.0 (0.1)          0.0 (0.0)
Attitudes [b]
   Booze buses                      4.6 (0.9)          4.8 (0.4) [d]
   Drunk driving                    4.5 (0.8)          4.7 (0.6) [d]
   Speed cameras                    4.4 (0.9) [c]      4.6 (0.6) [c]
   TAC advertisements               4.6 (0.9)          4.7 (0.7) [d]
   Driver training                  4.0 (1.1)          3.8 (0.8)
   Young driver safety              4.3 (0.9)          4.4 (0.9)
Alcohol use
   Drinking occasions per month     3.6 (4.4)          2.2 (3.2) [c,g]
   Drinks per occasion              3.9 (3.1) [c,f]    2.0 (2.3) [c,e,g]
   Total quantity per month        17.9 (22.8) [c,g]   8.1 (13.3) [c,g]
Other drug use
   Use prohibited drugs               22%                   7%
   Ever smoked tobacco regularly      40%                  20%
   Smoke tobacco regularly now        20%                  14%
Demographic variables
   Age                             18.1 (0.7)         17.9 (0.7)
   Sex (% male)                    50%                40%
                                       5                       F-or
Variable                            (n = 21)            [[chi].sup.2]-value
Driving style [a]                   3.0 (0.7) [c,d,e,f]        43.1 [**]
Driving record
   Involved in an accident (%)        19.0                      n/a
   Responsible for an accident (%)     9.5                      n/a
   Moving violations                0.9 (3.3)                   2.5 [*]
Attitudes [b]
   Booze buses                      4.8 (0.4)                   3.3 [*]
   Drunk driving                    4.4 (0.9)                   2.4 [*]
   Speed cameras                    4.4 (1.0)                   4.4 [*]
   TAC advertisements               4.9 (0.3) [d]               3.1 [*]
   Driver training                  3.8 (1.1)                   0.5
   Young driver safety              4.1 (1.0)                   0.5
Alcohol use
   Drinking occasions per month     5.5 (5.4) [f]               3.6 [**]
   Drinks per occasion              5.3 (3.6) [f]              13.5 [**]
   Total quantity per month        42.8 (49.8) [e,f]            6.9 [**]
Other drug use
   Use prohibited drugs                33%                     13.6 [**]
   Ever smoked tobacco regularly       52%                     12.0 [*]
   Smoke tobacco regularly now         43%                     12.4 [*]
Demographic variables
   Age                             18.1 (0.4)                   1.0
   Sex (% male)                    76%                         18.4 [*]


(a.)Higher scores indicate a more risky driving style.

(b.)Scores range from 1 (strongly disagree) to 5 (strongly agree).

(c.)Significant difference with Cluster 1.

(d.)Significant difference with Cluster 2.

(e.)Significant difference with Cluster 3.

(f.)Significant difference with Cluster 4.

(g.)Significant difference with Cluster 5.

(*.)p[less than].05,

(**.)p[less than].01
                 Mean Performance Measures (with Standard
                   Deviation in Parentheses) during the
                     Three Attentional Control Drives
                                     Young Novice Drive Cluster
                                                 1                   2
Attentional Control Drive 1
 Driving performance
  Maximum acceleration [m/[s.sup.2]]        1.18 (0.40)         1.03 (0.29)
  Time to desired speed [s]                 8.74 (2.16)         9.66 (2.79)
  Time to initiate speed change [s]         3.98 (1.15)         3.87 (2.54)
  Standard deviation of speed               4.60 (1.24)         5.00 (1.13)
 Numeric calculation task
  Correct responses (%)                     86.6 (17.2)         83.3 (13.4)
  Response time [s]                         4.70 (1.98)         4.12 (1.71)
Attentional Control Drive 2
 Driving performance
  Maximum acceleration [m/[s.sup.2]]        1.06 (0.33)         0.87 (0.15)
  Time to desired speed [s]                 8.01 (1.70)         7.23 (2.67)
  Time to initiate speed change [s]         3.74 (1.52)         3.96 (2.29)
  Standard deviation of speed               4.41 (0.75)         4.30 (1.03)
 Numeric calculation task
  Correct responses (%)                     93.1 (11.3)         95.5 (10.1)
  Response time [s]                         3.88 (1.90)         4.19 (1.08)
Attentional Control Drive 3
 Driving performance
  Maximum acceleration [m/[s.sup.2]]        1.03 (0.34)         0.99 (0.29)
  Time to desired speed [s]                 7.08 (2.44)         6.98 (3.50)
  Time to initiate speed change [s]         4.29 (1.65)         3.52 (1.97)
  Standard deviation of speed               4.73 (1.01)         5.20 (1.35)
 Numeric calculation task
  Correct responses (%)                     81.8 (22.2)         88.6 (13.1)
  Response time [s]                         5.14 (2.38)         5.57 (2.32)
                                          3           4           5
Attentional Control Drive 1
 Driving performance
  Maximum acceleration [m/[s.sup.2]] 0.96 (0.23) 1.10 (0.19) 1.05 (0.24)
  Time to desired speed [s]          9.76 (1.93) 9.76 (2.26) 8.84 (1.95)
  Time to initiate speed change [s]  4.88 (2.50) 3.26 (1.41) 3.21 (1.27)
  Standard deviation of speed        5.00 (1.49) 5.98 (1.62) 5.28 (1.16)
 Numeric calculation task
  Correct responses (%)              89.6 (19.8) 75.0 (31.6) 74.0 (21.8)
  Response time [s]                  3.76 (1.24) 4.87 (2.38) 6.01 (2.02)
Attentional Control Drive 2
 Driving performance
  Maximum acceleration [m/[s.sup.2]] 1.02 (0.21) 1.19 (0.62) 1.05 (0.21)
  Time to desired speed [s]          8.06 (2.55) 6.89 (2.38) 7.90 (2.72)
  Time to initiate speed change [s]  3.89 (1.70) 3.29 (1.23) 3.62 (1.21)
  Standard deviation of speed        4.26 (0.69) 4.72 (1.02) 5.11 (1.33)
 Numeric calculation task
  Correct responses (%)              77.1 (22.5) 91.7 (27.8) 66.6 (28.0)
  Response time [s]                  3.36 (1.02) 4.34 (1.52) 4.88 (1.44)
Attentional Control Drive 3
 Driving performance
  Maximum acceleration [m/[s.sup.2]] 0.91 (0.37) 0.99 (0.24) 0.79 (0.17)
  Time to desired speed [s]          6.82 (2.79) 6.55 (2.67) 6.54 (1.77)
  Time to initiate speed change [s]  3.79 (1.61) 3.04 (1.75) 3.65 (2.39)
  Standard deviation of speed        4.41 (1.04) 4.91 (1.29) 4.44 (1.27)
 Numeric calculation task
  Correct responses (%)              93.8 (11.3) 70.8 (20.4) 74.1 (28.1)
  Response time [s]                  4.24 (1.33) 5.70 (1.94) 7.22 (1.79)
COPYRIGHT 1999 Human Factors and Ergonomics Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1999 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Deery, Hamish A.; Fildes, Brian N.
Publication:Human Factors
Article Type:Statistical Data Included
Geographic Code:8AUST
Date:Dec 1, 1999
Words:9719
Previous Article:The Design of a Visual Display for the Presentation of Statistical Quality Control Information to Operators on the Plant Floor.
Next Article:Work Organization, Job Stress, and Work-Related Musculoskeletal Disorders.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters