MUARC's Speed Enforcement Research: Principles Learnt and Implicationsfor Practice

Monash University Accident Research Centre - Report No. 200 - 2003

Full report in pdf format [215KB]

Authors: Delaney, A., Diamantopoulou, K. & Cameron, M. 

Abstract

The effectiveness of speed enforcement programs in Victoria has been the subject of a significant volume of research conducted by the Monash University Accident Research Centre (MUARC). This report provides a coherent review of MUARC's speed enforcement research to date and highlights the enforcement principles established by the research. The report also examines the practical implications of the research for future enforcement operations. 

The MUARC research reviewed in this report includes evaluations of the effectiveness of a range of enforcement technologies including mobile speed cameras, hand-held laser speed detection devices, mobile radar speed detectors and fixed speed cameras. The effectiveness of these technologies has been evaluated in terms of both the effect on casualty crash frequency and severity and in some instances the effect on driver behaviour. Principles and implications arising from this research relating to the timing, intensity and location of enforcement operations are provided where appropriate. In addition, the role of mass media publicity and driver perceptions of enforcement activities are examined in the context of future enforcement operations. 

Executive Summary 

The effectiveness of speed enforcement programs in Victoria has been the subject of a significant volume of research conducted by the Monash University Accident Research Centre (MUARC). This research has evaluated specific speed enforcement programs in terms of their effect on casualty crashes and in some instances their effect on driver behaviour. The role of mass media publicity supporting the speed enforcement operations has often figured in this research. More general research has also been conducted that examines driver perceptions of enforcement activities and the reported behaviour of specific groups of drivers. 

This report provides a coherent review of speed enforcement research conducted by MUARC to date and practical measures that can be used in planning future enforcement operations. 

General Principles for speed enforcement 

All types of speed enforcement examined by MUARC have led to at least some positive impact on either casualty crash frequency, crash severity or driver behaviour. In most cases this effect has been significant. However, the effectiveness of the enforcement program may depend on the appropriateness of the speed enforcement device or scheduling method for the road type on which the program operates. The research suggests that the following conclusions are relevant in this context: 

• Speed cameras are effective in reducing casualty crash frequency on arterial roads in metropolitan Melbourne and country towns and on highways in rural areas. These devices are also effective in reducing crash severity in Melbourne. 
• Mobile radar devices are effective in reducing casualty crashes in rural areas on undivided roads in 100 km/h speed zones. Analysis of the effects of the mobile radar devices in outer metropolitan areas was inconclusive. 
• Hand-held laser speed detection devices are effective in reducing casualty crash frequency, but not severity, on arterial roads in metropolitan Melbourne. 
• The Random Road Watch program, which involves the allocation of overt enforcement resources in a random way with respect to time and space, is effective in reducing crashes of all severity levels in non-metropolitan areas. The effect of the program in metropolitan areas is unclear. 

It is noted that the use of laser devices was intended to alleviate problems associated with speed detection using hand-held radar devices on arterial roads. In addition, mobile radar devices were primarily intended for use on two-way, undivided, 100 km/h speed limit roads in rural areas where it is difficult to use speed cameras covertly. It is therefore evident that speed detection devices work best when used on the road type for which they were intended. This indicates the following principle: 

Principle 1: The use of different types of speed enforcement devices should concentrate on the road type or geographical region for which their use is intended. 

The second important general result relates to the duration of the impact of speed enforcement. The analysis of speed cameras and mobile radar devices has produced important results in this area. 

It is important to note a number of key differences between these devices. First, the use of speed cameras in Victoria is largely covert and infringement notices issued as a result of speed camera operations are usually received one to two weeks after the offence occurs. In addition, speed cameras are able to detect large volumes of speeding motorists per enforcement hour. In contrast, an offender detected speeding by a mobile radar device is issued with an on-the-spot fine in a deliberately overt manner. Also, the volume of speeding motorists detected by mobile radar devices per enforcement hour is lower than that for speed cameras. 

Differences in the duration of the enforcement effects are also evident. Speed camera enforcement operations have been shown to impact on casualty crash frequency during the two weeks after the infringement notices are received. The exact duration of the effect is unclear. On the other hand, mobile radar enforcement was found to have the greatest effect during the four days immediately following the enforcement operations. Differences in the duration of the enforcement effects are important issues to consider when planning future speed enforcement programs. However, it is unclear which of the differences between the two enforcement programs result in the different duration of the enforcement effects. It is therefore difficult to extend these findings to speed enforcement programs operating under different conditions. 

Principle 2: The duration of the effect of overt speed enforcement programs resulting in visible issuance of on the spot fines issued at the time of the offence is likely to be more immediate but shorter than the effect generated by a delayed issuance of infringement notices resulting from covert operations which detect many offences per enforcement hour. 

Effects of Overt Speed Enforcement 

As with any form of enforcement program, the effectiveness of overt speed enforcement may be due to the effect of specific deterrence, general deterrence or some combination of the two. There appears to have been some conflict as to whether overt enforcement has a significant general deterrence effect. Given the extensive use of overt speed enforcement in Victoria, some discussion of this conflict is necessary. 

First it is noted that the initial introduction of the speed camera program in Victoria involving the overt use of cameras had no overall impact on casualty crash frequency. In addition, some research indicated that no relationship between the amount of enforcement seen by drivers and a driver's perceived risk of detection could be found. These results indicate that speed enforcement operations may not have a general deterrence effect. 

On the other hand, more recent research relating to laser speed detection devices found that the use of these devices resulted in a significant decrease in casualty crashes. This supports the proposition that overt enforcement activity can have a significant impact on crashes and has a general deterrence effect that may be localised in space. 

In addition, an evaluation of the New Zealand speed camera program found that the use of covert speed cameras to supplement overt cameras in signed areas generalised the effect of the program beyond the speed camera sites. It is concluded that the use of visible symbols, such as signs warning of potential camera presence and marked enforcement vehicles, may help to remind drivers of the unseen threat of covert operations, thus increasing general deterrence. 

These results suggest that the general effect of overt speed detection programs can be increased by expanding the geographical coverage of the enforcement sites. In terms of future enforcement operations this suggests the following principle: 

Principle 3: To maximise the general deterrence effect of overt enforcement programs, these programs should involve low to medium intensity speed enforcement at multiple sites on the road network. 

In addition to the localised, general deterrence effect of overt speed enforcement, it has been suggested that some forms of speed enforcement have a specific deterrence effect. This effect is primarily used in the speed camera program where the receipt of Traffic Infringement Notices (TINs) has been shown to impact on subsequent speeding behaviour, and consequently the number of casualty crashes occurring. Early research relating to the period from 1990 to 1993 showed that the receipt of TINs for speeding offences resulted in reductions in casualty crashes. However, similar analysis using data from 1994 to 1996 found evidence of reductions in crash severity but not casualty crash frequency as a result of the receipt of TINs. 

The change in the effect of the receipt of TINs suggests that the specific deterrence effect of speed camera enforcement may have declined over time. This decline may be in part due to changes in the public's attitudes towards receiving TINs. The public may have become accustomed over time to receiving TINs and consequently may no longer adjust their driving behaviour. Alternatively, the magnitude of the penalties may not have been sufficient to deter speeding behaviour. Further research into this area is required to determine whether speed enforcement programs continue to have adequate specific deterrence effects. 

Mobile Speed Cameras 

1. The initial use of a small number of overtly operated speed cameras in Victoria could not be shown to reduce casualty crash frequency and its effect on speeds was limited to within 1-2 km of the enforcement site. This implies that the overt use of speed cameras is only effective in reducing speeds in the vicinity of the camera site and may produce small, localised casualty crash reductions, but has no general impact on casualty crash frequency. 

2. Significant reductions in low alcohol hour 1 casualty crashes were experienced during the initial implementation phase of the new, covert speed camera program in 1990 when publicity was high but actual enforcement levels were relatively low. This suggests that, even with low enforcement levels, high profile media activity can establish and maintain a threat of detection in the short term. 

3. Following the full implementation of the speed camera program, statistically significant reductions in low alcohol hour casualty crashes were found across arterial roads in Melbourne and country towns and on rural highways. This demonstrates that, the intense, covert use of speed cameras can lead to long term reductions in low alcohol hour casualty crashes across a number of road types when accompanied by high-profile publicity (i.e. a general effect much broader than any localised effects). 

4. Initially, no casualty crash reduction effects were found within one kilometre of enforcement sites in the week following enforcement sessions. Localised reductions in high alcohol hour casualty crashes were experienced during the two weeks following the receipt of TINs. This indicates that, in addition to the general effect outlined above, the speed camera program has a localised effect on high alcohol hour casualty crashes corresponding to the two-week period after the receipt of TINs. The exact duration of this effect is unknown. 

5. Later research identified additional casualty crash reductions within one kilometre of enforcement sites in the week following enforcement sessions. This demonstrated that the localised effect of the speed camera program was also linked to the enforcement presence although perhaps less strongly than the receipt of TINs. Recent changes to make camera operations more covert and unpredictable may have weakened this localised, additional effect. 

6. From the research relating effects to the receipt of TINs emanating from camera-detected speeding offences, it can be concluded that specific deterrence, operating through the actual detection and punishment of offenders, works effectively to reduce high alcohol hour casualty crashes. 

7. Links between reductions in casualty crashes and road safety advertising relating to speed and drink-driving enforcement operations have been established. Enforcement related advertising effectively magnifies the deterrent effects of enforcement operations. 

8. During the period 1990-1993, both the TINs issued and publicity with speeding-related themes contributed to reductions in serious casualty crashes. This implies that, both specific deterrence (achieved through the issuing of TINs) and general deterrence (achieved through media advertising of speeding offences) are effective in reducing serious casualty crashes. 

9. An economic analysis of speed camera operations in Melbourne during low alcohol hours indicates that effective speed enforcement programs are highly cost beneficial. The benefits were measured by the reduced social costs due to savings in road crashes and injuries, and the costs were those required to operate the speed camera program. Fine revenue was not included as a benefit (or a cost). The analysis indicated that the level of speed camera activity used up to 1998 could be increased with further economic benefits. The level of camera activity was increased by 50% during 2001/02, consistent with this finding. 

Hand-Held Laser Speed Detection Devices 

1. The laser speed detection program implemented during 1997 was successful in reducing casualty crashes on arterial roads in metropolitan Melbourne when conducted at low to medium intensity levels (sessions typically less than one hour, for up to 15 hours per site per year). Therefore, overt forms of speed detection can be effective in reducing casualty crashes when enforcement operates at low to medium intensity levels. 

2. The effect of laser detection devices may be localised in space suggesting that overt forms of speed enforcement may have a general deterrence effect that is limited to the location at which enforcement activity is observed. 

3. To expand the general effect of overt speed detection programs, a more extensive geographical coverage may be required. 

4. It follows from points 1 to 3 that low to medium intensity speed enforcement at multiple sites on the arterial road network will have a greater effect on crashes than high intensity enforcement at fewer locations. 

5. The casualty crash reduction effects identified were found to be statistically significant for arterial roads only. Current knowledge does not allow firm conclusions concerning the effectiveness of laser enforcement on non-arterial roads. 

Mobile Radar Speed Detectors in Marked and Unmarked Cars 

1. The effect of mobile radar enforcement and associated publicity was found to be greatest during the four days immediately following the enforcement, with the effect diminishing between five and seven days after the enforcement presence. This suggests the existence of a four-day residual enforcement effect on casualty crashes. 

2. Casualty crash reductions were found following speed enforcement operations involving either covert operations (unmarked cars) or a mix of overt and covert enforcement activity (marked and unmarked cars operating together). 

3. No statistically significant effect on casualty crashes was found on the actual day of enforcement activity in rural Victoria. Therefore, it is not clear that mobile radar activity is effective in reducing casualty crashes in rural Victoria on the actual day of enforcement. 

4. The most significant impact on casualty crashes occurred when high public awareness of media publicity accompanied radar enforcement. However, the effect was greater when high publicity awareness levels were specific to mobile radar enforcement as opposed to when the high publicity awareness levels related to general, speed-related road safety themes. 

Therefore, publicity relating to specific enforcement programs may be more effective in reducing casualty crashes when accompanying the enforcement than publicity relating to speeding generally. 

5. The strongest effect on casualty crashes was found on the same day as enforcement when high levels of mobile radar publicity accompanied mobile radar enforcement operations involving both overt and covert activity. 

Fixed Speed Cameras On Citylink 

1. Fixed position, automated speed cameras located in the Domain tunnel have successfully reduced the average speed of motorists travelling in the tunnel and the proportion of vehicles exceeding speeds of 80, 90 and 110 km/h. 

2. Reductions in the measures of effect described in point 1 can be expected to translate into reductions in crash risk and injury risk in the tunnel. 

3. The fixed position speed cameras have had varied effects depending on the lane the vehicle is in, the day of the week and the time of day that the vehicle travels. In particular the cameras have: 

• Been more effective in reducing vehicle speeds in the left lane than in the right; 
• Been more effective in reducing vehicle speeds on weekends than during the week; and 
• Been most effective in reducing vehicle speeds on weekdays during the peak afternoon period. 

No time of day pattern is evident on the weekends. 

Enforcement Scheduling: The Random Road Watch (RRW) Program 

1. The RRW program has been successful in producing statistically significant crash reductions at all severity levels in non-metropolitan Queensland. The effect of the program in metropolitan areas has been more difficult to assess. 

2. The effect of the RRW program has increased over time for all crash types (except those involving fatalities, for which the effect was approximately constant). 

3. From points 1 and 2 above it is concluded that, randomising the time and location of police road safety enforcement can have significant positive effects on the number of crashes reported and this effect has been shown to increase over time. 

4. The positive effect of the RRW program on crashes of various severities in both rural and urban areas indicates that, such a program may be as effective on lower speed roads as on open highways. 

5. The statistically significant association between program coverage and crash effects indicates that in operational terms, wide spread but perhaps less intense, randomised enforcement will result in greater crash effects than more intense but less diverse coverage. That is, the coverage of a randomised enforcement program is a key contributor to its effectiveness. 

6. Analysis of similar programs in other jurisdictions suggests that the results experienced in Queensland could be reproduced elsewhere. 

Specific Groups at Risk 

A number of specific groups have been identified as presenting particular road safety problems in regard to speeding. Having identified these groups it is then necessary to formulate enforcement programs or other measures that can be targeted towards these groups and have positive impacts on their behaviour and crash involvement. 

The following groups of drivers may need to be the target of specific programs in future. First, it is noted that for two particular groups of drivers, enforcement may not be sufficient to alter their behaviour. It is also noted that there is some overlap between these two groups. 

The first group is those classified as younger drivers. For this group, it was found that enforcement is unlikely to be the dominant determinant of behaviour. Therefore, enforcement alone is unlikely to be sufficient to address the behaviour of this group. This suggests the following principle: 

Principle 4: Public education campaigns should play an important role in targeting younger drivers particularly when accompanied by a significant increase in speed enforcement activities during the evenings and late at night. 

The focus of education campaigns should be on the particular biases of those young people involved in crashes including biases towards males, car owners, those who drive for social reasons at night and people with offence histories. In addition, advertising specific to the road conditions in which these drivers face difficulties would be useful. 

The second group of drivers for whom enforcement is unlikely to affect behaviour is also likely to benefit from an education campaign. This group is comprised of those drivers under the age of 20 and those drivers between the ages of 20 and 29 who perceive their personal risk of detection as high but are nevertheless prone to speeding behaviour. 

Three further groups were identified as possible targets of future enforcement initiatives. The first of these groups had a high recall level for road safety material presented on the radio. This group is more likely to be male, live in metropolitan areas, drive as part of their occupation and have higher levels of exposure to driving risk and enforcement. 

Principle 5: Targeting of drivers who drive as part of their occupation and possess the other specific characteristics detailed above should involve radio advertising. 

The second group identified has a low perceived risk of detection combined with a relatively high likelihood of alcohol consumption. Members of this group are more likely to be classified as young drivers, drive in built-up areas during the daytime and experience lower levels of exposure to enforcement activities. 

The final group comprised mainly of 20-29 year olds, was predominantly male and believed that speed enforcement was overt and that detection could be avoided. 

Principle 6: Targeted enforcement programs aimed at the identified groups of drivers who are either exposed to lower enforcement levels or believe that enforcement can be avoided must involve an increase in the actual detection of offences. 

Given the characteristics of these two groups, increasing the actual level of offences detected amongst these groups should alter perceptions of the levels of enforcement and the risk of detection when speeding. 

Threshold level of enforcement and Driver Perception of Overt and Covert Enforcement Initiatives 

1. Some research suggests that an increase in visible enforcement must achieve a threshold level for that increase to be effective in reducing casualty crash frequency. The exact threshold required is unclear, but few studies have shown effects on crashes when visible enforcement is increased by a factor of less than three. 

2. Later research of both manual and automatic enforcement techniques does not indicate a threshold with a smooth relationship between the amount of enforcement and the level of road safety. 

3. Overt Police enforcement operations are likely to have a general deterrence effect. On the other hand, covert operations need to rely on high levels of detection to ensure that specific deterrence is effective. 

4. Two groups of drivers have particularly negative speed-related attitudes and behaviours, which may need to be targeted in future speed enforcement initiatives. 

5. The first risk group comprised mainly of 20-29 year olds was predominantly male, and believed that speed enforcement was overt and that detection could be avoided. Effects on this group could benefit from an increase in their actual risk of detection. 

6. The second group of drivers perceived their personal risk of detection as high but were nevertheless prone to speeding. Both genders are represented in this group and they were predominantly comprised of drivers under 20 or between 20-29 years of age. Given that the second risk group already reports a high risk of detection, the attitudes and behaviours of this group may not be affected by enforcement activities alone. 

Speeding and Young Drivers 

1. Speed related crashes are more likely to involve younger drivers, particularly younger male drivers. This suggests that younger drivers present particular problems in relation to speeding, and should be a target for speed enforcement. 

2. Younger drivers are over-represented in crashes occurring on sections of curved road especially in the evening in wet road conditions and late at night. This corresponds with low levels of speed enforcement during the evening and late at night. 

3. Point 2 above suggests that, an increase in speed enforcement activities during the evenings and late at night may be necessary to have an impact on the incidence of crashes involving younger drivers. 

4. No clear relationship between either exposure to enforcement or the perceived risk of detection and other speed-related attitudes has been found. This indicates that, future enforcement programs aimed at younger drivers need to do more than rely on high level of enforcement. 

5. Point 4 above is supported by the fact that younger drivers who have previously been detected speeding continue to have unsafe attitudes towards speeding behaviour. 

6. Previous MUARC research suggests that the following characteristics may be useful elements of future enforcement programs aimed at younger drivers: 

• A focus on personal (actual) risk of detection would be expected to have a greater effect on behaviour than a more general perceived risk of detection. 
• Public education programs accompanying increased enforcement could focus on those younger drivers more likely to be involved in crashes, including: males, people who drive for social reasons at night, and drivers with histories of offending. 

General Principles for Supporting Publicity 

TAC road safety publicity aimed at increasing both awareness of specific enforcement programs and general public awareness of safety issues has been integral to the implementation of many speed enforcement programs in Victoria. Such publicity is generally run to increase the general deterrence effect associated with new or existing enforcement programs. 

Results from the research relating to the effectiveness of publicity indicate a number of underlying themes: 

1. Public education campaigns should play an important role in targeting younger drivers particularly when accompanied by a significant increase in speed enforcement activities during the evenings and late at night. 

2. Targeting of drivers who drive as part of their occupation and possess the other specific characteristics (males living in metropolitan areas) should involve radio advertising. 

3. Targeted enforcement programs aimed at the identified groups of drivers who are either exposed to lower enforcement levels or believe that enforcement can be avoided must involve an increase in the actual detection of offences. 

4. The use of enforcement related advertising in future enforcement programs should increase the effectiveness of such programs and will contribute to the cost beneficial nature of those programs. 

5. In the short-term, high profile publicity can be used as an effective interim measure to reduce crash frequency when enforcement levels are low. 

First, in all instances examined, publicity supporting enforcement programs has apparently been effective in magnifying the effects of the enforcement program. These magnified effects have been evident over a range of road types. Also, the combined effect of speed enforcement programs and publicity is highly cost beneficial. This leads to the following principle: 

Principle 7: The use of enforcement-related advertising in future enforcement programs should increase the effectiveness of such programs and will contribute to the cost-beneficial nature of those programs. 

It is noted that publicity supporting speed enforcement programs is more effective in reducing crashes when it is specific to the enforcement rather than when it encompasses more general speed-related themes. 

High profile media activity has also been found to be effective in reducing casualty crashes even when enforcement levels are low. 

Principle 8: In the short-term, high profile publicity can be used as an effective interim measure to reduce crash frequency when enforcement levels are low. 

Conclusion 

MUARC's evaluations of speed enforcement to date have shown that speed enforcement initiatives conducted in Victoria in recent years have successfully contributed to reductions in road trauma. However, the current study highlights a number of areas in which further research is required to more fully understand the mechanisms at work and the effectiveness of enforcement operations in a variety of environments. In particular, it would be valuable to consider the following issues in the context of future speed enforcement research:

• The interaction effect between publicity and enforcement operations
• The duration effect of enforcement in its variety of forms 
• The use of fixed-location automatic surveillance techniques outside of a tunnel environment to determine the efficiency and effectiveness of the enforcement 
• The influence of reduced speed offence threshold levels on speed behaviour and casualty crashes 
• The relative effect of randomised scheduling of speed enforcement operations in Victoria compared with current practices. 

Further, there are a number of issues that have not been examined in detail that would add to current knowledge on speed enforcement. Specifically, issues relating to the issuance and processing of traffic offences within the administrative and judicial systems are relevant to the effectiveness of enforcement operations. The perception of these systems by road users may provide an understanding of some of the mechanisms behind effective operations. 

Finally it is noted that the effect of speed enforcement initiatives may change over time and the outcomes of this review should be updated and supplemented by future research in this area. 

---

1 Low-alcohol hours are times of the week when alcohol related crashes are less likely to occur, whereas high-alcohol hours of the week are those periods when alcohol related crashes are more likely to occur.

Sponsoring Organisation: Baseline Research Program - Department of Justice, Transport Accident Commission, VicRoads