Monash University Accident Research Centre - Report #16 - 1991
Authors: B. Fildes, G. Rumbold & A. Leening
Four studies were undertaken at rural and urban road sites aimed at relating motorists' attitudes to speed with their actual on-road speed behaviour. Those travelling at excessively fast and slow travel speeds were of particular interest. Drivers' speeds were measured at each site and target vehicles were stopped at a set of traffic lights further along the road and asked to participate in a general road safety survey. Those who agreed pulled off into an off-road parking area and were questioned about a number of driver, vehicle, trip purpose, speed attitude, and accident history factors. Significant travel speed relationships were found at all sites for driver age, number of occupants, purpose of the trip, intended speed, safe speed, accident involvement and the total number of accidents. In addition, vehicle type, age, and whether the vehicle was towing or not was related to travel speed in rural areas, while amount of weekly travel was associated with travel speed in urban areas. A surprisingly high number of motorists at all speed levels did not believe it to be dangerous to travel 30km/h above the posted speed limits and most thought the chance of being stopped by the police for speeding at these sites to be low. Multivariate analysis was also undertaken to examine the relative importance of (and important interactions between) the variables. While these findings provided some additional useful and novel information, care needs to be taken with these results because of the relatively small amount of variance explained and hence, predictive power (other factors and interactions are clearly important for a drivers' speed decision on the road). The findings of the study point to a number of potential countermeasures against speeding and some additional research required in this area.
Four studies were carried out to assess speed behaviour and drivers' attitudes to speeding on straight and curved roadways in urban and rural Victoria.
This research was aimed at describing the relationship between driver and vehicle characteristics, travel distances and times, and driver attitudes to speeding (their stated intentions to behave in a particular manner) with their actual on-road speed behaviour. The relationship between observed speed behaviour and five year accident history was also of interest.
The study was to provide new knowledge in this area as well as to identify potential countermeasures against excessive speeding.
Drivers who travelled at excessively fast and slow travel speeds (upper and lower 15% of the speed distribution) were of particular interest in this research program. The executive summary, however, emphasises the findings for excessively fast drivers only, for reasons apparent in the results.
These studies were undertaken prior to the extensive speed camera program that commenced in this state in April 1990.
A technique was developed which involved unobtrusively measuring vehicle speeds manually at the test sites and then stopping target vehicles further down the road for interview. Selected drivers were asked a series of questions aimed at eliciting information on driver demographics, trip and vehicle data, as well as speed attitude and accident history details.
These responses were subsequently normalized and combined across similar speed zones (10Okm/h in rural and 60knilh in urban areas) and analysed in terms of their relationship with travel speed. The relative importance of, and some interactions between, the variables were also examined.
There were a number of significant findings from this research.
DRIVER AGE - Younger drivers (those aged under 34 years) were more likely to exceed the speed limit and be excessively fast drivers at all locations. Older drivers (those aged 45 years or more) were more likely to be below the speed limit and to be excessively slow motorists.
NUMBER OF OCCUPANTS - Vehicles with single occupants (driver only) were more likely to exceed the speed limit and be excessively fast vehicles, while those with 2 occupants were less likely to exceed the speed limit. This was consistent for urban and rural settings and for straight and curved sections of roadway. There was no significant relationship between vehicles with 3 or more occupants and travel speed.
PURPOSE OF THE TRIP - Business travellers were more likely to exceed the speed limit and to be excessively fast drivers at all locations. Conversely, those travelling for recreation or domestic purposes were more likely not to exceed the speed limit. Those travelling behind schedule were more likely to exceed the speed limit and to be excessively fast motorists than all other travel schedules categories. There was a tendency for those not on any particular travel schedule to be less likely to exceed the mean traffic speed.
TYPE OF VEHICLE - The type of vehicle being driven was associated with excessively slow travel speeds in rural, but not urban, areas. Those driving vans and light commercial vehicles were more likely to be very slow travellers than were passenger car drivers. There was no statistical difference in travel speed behaviour, however, between the different categories of passenger cars.
VEHICLE AGE - Drivers of recent vehicles (4 years old or less) were more likely to exceed the speed limit and travel at excessively fast travel speeds than drivers of older cars in both urban and rural areas.
TOWING A TRAILER- Vehicles observed not towing a trailer in rural areas were more likely to exceed the speed limit and to travel at excessively fast speeds than those towing. Again, this was not apparent in urban areas, although the numbers of vehicles towing were considerably less here.
WEEKLY TRAVEL - Those who reported travelling high distances each week were more likely to exceed the speed limit and travel at excessively fast speeds in urban areas than low weekly distance travellers. There was no such finding observed in rural environments.
SPEED JUDGEMENTS - Motorists' travel speeds were highly correlated with what speed they nominated they would travel at for all locations. Most motorists correctly nominated the posted speed limit for both the rural (1 00knVh) and urban (60knVh) sites and this judgement was not related to their travel speed.
SAFE & DANGEROUS SPEED - Estimates of what constituted a safe travel speed were positively correlated with observed speed and own speed estimates at all locations. There was also a significant negative correlation between travel speed and what the driver nominated as a dangerous speed at the urban sites.
EXCESSIVELY FAST SPEED - A substantial number of motorists interviewed believed it was not dangerous to exceed the posted speed limit by 3Okm/h at both rural and urban locations (the rates were higher for straight than curved road settings). A surprising proportion of those travelling at excessively slow speeds (up to 30%) also did not believe travelling 3Okm/h above the speed limit to be dangerous.
SPEED DETERRENCE - Most motorists assessed the likelihood of being stopped by the police for exceeding the speed limit by 2Okm/h at these sites to be less than 50 percent. Moreover, there was no correlation between observed travel speed and the driver's assessment of the likelihood of being stopped by the police at these locations.
ACCIDENT INVOLVEMENT - Those who reported having been involved in a crash (and multiple crashes) over the past 5 years were more likely to be travelling above the mean speed of the traffic and the speed limit (there were insufficient numbers to statistically test crash involvement with excessively fast and slow travel speeds). Contrary to earlier overseas findings, there was no indication of any increased crash involvement for those travelling below the mean traffic speed in these data. There were relatively small numbers of accidents reported overall in these studies.
INJURY SEVERITY- There was a trend for those travelling above the mean traffic speed in all locations to report more severe injuries in previous crashes than those below. Those travelling at excessively fast speeds were more likely to report injuries requiring hospital or medical treatment, while nobody travelling at excessively slow speeds reported severe injuries.
FINDINGS WHICH WERE NOT SIGNIFICANT
There were a number of findings which were not significantly related to travel speed by themselves at any of the test sites in this research program. These included the following variables:
- sex of the driver,
- distance travelled prior to interview,
- distance still to travel,
- time since last rest stop,
- how tired the driver felt, and
- who owned the vehicle.
It was not possible to conduct a reliable analysis of travel speed for drivers not wearing their seat belts and those vehicles displaying P-Plates, given the very small numbers of observations recorded for these factors at all sites.
RELATIVE IMPORTANCE OF THE VARIABLES
Multivariate statistical analyses were undertaken to demonstrate the relative importance of (and important interactions between) the above variables for speed behaviour in rural and urban areas.
The rural analysis showed that vehicles observed exceeding the mean traffic speed and the posted speed limit were more likely (than slower speed vehicles) to have one or more of the following characteristics, in order of importance:
- not towing a trailer,
- young driver (under 34 years) with a high accident history,
- drivers reporting a high safe travel speed,
- male drivers travelling over long distances not for domestic purposes,
- vehicles with a driver only, travelling for business purposes, who drive high weekly mileages, and do not own the vehicle they drive.
Those travelling below the mean speed would be expected to have one or more of the opposite characteristics.
The urban analysis revealed that vehicles exceeding the mean traffic speed and the posted speed limit were more likely to have one or more of the following characteristics (also in order of their importance):
- a young driver (under 34 years) with a high accident history,
- drivers reporting a high safe travel speed,
- drivers in vehicles less than 5 years old,
- business travellers (not on recreation journeys) who travel high mileages each week.
There was reasonable consistency in the relative importance of the variables across the two sites, although the relationships were weaker for the urban curved arterial road in a highly residential area.
Not all of the factors found to be significantly related with travel speed were related to travel speed in the multivariate analysis. Furthermore, other factors not significant in their own right reached prominence with regression. These apparent anomalies can be explained by the relative importance of each variable and the interactions that occurred between these variables.
Care should be taken in interpreting predictive power from these analyses, given the relatively low amount of variability in travel speed explained by the variables collectively (up to 32%). Drivers' decisions about travel speed on the road are clearly multi-factorial and may also involve additional important factors not tested here.
There were several other interesting findings that arose from the research.
FREE SPEEDS - Mean vehicle speeds measured at both the rural and urban sites were higher than the posted limit on straight sections (+5.9km/h rural and +12.3km/h urban) but generally lower on curved section of roadway (-7.6km/h rural and +2.3km/h urban). There were very few differences in mean travel speed across the different days of the week during the study period.
SAMPLE CHARACTERISTICS - The sample of drivers interviewed was generally quite representative of the population of motorists using the roadway at each study she. There was a slight tendency for interviewees to be in less of a hurry at all locations, to drive smaller cars and light commercial vehicles at the rural sites, and to be older at these locations. Moreover, interview rates were higher in country areas, suggesting that these drivers had more time available (or were more approachable) than those at city locations.
MEASUREMENT - The technique developed here was a particularly useful means of relating drivers' characteristics and attitudes to speed with their on-road driving behaviour. The usefulness of the accident data is somewhat limited by the small (and rather costly) amount of data available, although R is advantaged by having details on property damage crashes, not available elsewhere at this stage.
IMPLICATIONS OF THESE FINDINGS
It should be remembered that these results were based on studies undertaken before the current speed camera program began in Victoria. Hence, some of these findings may need re-testing as a consequence of this program.
The studies showed a sizeable number of motorists travelling at excessive speeds at both the rural and urban locations. This practice is clearly unsafe from the accident data presented and needs to be discouraged. There are several measures available for reducing the incidence of excessive speeding.
EDUCATION & ENFORCEMENT- Drivers at all speed levels, especially those likely to travel at high speeds, need to be convinced of the increased danger travelling at speeds markedly greater than the mean traffic speed, both in terms of crash involvement and greater severity of outcome.
Moreover, other research has shown that speed education or promotion effort needs to be supported by a fully integrated speed enforcement campaign if 0 is to be successful.
TARGETING UNSAFE SPEED BEHAVIOUR - The research conducted here found a number of factors which identify drivers who should be given priority in targeting unsafe speed practices in any future campaigns against excessive speed.
DRIVER ATTITUDE CHANGE - Driver attitudes towards what constituted a safe travel speed were related to speeding behaviour, although not as strongly as driver age and accident history. This is further evidence of the need for an educational campaign aimed at changing this attitude. Previous experience suggests that achieving a change in attitude may require a long-term program of measures using a mufti-facet approach.
INCREASED RISK OF SPEED DETERRENCE - Motorists in this study had a low perception of the probability of being detected by the police for exceeding the speed limit by 20km/h. This needs to be increased substantially if police enforcement is to be used as a speed deterrent measure. Highly visible police enforcement effort is required to change this perception.
SPEEDING PENALTIES- Current speeding penalties were not seen to be particularly severe by the motorists interviewed in this study. If speeding penalties are to act as a speed deterrent, there may be a need to increase their severity.
It should be noted, though, that other research has shown that increases in perceived detection rates are likely to be more influential than increased penalties alone.
IN-VEHICLE SPEED CONTROL- Education and enforcement campaigns may not be totally successful in reducing excessive speed. Top speed limiting devices are being introduced for certain heavy vehicles to prevent excessive speeds. It may be necessary for these devices to become standard equipment on all vehicles to stop these deviant practices. They could be introduced in the first instance as a recidivist device.
In addition, technology for extending the operation of speed limiters for use in urban areas exists today, although current designs generally require substantial road infrastructure equipment and costs. These devices need further investigation and development at this time.
PERCEPTUAL COUNTERMEASURES - Previous studies support the finding here that motorists' perceptions of the road and their environment have a marked influence on their speed behaviour. A number of on-road speed perception countermeasures are available for use at specific hazardous locations to change this perception and ultimately reduce travel speed. However, many of these measures still need to be evaluated.
OTHER POSSIBILITIES - The data collected here suggested that consideration might also be given to lane separations to limit the interaction between fast and slow vehicles and what is an appropriate design speed for rural highways in this state.
ADDITIONAL RESEARCH REQUIRED
A number of areas requiring further research in speed behaviour and accident involvement were identified in this research program.
ACCIDENT INVOLVEMENT - There was a suggestion in these data that motorists travelling below the mean traffic speed were not over-involved in accidents, and that fast driver involvement rates are of a different form and may not be as large, as has been reported previously. These findings may be a function of the small amount of data collected in this study and need to be examined further, given the lack of information currently available on the speed and accident causation relationship and the urgent need for accurate and detailed information in this area.
LIMITING VEHICLE SPEEDS - There is a need for a proper review of devices for limiting vehicle speeds in urban and rural areas, including their availability, suitability and likely costs and benefits. The question of how they could be introduced into the vehicle fleet also needs to be addressed.
BETTER SPEED DATA - Past research shortcomings in this area stem from the lack of available and accurate information on the speeds of vehicles involved in accidents. An on-board speed measurement device (a "black-box" recorder of lower cost than a tachograph) would be extremely helpful for better understanding this phenomenon and could be useful for ascribing culpability in road crashes. However, research and development effort is still required at this stage for developing such a device.
FURTHER SPEED ASSESSMENT - Drivers' speed behaviour and attitudes to speeding need to be assessed for a number of other (additional) road and environment conditions. In particular, night-time and poor weather conditions are likely to have an influence on these judgements and need to be examined further.
SPEED CAMERA PROGRAM - It might also be useful to re-test some of these findings at existing sites to assess the effect of changes in risk perception and behaviour as a result of the current speed camera program in this state.