In the driver's seat II: Beyond the early driving years

Research report No. 17

4. Consistency of driving behaviour over time

Download (PDF 358KB)

This next chapter focuses on an important issue - the stability of driving behaviour over the early years of young people's driving careers. Do young people maintain the same tendencies to drive safely or riskily, or do they change in style as they gain maturity and more years of driving experience?

The scant research available suggests that risky driving tends to decrease as young people make the transition from adolescence to adulthood (Bingham, Shope, Zakrajsek, & Raghunathan, 2008; Jessor, Turbin, & Costa, 1997). For instance, Jessor and colleagues noted a decline in levels of risky driving among young drivers between the ages of 18 and 25 years. Similarly, Begg and Langley (2001) reported a decrease in the prevalence of almost all types of risky driving between the ages of 21 and 26, although this decrease was only apparent among males.

Explanations of this trend point to the developmental changes that typically occur during early adulthood (Bingham et al., 2008; Jessor et al., 1997). The transition from adolescence to adulthood is generally a time of considerable change in multiple aspects of life. For many, this period witnesses the transition from school to employment and/or further education, a move out of the parental home, as well as the development of intimate relationships, marriage and parenthood. Some argue that adolescents engage in problem behaviours such as risky driving to gain adult-like status (Jessor et al., 1997). The taking on of adult occupational, marital and parental roles has been linked to decreased involvement in problem behaviours (Jessor et al., 1997). For example, Bingham and colleagues found that both the degree and frequency of risky driving decreased with increasing levels of psychosocial maturity (as reflected by the adoption of adult roles and completion of developmental tasks associated with becoming an adult). It has also been suggested that a decrease in immature characteristics such as sensation-seeking and self-centredness that may underlie risk-taking contributes to the decline in problem behaviour often seen during early adulthood (Arnett, 2001).

Another explanation points to the benefits that increased experience on the road brings to driving skills. Some forms of risky driving, such as unsafe lane changes or following too closely, may not be intentional, but instead reflect inexperience or inadequate driving skills. It is possible that young people drive in a less risky manner as they grow older because their driving skills and appreciation of the dangers of risky driving improve with experience (Bingham et al., 2008). On the other hand, greater confidence could also lead to an increase over time in risky driving among young people who were initially safe drivers.

Finally, recent research suggests that brain development continues through adolescence and early adulthood, with capacities associated with impulse control, decision-making, planning and judgement being the last to evolve (Giedd, 2004; Sowell, Thompson, Holmes, Jernigan, & Toga, 1999). Hence, it is possible that cognitive maturation may also account for the decline in risk-taking observed over this period.

Information collected at 19-20 and 23-24 years of age on young people's engagement in a range of driving practices is used here to investigate the stability of driving tendencies over this age period. Two main aspects are explored: (a) the stability of particular types of risky driving, and (b) the continuity of young people's overall propensity to engage in risky driving between 19-20 and 23-24 years.

Four main questions are addressed:

  • Did rates of risky driving increase, decrease or remain the same across the ATP sample between the ages of 19-20 and 23-24?
  • Were the same individuals engaging in specific types of risky driving at 19-20 and 23-24 years (e.g., speeding, failure to wear seatbelts, driving under the influence of alcohol or other drugs)?
  • How similar were the driving profiles of 19-20 and 23-24 year olds who were identified as high-level risky drivers?
  • Did young people identified as low-, moderate- and high-level risky drivers at 19-20 years show similar driving tendencies at 23-24 years?

4.1 Stability of different types of risky driving among the sample

We were first interested in looking at whether or not there had been a change in the percentage of young people who engaged in different types of risky driving at 19-20 and 23-24 years in this sample of young Australians. As a wider range of risky driving types was measured at age 23-24 than at 19-20, only those behaviours that were assessed at both time points were included in these comparisons.

Figure 13 shows the rates of occurrence of each form of risky driving at 19-20 and 23-24 years for this sample (risky driving was defined as the percentage who had engaged in the risky driving practice on at least one of their past ten trips).

Figure 13 graph of engagement in risky driving in previous 10 trips, described in text.

Figure 13. Engagement in risky driving during previous ten trips, at 19-20 and 23-24 years

As this figure shows, the percentage of young people who engaged in moderate-level speeding (between 11 and 25 km/h over the limit) decreased somewhat between the ages of 19-20 and 23-24 years (from 49% to 45%), as did the percentage who reported high-level speeding (driving at more than 25 km/h above the limit; this decreased from 19% to 14%). Driving without a seatbelt (or helmet) for all of a trip, while uncommon in this sample, also dropped, with 9% reporting this behaviour at age 19-20, compared with 5% at 23-24 years.

Rates of most other risky driving behaviours (e.g., speeding up to 10 km/h over the limit, failure to wear a seatbelt (or helmet) for part of a recent trip, driving when very tired, and driving when affected by an illegal drug) remained fairly constant among the sample over this time period. However, the percentage who drove when affected by alcohol increased markedly from 14% at 19-20 years to 23% at 23-24 years.

As well as looking at whether there had been a change in the percentage who had engaged in differing types of risky driving at ages 19-20 and 23-24, we were also interested in determining whether there had been a change in how frequently risky driving practices occurred. Figure 14 shows the average number of trips on which this sample of young people had engaged in each type of risky driving at 19-20 and 23-24 years.

Figure 14 graph of frequency of risky driving behaviours, described in text.

Figure 14. Frequency of risky driving behaviours, at 19-20 and 23-24 years

As illustrated in Figure 14, young people tended to engage in speeding slightly less often at age 23-24 than they did at age 19-20. For instance, the average number of trips during which young people engaged in low-level speeding decreased from 4.1 (or 41%) of their past ten trips at age 19-20 to 3.7 (or 37%) of their past ten trips at age 23-24; and moderate-level speeding decreased from 1.6 (16%) of trips to 1.3 (13%) of trips. The frequency of all other types of risky driving was similar at the two ages (driving when very tired, not wearing a seatbelt or helmet for part or all of a trip, or driving when affected by alcohol or illegal drugs).

Summary: Rates of risky driving over time

The stability of different forms of risky driving over the early adult years was examined by looking at whether there had been a change in (a) the percentage of people who engaged in each form of risky driving at 19-20 and 23-24 years, and (b) how frequently they had done so.

These comparisons revealed a modest decrease in speeding over this time period, with fewer exceeding the speed limit (particularly by moderate or high margins) at age 23-24, and this behaviour occurring on fewer trips than at 19-20 years.

Rates of driving when affected by alcohol, on the other hand, increased substantially over this same period, with more young people engaging in this behaviour at 23-24 years than at 19-20.

There was little change in rates of other risky driving behaviours between 19-20 and 23-24 years, with the exception of driving without a seatbelt or helmet. While relatively uncommon at both ages, fewer young people had driven without a seatbelt or helmet for an entire trip in their mid-20s than at 19-20.

4.2 Stability of different types of risky driving among individuals

We also sought to investigate (a) whether the same individuals who engaged in risky driving practices at 19-20 years were still doing so at 23-24 years, and (b) how often individuals who engaged in risky driving at 19-20 did so at 23-24 years.

To answer these questions, young people were divided into groups according to whether or not they had engaged in a particular form of risky driving in their last 10 trips at 19-20 years of age. Their risky driving profile at 23-24 years on five key indicators was then examined: high-level speeding, driving without a seatbelt or helmet, driving when fatigued on multiple occasions, driving when affected by alcohol, and driving when affected by an illegal drug. The criteria used to form the five separate groups and the numbers in each group are summarised in Table 13.

Table 13. Risky driving sub-groups, at 19-20 years
  Sub-group Description n %
High-level speeding (> 25 km/h over the limit) No Did not exceed speed limit by more than 25 km/h 871 80.8
Yes Exceeded speed limit by more than 25 km/h 207 19.2
Driving without a seatbelt or helmet for entire trip No Wore a seatbelt/helmet for entire trip 987 91.4
Yes Did not wear seatbelt/helmet for entire trip 93 8.6
Driving when fatigued on multiple occasions a No Drove when very tired on 1 or fewer recent trips 673 62.3
Yes Drove when very tired on 2 or more recent trips 407 37.7
Driving when affected by alcohol No Did not drive when affected by alcohol 928 85.9
Yes Drove when affected by alcohol 152 14.1
Driving when affected by an illegal drug No Did not drive when affected by an illegal drug 981 90.8
Yes Drove when affected by an illegal drug 99 9.2

Note: a Due to the high prevalence of fatigued driving in the sample, this criterion differentiates those who drove when very tired on multiple occasions from those who only once or never drove when very tired.

Most of the young people who did not engage in a particular risky driving practice at 19-20 years also did not engage in this behaviour at age 23-24 (Figure 15). This was particularly true for driving when affected by an illegal drug, high-level speeding and driving without a seatbelt or helmet, with 92% to 96% of young adults who refrained from these behaviours at 19-20 years continuing to do so at 23-24. However, 30% of those who had not driven when very tired at 19-20 did so at 23-24 years, and 18% of those who had not driven when affected by alcohol at the younger age did so at the older age.

Figure 15 graph of engagement in risky driving behaviours, described in text.

Figure 15. Engagement in risky driving behaviours at 23-24 years, by whether engaged in by the same individuals at 19-20 years

There was much less stability in the later driving behaviour of those who had engaged in each type of risky driving at 19-20 years (Figure 15). For instance, the majority of individuals who had at age 19-20 driven when affected by an illegal drug, engaged in high-level speeding or who had driven without a seatbelt no longer did so at 23-24 years. About half of those who drove when affected by alcohol and/or drove when very tired at age 19-20, continued to do so four years on.

Furthermore, those who engaged in risky driving at 19-20 years did so less frequently at 23-24 years than at 19-20 (Figure 16). For example, high-level speeders had driven more than 25 km/h above the legal limit on 2.6 (or 26%) of their past ten trips at 19-20 years, in contrast to only 1.0 (10%) of their ten most recent trips at 23-24 years. Similarly, among those who had driven when very tired at 19-20 years, this had occurred on 34% of their recent trips at 19-20 years, compared with 21% of their ten most recent trips at 23-24 years. Trends were similar for failure to wear a seatbelt or helmet, and driving when affected by alcohol or an illegal drug.

Figure 16 graph of frequency of engagement in risky driving behaviour by the same individuals, described in text.

Figure 16. Frequency of engagement in risky driving behaviours by the same individuals, at 19-20 and 23-24 years

Summary: Stability of different types of risky driving among individuals

We were interested in determining whether the individuals who engaged in each form of risky driving at 19-20 years were continuing to do so at age 23-24, and whether they were engaging in these driving practices as frequently as in the past.

We found that the great majority of young people who did not engage in a particular form of risky driving at age 19-20 continued not to do so at 23-24. This was particularly true for high-level speeding, driving when affected by an illegal drug, and driving without a seatbelt or helmet, with only 4-9% of drivers who did not engage in these behaviours at 19-20 years, engaging in them at 23-24.

There was less continuity in the behaviour of those who had engaged in risky driving at 19-20 years, with at least half of those who reported involvement in a risky driving behaviour at 19-20 years refraining from this behaviour at 23-24. Furthermore, by 23-24 years, they tended to engage in this behaviour much less often.

4.3 Similarity of 19-20 year old and 23-24 year old risky drivers

The previous sections investigated the persistence of driving tendencies over time. Section 4.1 showed that across the whole sample there was a modest decrease of 3-5% in three of the eight types of risky driving examined, a sizable increase in a fourth type and minimal change in the other four indicators. Section 4.2 showed that individuals who did not engage in a particular form of risky driving at 19-20 years were generally unlikely to do so at 23-24, while risky driving had greatly decreased at 23-24 years among those who had engaged in a risky driving practice(s) at 19-20 years.

Nevertheless, risky driving remained quite prevalent at 23-24 years. This next section looks at whether groups of individuals who were identified as risky drivers at 23-24 years showed similar propensities for risky driving as groups of individuals who were identified as risky drivers at 19-20 years. To answer this question, five separate risky driving sub-groups were identified at 23-24 years, using the same classifications as at 19-20 years. The sub-groups identified at 23-24 years are shown in Table 14.

Table 14. Risky driving sub-groups, at 23-24 years
  Sub-group Description n %
High-level speeding (> 25 km/h over the limit) No Did not exceed speed limit by more than 25 km/h 828 86.3
Yes Exceeded speed limit by more than 25 km/h 131 13.7
Driving without a seatbelt or helmet for entire trip No Wore a seatbelt/helmet for entire trip 912 94.8
Yes Did not wear seatbelt/helmet for entire trip 50 5.2
Driving when fatigued on multiple occasions a No Drove when very tired on 1 or fewer recent trips 599 62.2
Yes Drove when very tired on 2 or more recent trips 364 37.8
Driven when affected by alcohol No Did not drive when affected by alcohol 745 77.4
Yes Drove when affected by alcohol 218 22.6
Driving when under the influence of illegal drugs No Did not drive when affected by an illegal drug 889 92.3
Yes Drove when affected by an illegal drug 74 7.7

Note: a Due to the high prevalence of fatigued driving in the sample, this criterion differentiates those who drove when very tired on multiple occasions from those who only once or never drove when very tired.

We then compared the levels of risky driving shown by the 23-24 year old risky driving sub-groups at the age at which they were identified as risky drivers (i.e., at 23-24 years) to the levels of risky driving of the 19-20 year old risky driving sub-groups at the age at which they were identified as risky drivers (i.e., at 19-20 years).

Figure 17 shows the trends for these comparisons. For example, with regard to speeding, the group of young people who engaged in high-level speeding at 19-20 years had done so on about 26% of recent trips (when 19-20 years old). By comparison, the group of young people who engaged in high-level speeding at 23-24 years had done so on 22% of their past 10 trips (at 23-24 years of age).

Figure 17 graph of frequency of risky driving behaviours by risky driving sub-groups, described in text.

Figure 17. Frequency of risky driving behaviours, 19-20 and 23-24 year old risky driving sub-groups

The frequency of most risky driving behaviours was similar across the 19-20 and 23-24 year old risky driving groups. For instance, young adults who had driven when affected by alcohol at age 19-20 had done so as often as those who engaged in this behaviour at age 23-24 years (an average of 16% of trips at both ages).

Nevertheless, one difference was found: the group of 23-24 year olds who engaged in high-level speeding did so slightly less often than their 19-20 year old counterparts (26% trips at 19-20 years vs 22% trips at 23-24 years).

Summary: Similarity of 19-20 year old and 23-24 year old risky drivers

Young people who engaged in risky driving at age 23-24 were compared with those who engaged in risky driving at 19-20, to see whether 19-20 year old high risky drivers displayed similar or different risky driving propensities as 23-24 year old high risky drivers.

The frequency of most forms of risky driving was similar at both time points (e.g., young people who had driven when very tired at age 23-24 tended to have done so as often as young people who had driven when very tired at age 19-20). There was, however, one exception: 23-24 year olds who engaged in high-level speeding did so less often than their 19-20 year old counterparts.

4.4 Consistency of patterns of risky driving from 19-20 to 23-24 years

In the first In the Driver's Seat report, groups exhibiting differing levels of risky driving at 19-20 years were identified using cluster analysis (see Smart & Vassallo, 2005, for further details). In brief, three groups of drivers were identified: (a) a large group that was lowest on all types of risky driving - the low-level risky driving group (n = 675, 64% of the sample); (b) a small group that was highest on all indicators of risky driving, particularly speeding - the high-level risky driving group (n = 74, 7% of the sample); and (c) a group with rates of risky driving that were intermediate to the low and high groups - the moderate-level risky driving group (n = 306, 29% of the sample). These groups differed significantly in gender composition (with more females than males in the low-level group and more males than females in the high-level group), and in their experiences of crashes and being detected for speeding (the high-level group had experienced more crashes and were more often detected speeding). Looking back in time, the high-level group could be differentiated from the other groups from mid- to late childhood on a range of personal, family, peer and school progress characteristics.

This next section follows forward the risky driving tendencies of these three groups. Two approaches are taken:

  • the low-, moderate- and high-level risky driving groups43 are compared on separate indicators of risky driving at 19-20 and 23-24 years to investigate whether their risky driving propensities have increased, decreased or remained similar over this time period; and
  • cluster groups displaying differing levels of risky driving at 23-24 years are identified, and the 19-20 year old clusters' placement in these groups is probed.

Profiles of risky driving clusters 4 years later

Figures 18 to 20 display the profiles of the three risky driving groups identified at 19-20 years on various aspects of risky driving four years later (at 23-24 years). In these figures, the average number of trips during which young people had engaged in each type of risky driving behaviour at 19-20 years are compared to the average number of trips in which the behaviour had occurred at 23-24 years. Figure 18 shows trends for the low-level risky driving group, Figure 19 the moderate-level risky driving group, and Figure 20 the high-level risky driving group. As a larger number of items were included at 23-24 years than at 19-20, comparison data is not available for all items.

Looking first at the low-level risky driving group (Figure 18), there was a trend for young people in this group to engage in most types of risky driving more often at age 23-24 than they had at 19-20. However, with the exception of low- and moderate-level speeding and fatigued driving, these increases were very small. Rates of most forms of risky driving were generally very low at 23-24 years (generally occurring on less than 5% of trips), with the exceptions of low-level speeding (29% of trips), driving when fatigued, and all types of mobile phone use when driving, which on average occurred on about 10% to 15% of trips.

Figure 18 graph of frequency of risky driving behaviours among low-level risky driving group, described in text.

Figure 18. Frequency of risky driving behaviours, low-level risky driving group, at 19-20 and 23-24 years

Comparison of the profiles of the moderate-level risky driving group at 19-20 and 23-24 years showed a reduction in speeding (particularly in the low- and moderate-level categories) and fatigued driving over this time period (Figure 19). There were only small changes in the frequency of other types of risky driving behaviour over this time. Nevertheless, at 23-24 years, this group had, on average, been involved in low-level speeding on about 50% of recent trips, and in moderate-level speeding, driving when fatigued and mobile phone use on approximately 20% of recent trips. All other type of risky driving had a very low occurrence.

Figure 19 graph of frequency of risky driving behaviours for moderate level risky driving group, described in text.

Figure 19. Frequency of risky driving behaviours, moderate-level risky driving group, at 19-20 and 23-24 years

There was a marked reduction in speeding (at all levels) and driving when very tired between the ages of 19-20 and 23-24 years for the high-level risky driving group (Figure 20). However, there was little change in the frequency of drink-driving, driving when affected by illegal drugs, and driving without a seatbelt or helmet (for all or part of a trip) among this group. Risky driving was still quite prevalent among this group at 23-24 years, with low-level speeding occurring on about two-thirds of recent trips; moderate-level speeding on 40% of recent trips; using a mobile function on 30% of trips; and high-level speeding, driving when fatigued and other mobile phone use on close to 20% of trips.

Figure 20 graph of frequency of risky driving behaviours high-level risky driving group, described in text.

Figure 20. Frequency of risky driving behaviours, high-level risky driving group, at 19-20 and 23-24 years

To sum up, while there was a decrease in levels of risky driving among the high- and moderate-level risky driving groups from 19-20 to 23-24 years, these groups' levels of risky driving were still noticeably higher than the low-level risky driving group, who showed a slight increase in risky driving from 19-20 to 23-24 years.

Stability of cluster group membership from 19-20 to 23-24 years

The second approach taken was to investigate whether cluster group members identified at 19-20 years would continue to be classified in a similar way at 23-24 years of age. To investigate this question, cluster analysis was used to identify groups with differing levels of risky driving at 23-24 years, as had been done at 19-20 years.44

The analyses suggested that a solution of between two and five clusters provided the best representation of risky driving patterns in the sample (see Table 15 for a description of these solutions).

Table 15. Risky driving patterns, 2- to 5-cluster solutions, at 23-24 years
No. of clusters Cluster label n Description
2 Low 673 Lowest on all behaviours: some low-level speeding, fatigued driving and mobile phone use when driving
High 275 Frequent low- and moderate-level speeding, moderate fatigued driving and mobile phone use, low levels of other behaviours
3 Low 630 As above
Moderate 261 Moderate levels of all behaviours: frequent low-level speeding, regular fatigued driving and mobile phone use when driving
High 57 Highest levels of all behaviours: frequent speeding, frequent mobile phone use when driving, moderate fatigued driving and non-seatbelt use, some drink-driving and driving when affected by illegal drugs
4 Low 611 As above
Moderate 262 As above
High (a) 59 High-level risky driving (a): high levels of speeding, fatigued driving and mobile phone use when driving
High (b) 16 High-level risky driving (b): very frequent non-seatbelt/helmet use, and highest levels of drink-driving and driving when affected by illegal drugs
5 Low 582 As above
Moderate 271 As above
High (a) 67 High-level risky driving (a): highest levels of speeding and driving when affected by marijuana
High (b) 19 High-level risky driving (b): high levels of non-seatbelt/helmet use
High (c) 9 High-level risky driving (c): very high mobile phone use when driving

For all solutions of three or more clusters, low- and moderate-level groups consistently emerged. In the four- and five-cluster solutions, differentiation of high-level risky drivers occurred. For example, a small group who frequently drove without a seatbelt or helmet was identified (cluster groups 4b and 5b).

While all solutions were meaningful, the three-cluster solution was considered the most appropriate for future analyses. The small number of cases in some of the clusters identified in the four- and five-cluster solutions (group sizes of 9, 16 and 19) precluded their use in further analyses, while the two-cluster solution did not provide sufficient differentiation.

As shown in Figure 21, the three-cluster groups significantly differed in their frequency of engagement in risky driving behaviours at 23-24 years.45

Figure 21 graph of frequency of risky driving behaviours, three-cluster groups, described in text.

Figure 21. Frequency of risky driving behaviours, three-cluster groups, at 23-24 years

As found at 19-20 years, there were significant gender differences (Table 16),46 with more young men than young women in the high-level risky driving group and far fewer young men than young women in the low-level risky driving group.

Table 16. Gender composition of risky driving groups, at 23-24 years
  Males Females
N % N %
Low-level risky driving 211 33.5 419 66.5
Moderate-level risky driving 130 49.8 131 50.2
High-level risky driving 33 57.9 24 42.1

Furthermore, consistent with the findings at 19-20 years, the groups significantly differed on the amount of police contact for driving-related offences47 and number of times they were detected speeding,48 with the high-level risky driving group experiencing the highest rates of apprehension for a driving offence. In terms of crash involvement, the high- and moderate-level groups had almost identical rates, which were higher than the low-level group, although only the moderate-level group significantly differed from the low-level group on this aspect (perhaps because the size of the high-level group, n = 57, was too small to detect significant differences).49

It also should be noted that there was considerable variability in the number of crashes and speeding violations reported by young people within each of the groups (as indicated by the large standard deviations). This within-group variability was particularly notable in the high-level risky driving group.

Next we examined the degree of commonality between the cluster groups identified at age 19-20 and those identified at 23-24 years, to determine whether the same individuals who engaged in low-, moderate- or high-level risky driving at 19-20 years were still doing so at 23-24 years.

This comparison revealed a great deal of stability in the low-level group, with three-quarters (76%) of those classified as low-level risky drivers at 19-20 being classified in the same category at 23-24 years (Table 17). Of the remaining quarter whose risky driving behaviour increased over this time period, almost all were classified as moderate-level risky drivers at 23-24 years. Very few (2%) changed from being low-level risky drivers at 19-20 years to high-level risky drivers at 23-24.

Table 17. Level of subsequent risky driving behaviour among 19-20 year old risky driving cluster groups
  n % of group % of ATP sample
Low-level group
Remained stable 395 76 49
Increased from low- to moderate-level 115 22 14
Increased from low- to high-level 10 2 1
Moderate-level group
Remained stable 90 37 11
Decreased to low-level 125 52 16
Increased to high-level 26 11 3
High-level group
Remained stable 13 28 2
Decreased to moderate-level 23 49 3
Decreased to low-level 11 23 1

The moderate- and high-level groups were less stable, with the majority of individuals in these clusters at 19-20 years being classified as less problematic at 23-24 (52% of the moderate-level group and 72% of the high-level group showed improvement). However, some had become more risky. About 11% of those identified as moderate-level risky drivers at 19-20 years showed an increase in their risky driving behaviour between 19-20 and 23-24 years.

It is interesting to note that while it was extremely uncommon for a low-level risky driver at age 19-20 to become a high-level risky driver at 23-24, a change in the opposite direction was not as unusual, with almost a quarter of those classified as high-level risky drivers at 19-20 years showing low levels at 23-24 years.

Looking next at trends across the entire sample, cluster group position remained stable for the majority, with 62% being classified as displaying the same pattern of risky driving (low-, moderate- or high-level) at age 23-24 as they had at 19-20. However, almost one in five (19%) showed an increase in risky driving between the ages of 19-20 and 23-24, while a similar proportion (20%) showed a decrease in risky driving behaviour over this time period.

Of particular interest was the progress of the high-level group. It had been hoped to compare those who had improved with those who remained stable to explore whether particular characteristics or experiences may have been instrumental in facilitating change. However, the small numbers precluded these additional analyses.

Summary: Consistency of patterns of risky driving from 19-20 to 23-24 years

The final question examined was how consistent young people's general patterns of risky driving were between 19-20 and 23-24 years. This issue was examined in two ways.

First, we looked at the risky driving profiles at 23-24 years of the low-, moderate- and high-level risky driving groups identified at 19-20 years to see whether there had been a change in their risky driving practices over this four-year period. The high- and moderate-level groups were found to be less problematic at 23-24 years than at 19-20, while the low-level group showed a slight increase in most risky driving practices over this time period. The behaviours that showed the greatest rates of change were low- and moderate-level speeding, and fatigued driving, with the high- and moderate-level groups engaging in these behaviours less frequently at age 23-24, and the low-level risky driving group engaging in them more often.

Second, we used cluster analysis to identify groups with differing patterns of risky driving at age 23-24. As at 19-20 years, three groups with differing patterns of risky driving were identified: a low-level risky driving group that reported the lowest rates of all forms of risky driving behaviour at 23-24 years; a high-level risky driving group that generally reported the highest levels of risky driving; and a moderate-level risky driving group which reported levels of risky driving intermediate to the other two groups.

The degree of overlap between the cluster groups identified at 19-20 and at 23-24 years was then examined, to determine whether individuals were consistently engaging in low-, moderate- or high-level patterns of risky driving at both ages. This examination revealed a high degree of stability in the risky driving behaviour of the 19-20 year old low-level group, with three-quarters being classified as low-level risky drivers once again at 23-24 years. There was less continuity in the risky driving behaviour of the 19-20 moderate- and high-level groups, with the majority of these groups assigned to less problematic clusters at 23-24 years. While it was uncommon for a low-level risky driver at 19-20 to become a high-level risky driver by age 23-24, a change in the opposite direction was not as unusual.

4.5 Discussion and implications

In order to determine the stability of risky driving behaviour over time, four sets of analyses were undertaken.

The first set of analyses focused on the whole sample of young Australians to determine whether there had been a general change in rates of different types of risky driving from 19-20 to 23-24 years. A modest decrease of 3-5% was found for three of the eight types of risky driving examined (moderate- and high-level speeding, failure to wear a seatbelt or helmet for all of a trip), a sizable increase in a fourth type (driving when affected by alcohol), and minimal change in the other four indicators.

The second set focused on risky driving sub-groups, and looked at the stability of specific types of risky driving over 4 years (e.g., speeding, driving when fatigued, driving when affected by alcohol). Individuals who did not engage in these types of risky driving at 19-20 years were very unlikely to do so at 23-24 years. In addition, among those who had been involved in these types of risky driving at 19-20 years, the occurrence of the risky driving practice had greatly decreased by 23-24 years.

The third set investigated whether 19-20 year old risky drivers showed similar patterns of risky driving as 23-24 year old risky drivers. Rates of most forms of risky driving were relatively similar across the two age groups.

The final set of analyses followed forward the high-, moderate- and low-level risky driving cluster groups identified at 19-20 years (classified using all risky driving indicators simultaneously) to determine whether they would again be classified as high-, moderate- or low-level risky drivers at 23-24 years. A high degree of stability was found for the low-level group, with three-quarters being classified similarly at the later age. However, less continuity was found for the moderate- and high-level groups, with the majority of individuals from these groups found to be in less problematic clusters at 23-24 years.

Some conclusions and implications that may be drawn from the findings are presented below.

A safe approach to driving was highly persistent

An important finding emerging from this research is that the majority of young people who did not engage in risky driving at 19-20 years continued to refrain from unsafe driving practices four years on. For instance, 92% to 96% of individuals who did not drive when affected by illegal drugs, speed at high levels or drive without a seatbelt at age 19-20, continued to not engage in these behaviours at 23-24. Similarly, three-quarters of young people classified as low-level risky drivers at 19-20 years continued to exhibit a pattern of low-level risky driving four years later. While some low-level risky drivers exhibited an increase in risky driving over this period, for the majority, the change was not large. Only 2% progressed from little or no engagement in risky driving at age 19-20 to high levels of unsafe driving at 23-24. Thus, the possibility that increasing experience on the road might encourage initially safe young drivers to subsequently engage in risky driving was not supported by these findings.

These findings reinforce the value of current road safety initiatives aimed at discouraging risky driving among novice drivers, as they suggest that avoidance of risky driving during the early years of a person's driving career is very likely to persist and be linked to safe driving practices at a later age. Given that risky driving is associated with a heightened risk of crash involvement (Crettenden & Drummond, 1994; Williams, 1998) and engagement in a range of other risky or problematic behaviours such as antisocial behaviour and substance use (Vassallo et al., 2008; Williams, 1998), such preventative initiatives have clearly been of great value.

It will be important to continue tracking the driving behaviour of these low-level risky drivers in order to confirm that their pattern of low-level risky driving is maintained, and hence that early avoidance of risky driving is highly persistent.

Risky driving declined among many high-level risky drivers

Another encouraging finding was that engagement in risky driving did not persist for many young risky drivers. For instance, the majority of individuals who engaged in high-level speeding, driving when affected by illegal drugs, or who failed to wear a seatbelt or helmet when driving at 19-20 years no longer engaged in these behaviours at 23-24 years. Additionally, the majority of young people identified as being in the moderate- and high-level risky driving clusters at 19-20 years showed a reduction in risky driving by 23-24 years. For a small number, this improvement was quite marked, with one in ten high-level risky drivers showing low levels of risky driving by 23-24 years. Due to the small numbers available for analysis, it was not possible to undertake further analyses to identify factors associated with persistence or improvement, but this is clearly an important next step for research.

These findings highlight the fact that young problem drivers are not destined to continue posing a road safety risk as they grow older. Rather, it would appear that improvement is not only possible, but common. This finding is supported by past research showing a decline in problem behaviours as young people mature and assume adult roles (Bingham et al., 2008; Jessor et al., 1997). Nevertheless, once again, it will be important to continue to follow the progress of the high-level risky driving group over the coming years to determine whether they maintain this improvement.

While it may seem important to concentrate road safety efforts on young drivers who exhibit a stable pattern of high-level risky driving behaviour, we do not yet have the tools to reliably identify those who persist and those who improve, and the factors that facilitate positive change. Further, while risky driving may be temporary, it still has the potential to result in considerable personal, social and economic costs to individuals, families and the wider community, and to have a long-term impact on young people's life trajectories. Thus, it will remain essential to continue targeting risky driving among young novice drivers, to take them out of known high-risk situations at a time when they are building their driving skills (for example, via publicity programs highlighting the dangers of peers as passengers, or the restrictions imposed by the Victorian graduated licence system).

Risky driving was a continuing concern for a small number in their mid-20s

For a small number, risky driving continued to be a concern in their mid-20s. For example, over a quarter of high-level risky drivers identified at age 19-20 (albeit less than 2% of the sample) continued to exhibit risky driving tendencies four years on. Another group of concern were the small number who engaged in low levels of risky driving at 19-20 years and progressed to frequent engagement in risky driving in their mid-20s (2% of low-level risky drivers, 1% of the sample). Both groups would seem to pose major road safety risks, and hence further research focusing on these young drivers would appear warranted.

Questions that could be addressed include:

  • What personal, social or environmental factors differentiate persistently high-level risky drivers from other drivers, especially those who cease engaging in such behaviour?
  • What factors may have caused the small group who progressed to high-level risky driving at age 23-24 to "buck the trend" and start taking risks on the road at a time when many others had or were desisting from this type of behaviour?
  • How entrenched is this behaviour? If we were to look at the driving behaviour of these two groups in another four years' time, would they continue to engage in high levels of risky driving?

As noted earlier, due to the small number of young people who displayed these patterns of driving within our study, the first two questions are beyond the scope of this report. Such knowledge would be extremely helpful in informing intervention and prevention efforts targeted at these high-level risky driving groups. Hence, further longitudinal research, employing larger samples, is needed in order to better understand the attributes and experiences of these individuals, and the long-term stability of their driving patterns.

There were changes in rates of risky driving over time

Between the ages of 19-20 and 23-24, there was a marked decrease in rates of speeding and driving without a seatbelt among the ATP sample. These findings are heartening, given the strong links between excessive speeding and crash involvement (Clarke et al., 2002; Engström, Gregersen, Hernetkoski, Keskinen, & Nyberg, 2003) and non-seatbelt use and risk of injury or death in the event of a collision (Data Analysis Australia, 2000).

However, troublingly, rates of other types of risky driving either increased or remained stable over this time period. Driving when fatigued remained very prevalent, and had not dropped in incidence (almost two-thirds of the sample at both ages reported this had occurred in their last 10 trips). Furthermore, the number of young people who drove when affected by alcohol increased markedly from 14% at 19-20 years to 23% at 23-24 years.

The increase in drink-driving over this period may reflect differences in licence restrictions at the two ages. For instance, at 19-20 years, all of the drivers in the study held probationary licences or learner's permits, and thus were required to have a zero blood alcohol content (BAC) level when driving. Hence, it is not surprising that rates of drink-driving were relatively low at this age. However, at 23-24, when rates of drink-driving were noticeably higher, most young drivers held full licences and thus were permitted to have a low level of alcohol in their system when driving (the legal BAC level in Victoria is 0.05 g/100ml). Hence, it is likely that the more stringent restrictions imposed on less experienced drivers had a stronger deterrent effect on drink-driving than those placed on fully licensed drivers, possibly explaining this increase.

Nonetheless, the increase in prevalence of drink-driving between 19-20 and 23-24 years reported here is still of concern, and parallels the higher rates of crashes associated with drink-driving and convictions for drink-driving found at this age in the accident and enforcement statistics for the state of Victoria. For example, of the 50 drivers and motorcyclists killed in Victoria in 2008 who had a BAC at 0.05 g/100ml or over, 82% were males, 34% were between 21 and 29 years of age, 20% were aged between 30 and 39 years, 28% were aged over 40, and the remaining 18% were 20 years of age or younger (Transport Accident Commission [TAC], 2009a).

The prevention of drink-driving has been the focus of a wide range of road safety initiatives in Australia in recent years. However, despite these efforts, it would appear that these campaigns and programs are not reaching all young drivers. Further efforts are needed to discourage young people from engaging in this potentially lethal practice, as is better understanding of why current initiatives targeted at drink-driving are not connecting with some young drivers, and how this group may be better targeted.

Risky drivers at both ages were equally problematic

While there were signs that a modest reduction in engagement in risky driving from 19-20 to 23-24 years had occurred as shown by the overall sample trends; when young people did engage in risky driving at the older age, they did so almost as frequently as their younger counterparts. Thus, there was little support for the notion that with increasing age, those who engage in risky driving would do so at lower intensities than younger drivers. The driving tendencies of risky drivers were equally problematic at both ages. This suggests that risky driving is as serious an issue (when present) in the mid-20s as in the late teens, and points to the importance of sustaining road safety efforts into the twenties. While drink-driving is rightly highlighted as a particular problem for young people in their mid-20s, other forms of risky driving are also salient and merit attention.

Overall, levels of risky driving remained fairly stable

As noted above, there were signs that a small decline in risky driving may have been occurring among this sample of young Australians, with decreases of 3-5% on three of the eight types of risky driving examined (i.e., moderate- and high-level speeding, and failure to wear a seatbelt or helmet for all of a trip). However, there was a sizable increase in drink-driving, and virtually no change in the remaining four indicators. Thus, to the extent that there was change, this was modest rather than dramatic. While an improvement was evident in the subsequent driving careers of many individuals who had been high- or moderate-level risky drivers at 19-20 years, the general sample trends suggested only a modest degree of change. Overall, then, these findings suggest a fair degree of stability in risky driving over the four-year period examined.

These findings are somewhat at odds with past research, which suggests that risky driving typically decreases over early adulthood (Bingham et al., 2008; Jessor et al., 1997). A number of explanations are possible, although we have no direct evidence that can speak to these possibilities.

Measurement differences may have contributed. The indicators of risky driving used in the current study (e.g., speeding, drink-driving, driving when affected by illegal drugs, fatigued driving and non-use of seatbelts) differed from those used in several other studies, which have tended to focus on behaviours such as tailgating, unsafe passing and running red lights or stops signs. The young people in the ATP study may be at a different stage in their driving careers to those in other international studies. Our young people have been able to hold licences from 18 years of age, whereas those in studies from other countries have often held licences from an earlier age. There may also be differences due to studies being conducted in different time periods: our study shows trends for 2006-2007, while some other studies report trends from the 1990s or earlier. Cultural effects may also be present, with different attitudes towards risky driving, values concerning appropriate behaviour on the road and tolerance for acts such as drink-driving influencing societal trends in risky driving among young people.

The relative stability found emphasises once again the need for road safety efforts to continue targeting young people beyond the first years of their driving careers. Broad-based initiatives targeting all young drivers (e.g., community campaigns, police and enforcement efforts) may be beneficial in reducing the general level of risky driving among young people.

Conclusions

In summary, this chapter examined the stability of driving behaviour over the early adult years. Two main approaches were taken. The first involved examining the stability of particular forms of risky driving between the ages of 19-20 and 23-24, while the second looked at the constancy of young people's propensities to engage in multiple forms of risky driving across this time period.

While modest declines in some forms of risky driving were noted, and some individuals showed a decrease in their risky driving tendencies over this time period, levels of engagement in risky driving remained fairly stable across the sample over early adulthood.

  • 43 The term "risky" is used in preference to "risk", because "risk" implies that the behaviours are likely to occur but are not yet evident. Here, the unsafe driving behaviours are evident, and are therefore termed "risky" rather than "risk".
  • 44 A two-step clustering procedure was undertaken of the 23-24 year olds' responses to the 14 risky driving items. The first step involved identifying the appropriate number of clusters in the data. Random samples of approximately 200 cases were hierarchically clustered using Ward's method, with squared Euclidian distance used to measure inter-object similarity. Examination of dendograms and agglomeration schedules showed that a three-cluster solution provided good differentiation between groups. However, solutions between two and five clusters were also appropriate. Therefore, in the second step, K-means cluster analyses, specifying 2-, 3-, 4- and 5-cluster solutions were performed on the sample, using SPSS QUICK CLUSTER, a procedure that groups cases into clusters once the number of clusters is provided. These different solutions were inspected to find the most meaningful one. A three-cluster solution using running means was selected as the most appropriate representation of the data.
  • 45 The three risky driving groups significantly differed in the frequency with which they exceeded the speed limit by up to 10 km/h: F(2,945) = 1111.03, p < .001; exceeded the speed limit by 11-25 km/h: F(2,945) = 540.89, p < .001; exceeded the speed limit by more than 25 km/h: F(2,945) = 118.20, p < .001; did not wear seatbelt/helmet for part of a trip: F(2,945) = 106.09, p < .001; did not wear seatbelt/helmet at all: F(2,945) = 51.55, p < .001; drove when very tired: F(2,945) = 78.51, p < .001; nearly fell asleep (or fell asleep) when driving: F(2,945) = 31.15, p < .001; drove when affected by alcohol: F(2,945) = 36.92, p < .001; drove when affected by marijuana: F(2,945) = 7.235, p < .001; drove when affected by ecstasy: F(2,945) = 3.06, p < .05; drove when affected by amphetamines: F(2,945) = 6.67, p < .001; talked on a hands-free mobile when driving: F(2,945) = 13.58, p < .001; talked on a handheld mobile when driving: F(2,945) = 13.58, p < .001; used a mobile phone function when driving: F(2,945) = 182.29, p < .001.
  • 46 χ2(2) = 29.24, p < .001.
  • 47 χ2(2) = 51.65, p < .001; 40% of the high-level risky driving group had been in contact with police for a driving-related offence in the past year at the age of 23-24 years, compared with 20% of the moderate-level group and 9% of the low-level group.
  • 48 F(2, 932) = 46.91, p < .001; the high-level risky group had been detected speeding, on average, almost four times during their driving careers (M = 3.88, SD = 5.70), whereas the moderate-level group had been detected speeding twice (M = 1.97, SD = 2.07), and the low-level risky driving group, once (M = 1.13, SD = 1.65).
  • 49 F(2, 910) = 12.28, p < .001; (high-level group average number of crashes = 1.23, SD = 1.48; moderate-level group average number of crashes = 1.25, SD = 1.19; low-level group average number of crashes = 0.87, SD = 1.00).

Next: 5. Risky driving and substance use

Top