Initiatives

Characteristics of fatal motorcycle crashes involving excessive and/or inappropriate speed

REPORT Executive summary Report

Title: Characteristics of fatal motorcycle crashes involving excessive and/or inappropriate speed

Organisation: Monash University Accident Research Centre

Author(s): Karen Stephan, Mark Symmons, Peter Hillard, Megan Bohensky, Carlyn Muir & Michael Lenné

Date: October 2008

Background
The Monash University Accident Research Centre (MUARC) was awarded a tender by VicRoads to investigate the role of speed and speeding in motorcycle crashes. Travel speed is acknowledged as a road safety risk factor – an increase in speed generally increases the likelihood of a crash due to a reduction in the time available to spot, interpret and react to a hazard. A crash at a higher speed is also likely to result in more injuries as well as an increased severity of any injury that does occur. In that sense, speed can be considered a contributory factor and the absolute travel speed at the time of the crash is the primary variable of interest.

While they are often used interchangeably in the literature, the term “excess speed” is generally used to denote travelling at a speed in excess of the prevailing speed limit and the term “inappropriate speed” indicates travelling at a speed unsuitable (or unsafe) for the prevailing conditions and road environment, or exceeding the capabilities of the driver, or exceeding the tolerances of the vehicle and its equipment. It is possible to exceed the speed limit but not be travelling at an unsafe speed on a well-engineered road, and it is possible to travel at an inappropriate speed within the speed limit. However, in many road environments, both rural and urban, a driver travelling in excess of the speed limit is also likely to be travelling at an inappropriate speed.

Speed as a crash risk factor is particularly pertinent for motorcycles and motorcycle riders. In many instances the quick acceleration and manoeuvrability, if not high speed, are features that attract many to motorcycling, especially leisure riders. But those very same features result in a smaller margin for error than might apply to most other vehicles, and the relative lack of protection offered when a crash does occur means that crash outcome is likely to be more severe for a motorcyclist involved in a crash.

Exceeding the speed limit and travelling at a speed inappropriate for the prevailing conditions can coincide or occur independently. Both increase the likelihood of being involved in a crash and increase the severity of a crash that does occur, for motorcycle riders and drivers. However, it is difficult to reliably separate these factors in crash investigations with certainty.

While previous research has determined that speed is a contributory factor in traffic crashes and that there is a relationship between speed and injury severity, few studies have investigated the proportion of crashes that involve the contribution of excessive (defined as travelling over the speed limit) and inappropriate (for the conditions) speed. The current research has addressed this question.
Aims
The aims and objectives of the study were:

To develop a method for determining the involvement of excessive and inappropriate speed from the information available in coroners records of fatal motorcycle crashes.
To apply this method to determine the proportion of motorcycle fatalities that involved excessive and/or inappropriate speed.
To apply the Safe System framework to investigate the contribution of the road-user, the motorcycle and the environment to the crash, and how these factors related to excessive and inappropriate speed involvement.

Method
Two existing data sources, the National Coroners Information System (NCIS) and the VicRoads Road Crash Information System (RCIS), were linked and used to identify 200 fatal motorcycle crashes which occurred during the period from July 2000 to December 2005. The frequency and proportion of motorcycle crashes which involved excessive and/or inappropriate speed were determined by reviewing the records for each crash. Initially, the involvement of excessive and inappropriate speed in each crash was judged from the coroner’s comments in the findings. In cases where there was no strong statement in the NCIS documents regarding the involvement of excessive or inappropriate speed, a member of the research team with experience in crash investigation judged the involvement of speed from evidence available in the crash reports, such as the length of brake marks, slide marks, ejection distance, the geometry of the crash site, reported weather conditions, knowledge of the vehicles involved, and other crash circumstances. Following this, analytical comparisons were made to determine if crashes involving excessive speed differed from crashes not involving excessive speed in terms of the rider, vehicle and environment factors noted. The same comparisons were made for crashes involving inappropriate speed.
Results
There were 115 (57.2%) crashes in which the riders’ speed relative to the speed limit could be confidently judged. Of these 115 crashes, almost half (56; 48.7%) involved excessive speed, that is, the rider was travelling faster than the speed limit. Whether or not the rider’s speed was inappropriate for the conditions was confidently judged for 135 (67.2%) crashes. Of these 135 crashes, more than two-thirds involved inappropriate speed for the conditions (95; 70.4%). The number of cases where excessive or inappropriate speed could and could not be judged is shown in Table 1.

Table 1. Binary classification of the involvement of excessive or inappropriate speed

		Excessive speed Frequency	Excessive speed Percent (% of known)	Inappropriate speed Frequency	Inappropriate speed Percent (% of known)
	Speed involved	56	27.9 (48.7)	95	47.3 (70.4)
	Speed not involved	59	29.4 (51.3)	40	19.9 (29.6)
	Total	115	57.2	135	67.2
Missing	not enough info	86	42.8	66	32.8
Total		201	100.0	201

Several individual rider, motorcycle, environment and crash characteristics were found to be related to excessive and/or inappropriate speed involvement. The magnitude of the association between these factors and speed involvement was determined using univariate logistic regression models. One of the limitations of univariate analyses is that the influence of each factor on excessive speed involvement is considered in isolation from each other. As some of the factors were correlated (e.g., marital status and age), multivariate logistic regression was conducted to simultaneously examine the relationship between a number of the potentially confounding characteristics. Of the rider characteristics, gender, marital status and rider experience were thought to be potentially correlated with age. There is also some mention in the literature that older riders are more likely to ride more powerful motorcycles, so the relationship between motorcycle engine capacity and speed involvement might also be confounded by age. As such, even though some of the univariate comparisons did not show a significant association between motorcycle engine capacity and speed involvement, this association was assessed, adjusting for age. Of the environment characteristics, there is likely to be some confounding between speed zone and location.

The association of various rider, motorcycle, environment and crash characteristics with excessive speed involvement is shown in Table 2, and for inappropriate speed involvement, in Table 3. Where necessary, potential confounders have been adjusted for.

Table 2. Factors associated with excessive speed involvement in a sample of fatal motorcycle crashes.

Risk Factor	Excessive speed involved n [%]	Excessive speed not involved n [%]	Odds Ratio [95% CI] *Adjusted odds ratio*	p-value
RIDER
Age - median [inter-quartile range]	27.5 [11]	39 [16]	0.93 [0.90, 0.97]	<0.001
Licence status
Valid motorcycle licence held No valid motorcycle licence	22 [42.3] 15 [75]	30 [57.7] 5 [25]	1 4.09 [1.29. 12.95]	0.02
Rider experience
Experienced Inexperienced	1 [8.3] 5 [45.5]	11 [91.7] 6 [54.5]	1 9.17 [0.86, 97.7] Adjusted for age 4.97 [0.33, 74.7]	0.07 0.25
Marital status
Married incl. de facto Not married	12 [27.9] 12 [54.6]	31 [72.1] 10 [45.5]	1 3.10 [1.06, 9.05] Adjusted for age 1.17 [0.29, 4.61]	0.04 0.83
Alcohol
Absent Present	34 [41.0] 18 [69.2]	49 [59.0] 8 [30.8]	1 3.24 [1.27, 8.31]	0.01
BAC = 0 BAC > 0 and <0.05 BAC> 0.05 and <0.15 BAC>0.15	34 [41.0] 4 [66.7] 11 [84.6] 3 [42.9]	49 [59.0] 2 [33.3] 2 [15.4] 4 [57.1]	1 2.88 [0.50, 16.63] 7.93 [1.65, 38.07] 1.08 [0.23, 5.14]	0.24 0.01 0.92
MOTORCYCLE
Engine capacity
<=250 cc >250 cc	7 [43.8] 22 [71.0]	9 [56.3] 9 [29.0]	1 3.14 [0.90, 11.03] *Adjusted for age* *7.37 [1.48, 36.66]*	0.07 0.02
<=500 cc >500 cc	7 [41.2] 22 [73.3]	10 [58.8] 8 [26.7]	1 3.93 [1.11, 13.85] *Adjusted for age* *10.69 [1.96, 58.37]*	0.03\ *0.006*
Riding own bike
No Yes	7 [100] 35 [44.9]	0 43 [55.1]	Cannot be computed
ENVIRONMENT
Speed zone
<=70 km/h > 70 km/h	33 [75] 22 [31.4]	11 [25] 48 [68.8]	1 0.15 [0.07, 0.36] Adjusted for location *0.31 [0.11, 0.90]*	<0.001 *0.03*
Location
City (incl. large town) Rural	41 [69.5] 15 [26.8]	18 [30.5] 41 [73.2]	1 0.16 [0.07, 0.36] Adjusted for speed zone *0.33 [0.12, 0.93]*	<0.001 *0.04*
CRASH
No. of other vehicles involved
0 1 2+	24 [63.2] 31 [44.9] 1 [12.5]	14 [36.8] 38 [55.1] 7 [87.5]	1 0.48 [0.21, 1.07] 0.08 [0.01, 0.75]	0.07 0.03
Fault of rider
Not at fault Partially at fault Fully at fault	4 [16.7] 10 [71.4] 31 [51.7]	20 [83.3] 4 [28.6] 29 [48.3]	1 12.50 [2.57, 60.7] 5.34 [1.63, 17.51]	0.002 0.006

Results in bold are those that met the conventional level for statistical significance (<0.05). Other results, where the p-value lies between 0.05 and 0.10 are also of interest).

Amongst the sample of fatally injured motorcycle riders, the following factors were significantly associated with being involved in a collision involving excessive speed:

Age: On average, as rider age increased by one year, the odds of the fatal crash involving excessive speed decreased by 7%.
Licence status: The odds of the crash involving excessive speed were increased by over three times if the rider did not have a valid motorcycle licence.
Marital status: The odds of the crash involving excessive speed were increased by over two times if the rider was not married, however, once the confounding effect of age was taken into account, this association disappeared. That is, for fatally crashed riders of a given age, marital status is not related to excessive speed involvement.
Use of alcohol: The odds of the crash involving excessive speed were increased by 224% if the rider had a positive alcohol reading. Looking at this in more depth, compared to those riders with a zero BAC, those with a BAC of between 0.05 and 0.15 were more likely to be involved in a crash involving excessive speed. However, riders with alcohol detected at levels less than the legal threshold (i.e. <0.05), and riders with very high BAC (over 0.15) did not differ from those with no alcohol detected in terms of the odds of the crash involving excessive speed.
Motorcycle engine capacity: The odds of the crash involving excessive speed were increased if the engine capacity of the motorcycle was over 250 cubic centimetres (cc), although the univariate association was not significant. However, older riders were more likely to ride more powerful motorcycles, and so a multivariate model was used, adjusting for age. For riders of a comparable age, the odds of the crash involving excessive speed were over six times higher when the motorcycle was over 250cc. Likewise, for riders of a comparable age, the odds of the crash involving excessive speed were almost ten times higher when the motorcycle was over 500cc.
Speed zone: The odds of the crash involving excessive speed were decreased in speed zones >70 km/h, compared to speed zones of 70 km/h or less. The effect of speed zone remained, even when the location (a potential confounder) was adjusted for. For a given location (i.e. urban or rural) the odds of the crash involving excessive speed decreased by 69% in speed zones >70 km/h, compared to speed zones of 70km/h or less.
Location: The odds of the crash involving excessive speed were decreased in rural areas, compared to city areas (including large country towns), even when the speed zone (a potential confounder) was adjusted for. For a given speed zone, the odds of the crash involving excessive speed decreased by 67% in rural areas compared to city areas.
Number of other vehicles involved: The odds of the crash involving excessive speed were decreased by 92% if there were two or more other vehicles involved, compared to when it was a single-vehicle collision.
Fault: The odds of the crash involving excessive speed were increased if the rider was partially (odds 11.5 times higher) or fully (odds over 4 times higher) at fault, compared to when they were not at fault.

Other factors, while not reaching the conventional level of statistical significance, may also be of interest for further research because there are indications that they might increase the likelihood of the crash involving excessive speed.

Rider inexperience, however, once the confounding effect of age was taken into account, this association disappeared.
Compared to single-vehicle collisions, the odds of the crash involving excessive speed may be decreased if there is one other vehicle involved.

Riding someone else’s motorcycle was also related to excessive speed involvement, however the magnitude of this association could not be computed. Time of day and the Defintion for Classifying Accidents (DCA) were also related. For those crashes where excessive speed could be judged, all crashes occurring between midnight and six am involved excessive speed compared to only 25% that occurred between six and nine am. In terms of DCA, a larger proportion of crashes classified as left off carriageway into object/parked vehicle and off right bend into object/parked vehicle involved excessive speed than for other crash types.

Table 3. Factors associated with inappropriate speed involvement in a sample of fatal motorcycle crashes

Risk Factor	Inappropriate speed involved n [%]	Inappropriate speed not involved n [%]	Odds Ratio [95% CI]	p-value
RIDER
Gender
Female Male	1 [25] 94 [71.8]	3 [75] 37 [28.2]	1 7.62 [0.77, 75.63] Adjusted for age 9.72 [0.96, 98.59]	0.08 0.05
Age median [iqr]	29 [16]	40 [15.5]	0.96 [0.93, 0.99]	0.007
Marital status
Married incl. de facto Not married	25 [54.4] 23 [74.2]	21 [45.7] 8 [25.8]	1 2.42 [0.90, 6.51] Adjusted for age 0.89 [0.25,3.181]	0.08 0.86
Tetrahydrocannabinol (THC)
Absent Present	64 [64.7] 19 [86.4]	35 [35.4] 3 [13.6]	1 3.46 [0.96, 12.53]	0.06
MOTORCYCLE
Engine capacity
<=250 cc >250 cc	11 [61.1] 32 [86.5]	7 [38.9] 5 [13.5]	1 4.07 [1.07, 15.50]	0.04
<=500 cc >500 cc	13 [65.0] 30 [85.7]	7 [35.0] 5 [14.3]	1 3.23 [0.86, 12.09]	0.08
Motorcycle registered
Yes No	34 [65.4] 11 [91.7]	18 [34.6] 1 [8.3]	1 5.82 [0.70, 48.78]	0.10
ENVIRONMENT
Speed zone
<=70 km/h > 70 km/h	43 [82.7] 51 [62.2]	9 [17.3] 31 [37.8]	1 0.34 [0.15, 0.80] Adjusted for location 0.52 [0.17, 1.58]	0.01 0.25
Location
City (incl. large town) Rural	57 [80.3] 38 [59.4]	14 [19.7] 26 [40.6]	1 0.36 [0.17, 0.77] Adjusted for speed zone 0.54 [0.20, 1.48]	0.009 0.23
CRASH
Cut off
No Yes	67 [74.4] 26 [60.5]	23 [25.6] 17 [39.5]	1 0.53 [0.24, 1.14]	0.10
No. of other vehicles involved
0 1 2+	34 [79.1] 58 [69.9] 3 [33.3]	9 [20.9] 25 [30.1] 6 [66.7]	1 0.61 [0.26, 1.46] 0.13 [0.03, 0.64]	0.27 0.01
Fault of rider
Not at fault Partially at fault Fully at fault	7 [28] 12 [92.3] 63 [79.8]	18 [72] 1 [7.7] 16 [20.3]	1 30.86 [3.35, 283.81] 10.13 [3.61, 28.39]	0.002 <0.00

Results in bold are those that met the conventional level for statistical significance (<0.05). Other results, where the p-value lies between 0.05 and 0.10 are also of interest).
Amongst the sample of fatally injured motorcycle riders, the following factors were associated with being involved in a collision involving inappropriate speed:

Age: As age increases by one year, the odds of the fatal crash involving inappropriate speed decreased by 4%.
Motorcycle engine capacity: The odds of the crash involving inappropriate speed were almost three times higher if the engine capacity of the motorcycle was over 250cc. Unlike for excessive speed, this relationship was not confounded by age.
Speed zone: The odds of the crash involving inappropriate speed were decreased in speed zones >70 km/h, compared to speed zones of 70 km/h or less. However, once the location of the crash (city or rural) was taken into account, this association disappeared.
Location: The odds of the crash involving inappropriate speed were decreased in rural areas, compared to city areas (including large country towns). However, once the speed zone of the crash was taken into account, this association disappeared.
Number of other vehicles involved: The odds of the crash involving inappropriate speed were decreased by 87% if there were two or more other vehicles involved, compared to when it was a single-vehicle collision.
Fault: The odds of the crash involving inappropriate speed were increased if the rider was partially (odds almost 30 times higher) or fully (odds over 9 times higher) at fault, compared to when they were not at fault.

Several other factors, while not reaching the conventional level of statistical significance, may also be of interest for further research because there are indications that they might increase the likelihood of the crash involving inappropriate speed. These factors were:

Gender (being male). This relationship became even stronger once age was adjusted for.
Not being married, however this potential association disappeared once age was adjusted for.
A positive THC result
Riding a motorcycle that was not registered
Riding a motorcycle with engine capacity greater than 500 cc.
Being involved in a crash in which the rider was not cut off by another road user.

Discussion
By combining and linking the National Coroners Information System and the VicRoads Road Crash Information System, this study has assimilated the most detailed information available for fatal motorcycle crashes that occurred in Victoria during a five year period. A wealth of information about the rider, the motorcycle, the road environment and characteristics of the crash itself was assembled. Having such comprehensive information enabled the application of the Safe System framework to investigate the contribution of these factors and how they are related to excessive and inappropriate speed involvement.

There were 115 (57.2%) crashes in which the riders’ speed relative to the speed limit could be confidently judged. Of the 115 crashes, almost half (56; 48.7%) involved excessive speed, that is, the rider was travelling faster than the speed limit. Whether or not the rider’s speed was inappropriate for the conditions was confidently judged for 135 (67.2%) crashes. Of these 135 crashes, more than two-thirds involved inappropriate speed for the conditions (95; 70.4%). It is impossible from the information provided to say with confidence whether or not excessive speed was involved in the other 85 crashes (involving 86 riders) or if inappropriate speed was involved in the other 65 crashes (involving 66 riders). There was simply not enough information provided in the coroners findings to judge speed involvement in these cases. As such, we cannot postulate whether the proportion of these cases with speed involved would be the same or different to the cases where there was enough evidence to make this judgement.

Potential solutions to reduce excessive or inappropriate speed will be multifactorial and complex and must be examined from the viewpoint of the motorcyclist rather than assuming that any measure that is effective for a car driver is bound to be effective for a motorcycle rider. Countermeasures can be grouped into three areas – those aimed at the rider, those targeting the motorcycle and associated equipment, and those focused on the road environment.

Targeted education through rider-specific anti-speeding advertising may be useful, especially to make it clear that anti-hooning legislation means that motorcycles can be impounded and forfeited. Remedial training and/or a demotion in licence status could be trialled for speeding recidivists. Pre- and post-licence training could be modified to include a greater insight component, which aims to reveal a rider’s misperceptions about their own abilities. Refresher insight courses could be required at each licence renewal and encouraged for those considering purchasing a more powerful motorcycle.
Regular enforcement blitzes focusing on speed and unsafe riding behaviours (including inappropriate speed choice) may be effective. These blitzes could also focus on other factors identified in the study as being related to speed-involved crashes, such as licence checks, random breath testing and checking that riders are not riding motorcycles too powerful for their licence conditions. In addition, the blitzes could focus on roadworthiness, particularly tyre wear. Tyre checks could be required on an annual basis before renewing motorcycle registration.

Technology, such as alcohol interlocks and intelligent speed adaptation systems (ISA), could play a large part. Future development of GPS based ISA systems could involve programming the system to warn motorcycle riders of approaching known motorcycle blackspots, corners with an unsafe camber or bad cornering line, and display advisory speeds in the approach to bends or other hazards.

The Guide to Traffic Engineering Practice: Motorcycle Safety (Part 15) (Austroads, 1999) provides guidance for design, construction and maintenance of the road and the environment to maximise motorcyclist safety and should be actively championed, particularly at local council level for minor works and installation of services. Sealing of roadside shoulders increases the margin for error, provides a buffer zone within which a recovery attempt can be made, and minimises debris and gravel on the road. The installation of perceptual countermeasures, e.g. by seemingly narrowing the approach to a corner, may also be effective in reducing speed-related crashes.

The current study highlighted the need to improve and standardise the data that is collected regarding traffic crashes in Victoria. In conducting retrospective crash investigations, however, it is difficult to collect evidence on driver behaviour (particularly human error and the pre-conditions) prior to the crash. Driver behaviour and human error are likely to be related to speed-involvement in a crash. Prospective studies should be conducted to enable the study of these important factors. Two potential study methodologies are recommended. The first is a case-control study of motorcycle crashes involving in-depth examination of the crash scene to estimate crash speed and interviews with the rider to reveal the circumstances surrounding the appropriateness of the speed at the time of the crash, to compare with a random sample of non-crashed motorcycle riders. The second is a long-term naturalistic riding study, akin to the 100 car study conducted by Virginia Tech. Naturalistic studies enable investigations to study not only crash situations where things have gone irretrievably wrong, but situations in which the road-user was able to recover from potentially disastrous situations. This approach is in line with current research in human error, in which the recovery from human error and avoidance of negative outcomes is a strong focus of research

A full copy of this report is available in the VicRoads Library at 60 Denmark Street Kew VIC 3101.

Characteristics of fatal motorcycle crashes involving excessive and/or inappropriate speed

Related Links