CRASH RISKS OF ROAD USER GROUPS IN VICTORIA

Transcription

1 CRASH RISKS OF ROAD USER GROUPS IN VICTORIA by Kathy Diamantopoulou Michael Skalova David Dyte MaxCameron MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE June 1996 Report No. 88

2 ii MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

3 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE REPORT DOCUMENTATION PAGE Report No. 88 Title and sub-title Dare ISBN June Pages x + 84 Crash Risks of Road User Groups in Victoria Author(s) Type of Report & Period Covered Kathy Diamantopoulou Michael Skalova DavidDyte Max Cameron GENERAL, 1994 Sponsoring Organisation(s) This project was funded through the Centre's baseline research program for which grants have been received from: Royal Automobile Club of Victoria (RACV) Ltd Roads Corporation (VicRoads) Ministry for Police and Emergency Services Transport Accident Commission Abstract This report presents the results of an analysis of two different data sets which were provided by VicRoads. The first of these was data collected during an exposure survey conducted by Amp Transportation Planning during 1994 on behalf of VicRoads. The second data set was a file of accident report data, originally collected by the Victoria Police during , and subsequently enhanced by VicRoads. Because the analysis made use of data prepared by others, it should be emphasised that the results depend on the validity of the original data collections. The results also depend on a range of assumptions made in the analysis. Subject to the data provided and the assumptions on which the analysis was based, this report shows some remarkable findings regarding casualty accident risks in Victoria during These new findings suggest that a new pattern of casualty accident risks may have taken shape in Victoria during the 1990's compared with earlier years. If this is the case, the implications for countermeasure development in Victoria are significant. Because of these implications, it is recommended that further research be conducted to confirm the apparent changes. Key Words: (IRRD except when marked*) Casualty Crash Risk, Driving Experience, Young Drivers, Older Drivers, Exposure, Confidence Limits Reproduction of this page is authorised Monash University Accident Research Centre, Wellington Road, Clayton, Victoria, 3168, Australia. Telephone: , Fax: CRASHRISKSOF ROADUSERGROUPSIN VICTORIAiii

4 IV MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

5 Table of Contents EXECUTIVE SUMMARY ix 1. INTR 0 D U eti ON 1 2. CRASH RISKS AND EXPOSURE EXPOSURE SURVEY DATA CRASH DATA CONFIDENCE LIMITS FOR THE RISK ESTIMATES.4 3. DRIVER CRASH RISKS IN MELBOURNE RISK ESTIMATES BY DRIVER AGE Driver Age by Driver Gender Driver Age by Vehicle Age Driver Age by Number of Passengers Driver Age by Day/Night Driving Driver Age by Time of Week Driver Age by Low/High Alcohol Times Seat Belt Usage Licence Category COMPARISON OF 1994 AND 1988 CASUALTY CRASH RISKS Risk by Driver Age Driver Age by Driver Gender Driver Age by Day/Night Driving Driver Age by Time of Week RISK ESTIMATES BY YEARS OF DRIVING EXPERIENCE Driving Experience by Driver Gender Driving Experience by Number of Passengers Driving Experience by Day/Night Driving Driving Experience by Time of Week Driving Experience by Time of Week by Driver Gender Driving Experience by Low/High Alcohol Times RISK ESTIMATES BY LICENCE TyPE Licence Type by Driver Gender Licence Type by Vehicle Age Licence Type by Number of Passengers Licence Type by Day/Night Driving Licence Type by Time of Week Licence Type by Low/High Alcohol Times RISKS DURING POTENTIAL NIGHT DRIVING CURFEW PERIODS Driver Age by Potential Night Cuifew Periods Driving Experience by Potential Night Cuifew Periods Comparison of Night Cuifew Timesfor 1988 and RISKS OF SERIOUS CASUALTY AND FATAL CRASHES Driver Age and Gender in Serious Casualty and Fatal Crashes Driver Experience in Serious Casualty and Fatal Crashes 38 CRASH RISKS OF ROAD USER GROUPS IN VICTORIA V

6 4. DRIVER CRASH RISKS FOR PROVINCIAL TOWNS IN VICTORIA RISK ESTIMATES BY DRIVER AGE Driver Age by Driver Gender Driver Age by Vehicle Age Driver Age by Number of Passengers Driver Age by Day/Night Driving Driver Age by Time of Week Driver Age by Low/High Alcohol Times Seat Be It Usage RISK ESTIMATES BY YEARS OF DRIVING EXPERIENCE Driving Experience by Driver Gender Driving Experience by Vehicle Age Driving Experience by Number of Passengers Driving Experience by Day/Night Driving Driving Experience by Time of Week Driving Experience by Low/High Alcohol Times RISK ESTIMATES BY LICENCE TYPE Licence Type by Driver Gender Licence Type by Vehicle Age Licence Type by Number of Passengers Licence Type by Day/Night Driving Licence Type by Time of Week Licence Type by Low/High Alcohol Times DRIVER CRASH RISKS FOR RURAL HIGHWAYS IN VICTORIA RISK ESTIMATES BY DRIVER AGE Driver Age by Driver Gender Driver Age by Vehicle Age Driver Age by Number of Passengers Driver Age by Day/Night Driving Driver Age by Time of Week Driver Age by Low/High Alcohol Time Seat Belt Usage RISK ESTIMATES BY YEARS OF DRIVING EXPERIENCE Driving Experience by Driver Gender Driving Experience by Vehicle Age Driving Experience by Number of Passengers Driving Experience by Day/Night Driving Driving Experience by Time of Week Driving Experience by Low/High Alcohol Times RISK ESTIMATES BY LICENCE TYPE Licence Type by Driver Gender Licence Type by Vehicle Age Licence Type by Number of Passengers Licence Type by Day/Night Driving Licence Type by Time of Week Licence Type by Low/High Alcohol Times 70 vi MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

7 6. ACCURACY AND BIAS OF DRIVER AGE 71 7 SUMMARY DIS CV SS ION CON CL USI 0 N REeo MMEND ATI0NS REFEREN CES 83 APPENDIX A. CRASH RISK ESTIMATES AND CONFIDENCE LIMITS, EXPOSURE ESTIMATES AND STANDARD ERROR ESTIMATES, CRASH INVOLVEMENT AND NUMBER OF SURVEYED DRIVERS ON MELBOURNE ARTERIAL ROADS APPENDIX B. CRASH RISK ESTIMATES AND CONFIDENCE LIMITS, EXPOSURE ESTIMATES AND STANDARD ERROR ESTIMATES, CRASH INVOLVEMENT AND NUMBER OF SURVEYED DRIVERS IN VICTORIAN PROVINCIAL TOWNS APPENDIX C. CASUALTY CRASH RISK ESTIMATES AND CONFIDENCE LIMITS, EXPOSURE ESTIMATES AND STANDARD ERROR ESTIMATES, CASUALTY CRASH INVOLVEMENT AND NUMBER OF SURVEYED DRIVERS ON VICTORIAN RURAL HIGHWAYS APPENDIX D. STANDARD ERROR METHODOLOGY FOR THE DISTANCE TRA VELLED IN AN AVERAGE WEEK APPENDIX E. METHODOLOGY USED TO DETERMINE CONFIDENCE LIMITS FOR THE CRASH RISK ESTIMATES CRASH RISKS OF ROAD USER GROUPS IN VICTORIA vii

8 V111 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

9 EXECUTIVE SUMMARY This report presents the results of an analysis of two different data sets which were provided by VicRoads. The first of these was data collected during an exposure survey conducted by Amp Transportation Planning during 1994 on behalf of VicRoads. The second data set was a file of accident report data, originally collected by the Victoria Police during , and subsequently enhanced by VicRoads. Because the analysis made use of data prepared by others, it should be emphasised that the results depend on the validity of the original data collections. The results also depend on a range of assumptions made in the analysis. Subject to the data provided and the assumptions on which the analysis was based, this report shows some remarkable findings regarding casualty accident risks in Victoria during These include: the generally higher casualty accident rate per kilometre driven by female drivers compared to males. For more severe crashes these differences tended to disappear and there were no statistically significant differences between female and male fatal crash risks the apparently high casualty accident rate per kilometre driven by drivers aged 75 and above the absence of statistically significant differences between casualty accident rates per kilometre when driving at night compared with the rate when driving during the day the substantial reductions in accident risk per kilometre driven in Melbourne for some types of drivers and circumstances between 1988 and The largest reductions appear to have been experienced by young drivers, particularly on weekends, and during both the night and day-time hours. These new findings suggest that a new pattern of casualty accident risks may have taken shape in Victoria during the 1990's compared with earlier years. If this is the case, the implications for countermeasure development in Victoria are significant. Because of these implications, it is recommended that further research be conducted to confirm the apparent changes, by repeating the 1994 exposure survey during 1996, and combining the results with the 1994 exposure data and accident data, to verify the new pattern of accident risks apparently present during the 1990's. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA ix

10 X MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

11 1. INTRODUCTION This study aimed to identify and assess the crash risk of high-risk road user subgroups, ultimately producing information for a Road Accident manual based on calculations of the relative crash risk for various combinations of vehicles and occupants in time and space. During , Monash University Accident Research Centre (MUARC) carried out an Accident Data Analysis Project with the objective of disaggregating the road accident problem using mass accident data to find groups of road users, vehicles and road segments which would be suitable targets for countermeasures (Cameron, 1992). A number of methods of analysis were developed and applied, but a general conclusion was that the analysis would be more satisfactory if appropriate exposure data was available to act as a denominator in calculations of accident involvement rates. Since then, VicRoads has conducted an exposure survey of motorised vehicle travel in Melbourne, selected country towns and on rural highways during a nonholiday period in July-August Using the VicRoads exposure survey data and data derived from Police accident reports, risk estimates (accident rates per kilometre) were calculated for all variables (and selected interactions) available in both data sets. However, an urgent need to produce risk analyses of young, old and inexperienced drivers led to this study being carried out in two stages: 1. An interim report which determined crash risk estimates for various road user groups, with emphasis on the age, driving experience and licence type of the driver. Separate risk analyses were conducted for Melbourne, Victorian provincial towns and Victorian rural highways (Diamantopoulou et ai, 1995). 2. This final report which has determined crash risks of high-risk user sub-groups as well as displaying accident rates in graphical form with confidence limits on the risk estimates, especially in cases where significant differences have been found. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 1

12 2 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

13 2. CRASH RISKS AND EXPOSURE Crash risk estimates are determined by dividing the number of crashes by the exposure, often defined as 'the opportunity to have a crash'. Further details of the concept of exposure are given by Cameron and Oxley (1995). Exposure can be measured as the distance travelled on the road, the number of trips of travel, the number of driver licence holders, the number of registered vehicles, fuel consumption or the number of people in the population. The most recent exposure survey for Victoria was conducted by Arup Transportation Planning (1995) in mid 1994 in conjunction with VicRoads. This survey estimated the distance travelled by vehicle occupants in cars and motorcycles in Melbourne, major provincial towns and major rural highways in Victoria. Hence, for this project, the distance travelled on the road in 1994 will be used as the measure of the exposure for calculation of the crash risks. 2.1 EXPOSURE SURVEY DATA The 1994 crash exposure survey estimated the distance travelled (in kilometres) for three areas of Victoria, namely metropolitan Melbourne, major provincial towns and selected rural highways during non-holiday periods in July-August Exposure information was obtained for occupants of cars and car derivatives, motorcycle riders and pillion passengers. Due to the sparseness of motorcyclists sampled, only drivers of cars and car derivatives were considered in the analyses for this report. Driver and vehicle information including driver age, sex and licence type, length of time licence was held, year of car manufacture, seat-belt usage and number of passengers was obtained by interviewers and observers at the surveyed locations. For Melbourne, drivers were interviewed on a sample of arterial roads at signalised intersections, with the estimated distance being the distance travelled along arterial roads in the samp1eab1e area, comprising 38 of the 56 Local Government Areas (pre 1994). The Victorian provincial towns sampled were Inner and Outer Gee10ng, Inner and Outer Ballarat, Inner and Outer Bendigo, Shepparton and Morwell. Drivers were interviewed on main and arterial roads at signalised intersections in these towns. Five major highways were selected as the sample of Victorian rural highways Ca1der Highway, Gou1burn Valley Highway, Princes Highway East, Princes Highway West and Western Highway, with driver interviews being conducted on these highways at permanent or temporary traffic signals. For each of the three samples, different weights were given to each driver to reflect their contribution to the total weekly distance travelled. Greater detail of the weighting procedures and the exposure survey methodology can be found in Arup Transportation Planning (1995). CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 3

14 2.2 CRASH DATA The VicRoads enhanced database derived from Police accident reports was used to obtain the crash data needed to estimate the crash risks. Casualty crashes that occurred in non-holiday periods (37 full weeks per year) during were extracted from the database. To ensure the crash data was as comparable to the exposure survey data as possible, the following selection criteria were used to extract the relevant crash data. For the Melbourne sample, drivers involved in casualty crashes that occurred on arterial roads in the same 38 Local Government Areas as the exposure survey were selected. Similarly, all casualty crashes occurring on arterial or main roads in the five provincial towns surveyed were chosen. In the exposure survey, the rural highway sample was restricted to 100 k/h and 110 k/h sections of the selected five highways between various kilometre posts. The casualty crashes were selected only for these highway sections including crashes that occurred in the intervening towns. For all three samples, only crashes involving vehicles registered in Victoria, with seating capacity less than ten seats, and the same vehicle types as those sampled in the exposure survey (cars, station wagons, utilities, panel vans and small vans) were selected. Instead of using the 1994 casualty crashes alone to estimate the 1994 crash risks, the crashes occurring in were used to increase the crash frequencies. Since the crashes occurred over a 185 non-holiday week period, but the exposure survey estimated the distance travelled in an average non-holiday week, the exposure measures were multiplied by a factor of 185 to be compatible with the crash data. 2.3 CONFIDENCE LIMITS FOR THE RISK ESTIMATES To investigate whether any apparent observed differences between the estimated crash risks were statistically significant differences, confidence limits or error bounds on the risk estimates needed to be calculated. In order to determine these confidence limits, it was necessary to find the approximate statistical distribution of the risk estimates. Methodology Suppose C denotes the number of casualty crashes and D denotes the distance travelled (or the exposure estimate), then for a particular road user group, the risk estimate, R is found by: R=C/D. 4 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

15 Assuming that the crashes follow a Poisson distribution with a mean crash frequency, m, then for m ~ 10, the crashes can be approximated by the following Normal distribution: C"" N(m,m), where an estimate of m is given by the observed number of casualty crashes for a particular road user group. The distance travelled in an average week can also be regarded as having an approximate Normal distribution as follows: where D is the estimate of the distance travelled and S. E(D) is the standard error estimate of the distance travelled. For the three separate data sources, estimation of the standard error of the distance travelled in an average week was based on the method used in Arup Transportation Planning (1995). This method calculated the mean of the site (sub-sample) exposure estimates. To take into account the different variability of the sub-samples, the exposure estimate for each site was weighted by the number of interviews at that site. An estimate of the standard error was then given by the standard deviations of the weighted sub-sample means. Further detail of the methodology can be found in AppendixD. Assuming C and D are independent and are distributed as above, then the crash risk estimate, R, can be regarded as the ratio of bivariate independent normal variables. To determine error bounds for the risk estimates, the distribution function given in Kotz and Johnson (1986) can be used, assuming that the exposure estimate is never less than zero. Further detail on the distribution function and methodological issues are given in Appendix E. Using this method, 95% confidence limits for each risk estimate were determined. The lower and upper limits together with the standard errors for the distance travelled in an average week can be found in Appendices A, B and C for the respective Melbourne, provincial towns and rural highway samples. It should be noted that the distribution function used to find the confidence could not be applied under the following conditions: limits i. the number of casualty crashes for a particular road user group was fewer than ten crashes; ii. the ratio of the standard error of the exposure estimate to the exposure estimate for a particular road user group was greater than one third; iii. the estimated distance travelled for a particular road user group was zero. In these circumstances, the lower and upper confidence limits for the risk estimates were not determined and are denoted by in the Appendices. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 5

16 The following sections of this report analyse the casualty crash risk involvement of drivers by their age, driving experience and licence type separately for the Melbourne, provincial towns and rural highway samples. Estimates of crash risks will be expressed in tables and charts as the number of casualty crashes per million kilometres travelled for the various driver groups. Crash risk estimates will be displayed in graphical form with 95% confidence limits placed on the estimates where appropriate. 6 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

17 3. DRIVER CRASH RISKS IN MELBOURNE The number of drivers involved in casualty crashes occurring on the selected Melbourne arterial roads during was 39,172, and 6,345 such drivers were surveyed during July-August On average, 38% of all casualty crashes in Melbourne occurred on the arterial road network in the 38 LGAs annually. Appendix A (Tables ALl to Al.8) gives details of the casualty crash involvement, number of drivers surveyed, exposure and risk estimates as functions of driver age and other variables including vehicle age, time of crash and sex of driver. All risk estimates are presented as the number of casualty crashes per million kilometres travelled. In addition, standard errors for the exposure estimates, and 95% lower and upper confidence limits for the risk estimates can be found in the appendix. 3.1 RISK ESTIMATES BY DRIVER AGE To assess the risk of casualty crash involvement for young drivers, risk estimates were determined for single ages, 18, 19, 20 and 21 years, whilst other driver ages were grouped into appropriate intervals. The exposure survey used the date of birth of the driver to obtain his/her age. However, this method produced only a 67% response rate for the Melbourne sample. The remaining 33% of sampled drivers had their age estimated by the interviewer or observer conducting the survey. This estimated age was recorded as an age group not as a single age and not necessarily one of the same age intervals used to group the known ages of 67% of the drivers. It was, therefore, considered inappropriate to combine the 33% of drivers with estimated ages with those of known age due to the inaccuracies involved in the age estimation. The drivers whose age was not known, contributed approximately 26% towards the total exposure (distance travelled in an average week on Melbourne arterial roads). Rather than not use this relatively large proportion of the total exposure, it was decided to distribute the distance travelled in an average week for drivers whose date of birth was not known proportionately amongst the driver age groups with known exposure as functions of other variables (such as driver gender, vehicle age, time of day and time of week). The methodological issues relating to the accuracy and bias of the driver age are discussed further in Chapter 6, and estimates of this "weighted" aggregate exposure for the 185-week period can be found in Appendix A (Table Al.8) as a function of driver age. Table 3.1 displays the casualty crash frequency, number of drivers surveyed, aggregate exposure for the 185 week period (with the distance travelled for drivers of unknown age distributed proportionately amongst driver age groups with known exposure) and the casualty crash risk estimates as a function of driver age. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 7

18 The casualty crash risk of drivers aged was relatively low (risk estimate of 0.58) compared with younger drivers. Amongst young drivers, those aged 18 and 21 had the greatest risk of 1.42 casualty crashes per million kilometres travelled, whilst drivers aged 19 and 20 years had lower risks of 1.38 and 1.00, respectively. As shown by the 95% error bars in Figure 3.0, the crash risks for drivers aged 18, 19 and 21 were significantly greater than the risks for all other age groups except for drivers aged 75+. As a combined group, drivers aged had an estimated risk of 1.27 casualty crashes per million kilometres travelled which was significantly greater than the risk for drivers aged Older drivers aged 75 years and above were the group most at risk of casualty crash involvement. Their estimated risk of 2.03 casualty crashes per million kilometres travelled was significantly greater than the risk for younger drivers. It should be noted that the risk estimate for drivers aged 75+ years is subject to greater variability because so few were involved in the exposure survey (Table 3.1), however their confidence limits take this fact into account. There was no statistically significant difference between the estimated crash risks for drivers aged 30-39, and years. These drivers also had the lowest risks amongst all age groups. For all drivers, the casualty crash risk was 0.82 (less than one casualty crash per million kilometres travelled). Table 3.1: Casual Crash Risk as a Function 0 Driver A Driver Casualty Number of Exposure Age Crash Drivers Estimate* (years) Involvement Surve ed (million km) 18 1, , , , , , , , , , , ,127 1,040 11, , , , , , , Total 39,172 6,345 47,780.9 *Exposure for drivers whose age was unknown was distributed proportionately e, Melbourne Casualty Crash Risk amongst the known driver ages. 1 Individual cells may not add to total due to the presence of drivers of unknown age and drivers aged 17 years and under. 8 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

19 Fif!ure 3.0: Casuali 2.5 CD 1.5 Co ~ CIl c:.s::: CIl l'll U ~(ij =' CIl l'll U 0.5 o Driver Age (Years) Driver Age by Driver Gender Due to the relatively small number of drivers sampled who were aged 18, 19, 20 and 21, it was decided to group these drivers into a single age group, 18-21, when considering the casualty crash risk as a function of driver age and as a function of other variables. For each age, the risk of casualty crash involvement was greater for female drivers than male drivers, with the largest female and male estimated risks occurring for drivers aged 75 and above (estimated risks of 2.94 and 1.76 for females and males, respectively). As depicted by the 95% confidence limits in Figure 3.1, the risk for females was significantly greater than the male risk for most age groups with the exception of drivers aged and For all drivers aged years, a relatively low risk was found, but the crash risk for females increased again from age 50 before an increase for males was observed. In fact, there was a statistically significant greater crash risk for females than males aged years. Table 3.2 and Figure 3.1 present the risk estimates by driver age and driver gender. Table 3.2: Casualty Crash Risks for Male and Female Drivers as a Function Driver Af!e. Melb ~ All Driver AJ,!e(Years) of CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 9

20 Figure 3.1: Casualty Crash Risk as a Function of Driver Age and Driver Gender, Melbourne 4 I [J MALE II FEMALE I Driver Age (Years) Driver Age by Vehicle Age Since the estimated risks have been calculated from casualty crashes occurring during , vehicles manufactured from 1990 onwards have not had the full period of exposure to crash. To allow for this bias, rather than weighting the weekly exposure data by 185 non-holiday weeks to cover the crash period, exposure estimates involving vehicles manufactured between were weighted by a factor of weeks whilst those manufactured between were weighted by 55.5 weeks. Casualty crash risk was greatest for drivers driving vehicles manufactured between (estimated risk of 1.71) and before 1970 (estimated risk of 1.14). Generally the risk of being in a casualty crash decreased with decreasing age of the vehicle, with the risk being significantly greater when driving a vehicle than any other vehicle for driver age groups, 18-21, and 75+ (Table 3.3 and Figure 3.2). However for the age group, 22-25, the risk of casualty crash involvement was significantly greater when driving a manufactured vehicle (and similar to that in a vehicle) than when driving one manufactured in For drivers aged years, the risk of casualty crash involvement was significantly greater when driving a vehicle than a vehicle manufactured in any other year. It is worth noting that no error bar could be placed on the pre-1970 vehicle risk for this age group since the ratio of the standard error of the exposure estimate to the exposure estimate was more than one third, thus violating one of the assumptions in section MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

21 ehicle Table 3.3: Casualty Crash Risk as a Function of Driver Age and All Driver A2e (Years) ,--~---- ~J Figure 3.2: Casualty Crash Risk as a Function Melbourne of Driver Age and Vehicle Age, 8 T IlilBEFORE ~ E ox: c 6 ~.~ 5 CD c ox: ~ 4.r: VI l!! () 3 ~ (ij :> ~ 2 () I o Driver Age (Years) Driver Age by Number of Passengers Table 3.4 and Figure 3.3 present the casualty crash risks as a function of driver age and the number of passengers travelling with the driver. For all drivers the casualty crash risk was greatest when there were at least two passengers in the car with the driver (estimated risk of 0.91). The least risk occurred when there was one passenger with the driver (estimated risk of 0.78). Generally drivers aged had casualty crash risks below 1.0 per million kilometres irrespective of the number of passengers in the vehicle. However, young drivers aged had risks above 1.2 in all categories with the greatest risk of casualty crash involvement occurring when there were at least two passengers with the young driver - a risk of 1.37 (although this risk was not statistically larger than the other risks). Older drivers (~ 75 years of age) displayed relatively high risks of at least 2 casualty crashes per million kilometres travelled, although no statistically significant differences were found between the risks as functions of number of CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 11

22 passengers. Drivers aged and had significantly larger risk estimates when travelling with two or more passengers than with fewer than two. However drivers aged were significantly more at risk of casualty crash involvement when travelling alone or with one passenger than with two or more passengers. Table 3.4: Casualty Crash Risk as a Function of Driver Age and No. of All Driver Age 5 Figure 3.3: Casualty Crash Risk as a Function of Driver Age and Number of Passeneers in Vehicle. Melbourne EJNo passengers 111passenger E:l2or more passengers E 4.><: c ~ 'E ~ 3.><: (fl if.s:: ~ o 2 ~ (ij :::J (fl <IS U 1 o Driver Age (Years) Driver Age by DaylNight Driving The risk of being involved in a casualty crash for all drivers was similar for day (6:00 am-5:59 pm) and night crashes (6:00 pm-5:59 am). Statistically there was no difference between these crash risk estimates (Tables A1.2 and A1.3 in Appendix A). Table 3.5 presents the day/night risks by driver age. An earlier study which estimated these same risks for 1988 crashes (Drummond and Yeo, 1992) found higher casualty crash risks at night than during the day. The reduced night casualty crash risk during may be a result of the reduced road toll and countermeasures such as speed cameras, increased Random Breath Testing (RBT) activity and publicity campaigns that were introduced in Victoria after MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

23 Table 3.5: Casualty Crash Risk for Day/Night Driving as a Function Driver Aee. Melb All Driver A~e (Years) of Figure 3.4: Casualty Crash Risk for Day/Night Driving as a Function Driver Aee. Melbourne of Cii E 3 III NIGHT I..>::c~ a. er: l!! '" <ll ::J ()~ 1.5.s:: EJDAV 2 Qi I t The day and night casualty crash risks decreased with age for drivers aged 18 to 49 years with day risks estimated to be marginally greater than night risks for these age groups. Generally a U-shaped trend emerged with the day and night risks increasing again as the driver age increased from 50 years upwards. The day risks were also greater than the night risks for drivers aged and 75+. Only for drivers aged were the night risks higher than the day risks. However, there were no statistically significant differences between the day and night casualty crash risks for any age group as depicted by the 95% confidence limits placed on the estimated risks in Figure 3.4. Drivers aged 75 years and above were estimated to have a relatively large night risk of 2.02 (the other night high-risk group was young drivers, aged 18-21, with the risk of casualty crash involvement at night for this group being 1.35 casualty crashes per million kilometres). CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 13

24 3.1.5 Driver Age by Time of Week The time blocks used in the crash risk analysis were the same as those presented in the exposure survey, namely: Weekday day (6am-6pm Monday-Friday); Weekend day (6am-6pm Saturday-Sunday); Weekday night (6pm-6am Sun.-Mon., Mon.-Tue., Tue.-Wed., Wed.-Thur.); Weekend night (6 pm-6am Thur-Fri, Fri-Sat, Sat-Sun). Table 3.6 depicts the casualty crash risks for each age group by time block. Drivers aged 75 years and above were most at risk in having a casualty crash on weekend days and weekend nights. The weekend risks for this age group were significantly larger than the weekday risks, with estimated weekend risks of 3.91 casualty crashes per million kilometres during the day and 4.06 during the night. Young drivers, aged 18-21, however were most at risk to have a casualty crash on weekdays (night or day), with estimated risks of 1.61 at night and 1.51 during the day. These risks were significantly larger than the weekend risks as shown by the 95% confidence limits in Figure 3.5. Drivers aged were significantly most likely to be involved in a casualty crash during weekday days than weekday nights, whilst those aged had a statistically greater risk of casualty crash involvement during weekday nights. Table 3.6: Casualty Crash Risk as a Function of Driver Age and Time of Week, Melbourne Time All Driver A2:e (Years) Weekday Weekend All 14 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

25 Figure 3.5: Casualty Crash Risk as a Function of Driver Age and Time of Week, Melbourne E..x: 7 6 ~ 5 "E a; 0.4..x: (f) ii:.c e 3 () ~ ~ 2 (f) III () liiweekday day Dweekend day IIIweekdaynight E;lweekendnight o Driver Age (Years) Driver Age by LowlHigh Alcohol Times The casualty crash risk estimates as a function of driver age and low/high alcohol times are presented in Table 3.7. The low and high alcohol times used were those presented in Harrison (1990), with low/high alcohol times corresponding largely to night/day categories. Generally, crash risks during high alcohol times appear even lower than those for the night time period (Table 3.5), with drivers aged 75+ having a significantly lower risk during high alcohol times than during low alcohol times. Details of 95% lower and upper confidence limits for these road user groups are given in Appendix A (Tables A1.2 and A2.3). Table 3.7: Casualty Crash Risk for Low/High Alcohol Times as a Function Driver Aee. Melb Alcohol All Driver A2;e (Years) of Seat Belt Usage The number of drivers surveyed not wearing a seat belt on Melbourne arterial roads was estimated to be 160 or 2.5% of all drivers sampled, and those involved in casualty crashes on arterial roads was reported to be 261 or 0.7%. However, previous studies (Fildes et ai, 1991) have found that seat belt wearing in Police reports is much greater than found in the investigation of serious casualty crashes. A 12% overreporting rate for seat belt wearing was found from the Police accident reports compared to the investigation. Hence, the calculations of casualty crash risks for seat CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 15

26 belt wearers and non-wearers unlikely to be correct. should be ignored, because their relative values are Licence Category The estimated risks were greatest for unlicensed drivers at 16.1 casualty crashes per million kilometres travelled and for learners at 2.2 casualty crashes. However these figures may be unreliable since only five unlicensed drivers and very few learner drivers were sampled in the exposure survey. Furthermore, the five unlicensed drivers were only those that admitted to being unlicensed. There were also 318 drivers recorded as being of unknown licence type. Since this data is self-reported, it could be inaccurate. Further analysis of casualty crash risks by licence type is given in section 3.4. (Appendix A gives risks and 95% confidence limits as a function of driver age and driver licence type). 3.2 COMPARISON OF 1994 AND 1988 CASUALTY CRASH RISKS An analysis of the risk of casualty crash involvement for 1988 in metropolitan Melbourne was presented in Drummond and Yeo (1992). This study used the same survey methodology as the 1994 exposure survey, however more than twice as many drivers were sampled on Melbourne arterial roads in The sampleable area consisted of the same 38 out of 56 Local Government Areas as the 1994 survey. The crashes used in the risk analysis were those that occurred in 1988 during non-holiday periods (37 full weeks). Driving experience was unavailable for the accident involved drivers when calculating accident risk estimates using the 1988 exposure survey, thus comparisons of 1994 and 1988 risk estimates will be restricted to crash risks as a function of driver age. Furthermore the survey was conducted only in metropolitan Melbourne so comparisons with crash risks for rural towns and rural highways were not possible. Comparisons will be made with the variables analysed in the 1988 study, namely driver gender, day/night, time of day and time of week, each as a function of driver age Risk by Driver Age To be compatible with the 1994 risk estimates, the distance travelled during 1988 for drivers whose age was not known was distributed proportionately amongst driver age groups with known exposure, although only 1% of drivers did not have their age recorded in the 1988 survey. For the 1988 survey, drivers were asked for their age not their date of birth, which may explain the relatively low proportion of drivers with unknown age compared to the 1994 unknown proportion. Table 3.8 presents the casualty crash involvement, 'weighted' exposure estimates and casualty crash risks for the 1988 and 1994 studies as a function of driver age using the age groupings presented in the 1988 study. The casualty crash risk for all drivers on Melbourne arterial roads has decreased since 1988 from 1.36 casualty crashes per million kilometres travelled to 0.82 crashes per 16 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

27 million kilometres in This reduction may be a consequence of the countermeasures introduced after 1989 to improve the road toll in Victoria, and may also have resulted from changes in exposure type such as purpose of travel. Across all age groups similar reductions occurred with the largest reductions occurring for young drivers aged 18, 19 and 20 years. The smallest reduction occurred for older drivers aged 60 years and above. Table 3.8: Casualty Crash Risk Estimates by Driver Age for 1988 and 1994, Melbourne Driver , , , ,302 47, , ,794 1,779 27, , , , ,673 7, , , ,684 11, , , , Casualty 1.77 Involvement(million Casualty Risk Crash Exposure* Estimate Crash km) *Exposure for drivers whose age was unknown was distributedproportionatelyamongst the known driver ages. Figure 3.6 presents the 1988 and 1994 age comparisons with the drivers aged 18, 19 and 20 years combined into one age group. Figure 3.6: Comparison of 1988 and 1994 Risk Estimates as a Function Driver A1?'e.Melbourne of 4.0 E ~ 3.0 ~ 'E [ CD Cl...l<: '" 0: 2.0.s::: '" ~ () ~ (ij :::J ~ 1.0 () Time of Week 2 Individual cells may not add to totals due to the presence of drivers of unknown age and drivers aged 17 years and under. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 17

28 A relatively large crash risk reduction has occurred amongst drivers aged since 1988, decreasing from 3.37 casualty crashes per million kilometres travelled in 1988 to 1.22 in Confidence limits have not been placed on the risk estimates in Figure 3.6 or on any other risk estimates in this section because of the unavailability of data required to calculate the 1988 limits Driver Age by Driver Gender Figure 3.7 compares the risk of casualty crash involvement for male and female drivers in 1988 and 1994 as a function of driver age. For all age groups male risks were greater in 1988 than in 1994 as were female risks. Furthermore, for each age group, females had higher casualty crash involvements than males in both 1988 and However, since 1988, accident risks for male drivers of all ages have fallen proportionately the same amount as for female drivers (a 41 % reduction for both genders). 4.0 Figure 3.7: Casualty Crash Riskfor 1988 and 1994 in Melbourne Driver Gender bv Driver A!!e 1:1Female Male t=:ifemale III Male I E ~ 3.0 ~ 'E Qi 0.. ~ Ul a: 2.0.s::: Ul e () ~ <ii ::J gj 1.0 () Driver Age (Years) Driver Age by Day/Night Driving Night time casualty crash risks have decreased since 1988 across all age groups, particularly for younger drivers aged years (Figure 3.8). However, when comparing all ages, night time driving still poses the greatest risk for young drivers in 1994 as it did in 1988 (although in 1994, drivers aged 60+ had the same night crash risk as drivers aged 18-21). Furthermore, the risk of casualty crash involvement during the day has decreased for all ages, although not as much as night crash risks. 18 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

29 4.0 Day Risks E..10:: ~ 3.0 'E Q5 Co..10:: Ul ii' 2.0..c: Ul ~ () ~ (ij ~ ~ 1.0 () Driver Age (Years) 4.0 Night Risks E..10:: ~ 3.0 'E Q5 Co..10:: Ul ii' 2.0..c: Ul ~ () ~ (ij ~ ~ 1.0 () Driver Age (Years) Driver Age by Time of Week For almost all time blocks and age groups the risk of casualty crash involvement has decreased since 1988 (Figure 3.9). The one exception was older driver (60+ years) crash involvement on weekend days. In 1988 the estimated risk was 1.00 casualty crashes per million kilometres travelled, whereas in 1994 it had increased marginally to 1.11 casualty crashes per million kilometres. The greatest reductions for drivers aged occurred on weekends during both the night and day. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 19

30 u ~ f~ 1.0 Weekend ~.1994 u El1988 'e~ ii: 15 u Co -" 'e0 E 4.0 a: 0 N Ig hi " 3.0 -".ci! 2.0 ~gj~ Weekday Weekend N Day Ighl I El1988 ElI MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

31 Weekend night-time driving was a major problem for year old drivers in 1988 with a risk of 4.16 casualty crashes per million kilometres travelled (Figure 3.10). By 1994 the weekend night risk for this age group had decreased substantially to 0.86 casualty crashes per million kilometres. There was also a large decrease in casualty crash risk during the weekend day time block for young drivers aged However, as depicted in Figure 3.10a, a general reduction in casualty crash risks has occurred for all drivers during all time blocks from 1988 to 1994, with the largest decrease between 1988 and 1994 being for weekend night crashes. 5.0 E 4.0 -" <: ~ 'E ID 3.0 c. -" '" ii:.s;;: '" l!! 2.0 () ~(ij El ] ::> '" ctl () Weekday Day Weekend Day Weekday Night Time of Week Weekend Night 5.0 E 4.0 -" <: ~ 'E ID 3.0 c. -" '" ii:.s;;: '" ~ 2.0 () ~ (ij El 1988 ::> '" ctl () Weekday Day Weekend Day Weekday Night Time of Week Weekend Night CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 21

32 3.3 RISK ESTIMATES BY YEARS OF DRIVING EXPERIENCE Driving experience is considered to be an alternative to driver age when comparing crash risks between driver groups because the focus is on new or inexperienced drivers who may not necessarily be young drivers. Much of the research in the past has focused on experience levels because, for licensing purposes, it is more usual to discriminate in terms of experience rather than in terms of age. The analyses by driving experience presented in the following section covers only drivers aged years since the accident database only included the years of driving experience for drivers in this age group. Consequently only drivers born between (a proxy group for those drivers aged years at the time of the survey) were considered in the estimates of exposure. Only for 2% of drivers surveyed was their years of driving experience not known. Because of this relatively small proportion, it was decided not to distribute the distance travelled for drivers with unknown driving experience amongst the known driving experience levels, unlike the risks presented in Sections 3.1 and 3.2 in which one third of drivers were of unknown age. Details of casualty crash risks, number of casualty crashes, exposure estimates and 95% lower and upper confidence limits as functions of driving experience can be found in Appendix A (Tables A2.l to A2.l6). Table 3.9 and Figure 3.11 present estimates of casualty crash risks as a function of driving experience. Table 3.9: Casuali Years of Driving Experience < 1 Year 1 ::;;Years < 2 2::;;Years < 3 3:::;;Years < 5 5:::;;Years < 8 ~ 8 Years Leamer* Non Victorian Lie. ~ Total Crash Risk as a Function of Driving Experience, Melbourne Casualty Number of Exposure Casualty Crash Drivers Estimate Crash Involvement Surveyed (million km) Risk 2, , , , , , , , , , ,129 1,636 12, ,814 3,301 24, *Learner driver risk is likely to be unreliable since casualty crash involvement may have been overestimated. The drivers significantly most at risk were those with the least years of driving experience, namely those who have had their licence for less than one year. The most experienced drivers (at least 8 years' driving experience) have the lowest risk of crash involvement. Statistically this risk was significantly lower than the other risks. The learner driver risk presented in Table 3.9 is likely to be unreliable, because of a coding error in the crash database which has mis-coded some non-learner drivers as learner drivers, thus overestimating the learner casualty crash involvement and 3 Individual cells may not add to total due to the presence of drivers with unknown years of driving expenence. 22 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

33 consequently the corresponding risk estimate. For this reason, learner driver risks will not be presented or discussed in the following sections. Fi!!ure 3.11: Casuall E.>0:: c: ~ E 2 :u Cl..>0:: CIl Cl: 1.5.c CIl ~ () ~ 1 <Il :::l CIl <Il () 0.5 o <1 year 1<=yr <2 2<=yr<3 3<=yr <5 Driving Experience 5<=yr<8 >=8 yr Driving Experience by Driver Gender Female casualty crash risks were generally greater than male casualty crash risks across all driving experience levels (Table 3.10). Confidence limits for these estimated risks can be found in Tables A2.2 and A2.3 in Appendix A. Male drivers with less than one year's driving experience were significantly most at risk of casualty crash involvement than males with more years of driving experience (Figure 3.12). Further, females in their second year of driving had a statistically significantly greater casualty crash risk than males with the same driving experience. This was also evident for driving experience groups, "2 ~ years < 3" and "~ 8 years", as depicted by the 95% confidence limits in Figure Table 3.10: Casualty Crash Risk as a Function of Driving Experience and Driver Gender. Melbourne Driver All S::O;yr< ::O;yr<S 2::O;yr< ~ <1 Years of Driving Experience l::o;yr<2 CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 23

34 Figure 3.12: Casualty Crash Risk as a Function of Driving Experience and Driver Gender. Melbourne 3 Cl MALE l1li FEMALE o <1 yr 2<=yr<3 3<=yr<5 Driving Experience >=8 yr Driving Experience by Vehicle Age Drivers with less than one year's experience were most likely to be involved in a casualty crash (4.23 casualty crashes per million kilometres travelled) if they were driving a vehicle manufactured between (Table 3.11). Their crash risk decreased with decreasing vehicle age. This was also generally true for drivers with more than one year's driving experience. The main exceptions were driving experience groups "3 :5; years < 5" and "5 :5; years< 8" who had relatively large risks when driving vehicles manufactured in It is worth noting that the number of drivers with less than 5 years driving experience who were surveyed driving pre 1970 vehicles were too small (under 10) to produce reliable risk estimates. Hence these figures are not presented in Table Table 3.11: Casualty Crash Risk as a Function of Driving Experience and ~ All ~yr< ~yr< ~yr< ~yr< ~ <11.38 Years of Drivin~ Experience 24 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

35 3.3.3 Driving Experience by Number of Passengers Table 3.12 and Figure 3.13 display the casualty crash risk estimates as a function of driving experience and number of passengers in the vehicle. Drivers with 1 to 2 years' driving experience who travelled with at least 2 passengers in their vehicle were most at risk of crash involvement (2.68 casualty crashes per million kilometres). This risk estimate was significantly greater than when driving alone or with one passenger for this driving experience group. Drivers who had their licence for less than one year, had similar (but relatively large) crash risks when they were travelling alone or with at least 2 passengers, however there were no statistically significant differences between these estimated risks. rs Table 3.12: Casualty Crash Risk as a Function of Driving Experience and All ~yr< ~yr< ~yr<8 ;:::8yr ~yr< <1 yr - - Years of Drivin~ Experience 4 [] No passengers 111 passenger 1:32 or more passengers o <1 year 1<=yr <2 2<=yr <3 3<=yr <5 >=8 yr Driving Experience CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 25

36 3.3.4 Driving Experience by Day/Night Driving Table 3.13 depicts the day and night casualty crash risks by the years of driving experience. For all levels of driving experience, the risk of casualty crash involvement is higher during the day than at night. However, the risk of driving in the daytime decreases more rapidly than that of driving at night for drivers with at least five years experience. New drivers were estimated to have the largest day crash risk at 2.46 casualty crashes per million kilometres travelled. Confidence limits for the estimated risks are given in Tables A2.2 and A2.3 in Appendix A. ht Table 3.13: Casualty Crash Riskfor Day/Night Driving as a Function of Drivinf! EXDerience. Melb Day vs ~yr< ~yr<8 3~yr<5 ~ ~yr<2 All < yr Years of Driving Experience Driving Experience by Time of Week Drivers who had held their licence for less than one year were most at risk of casualty crash involvement on weekday days (estimated risk of 3.18) and weekday nights (estimated risk of 3.0). Their weekend night risk was half their risk of weekday night driving. However, in comparison to drivers with more years' driving experience, night time driving on weekends for new drivers had a relatively high risk (Table 3.14). Drivers with 2 to 3 years' driving experience exhibited similar crash involvement risks when driving on weekday days and weekend nights (estimated risks of 1.6). Experienced drivers (at least 8 years driving experience) had relatively low risks during all time blocks. Details of the 95% confidence limits for these estimated risks are given in Appendix A. Table 3.14: Casualty Crash Risk as a Function of Driving Experience and -. ~ All ~yr< ~yr< ~yr< ~yr<3 ~ <1 yr ~ Years of Drivine Experience Driving Experience by Time of Week by Driver Gender Male drivers with less than one year's driving experience were most at risk of crash involvement when driving on weekday nights (estimated risk of 3.03). However 26 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

37 females with the same driving experience were most at risk during weekday days (estimated risk of 3.65). A similar pattern of results occurred for drivers with 1 to 2 year's driving experience. Females were most at risk during weekday days but males were most at risk during weekday nights (Table 3.15). In 1994, during the first year of licensed driving there was a higher risk of casualty crash during the weekday periods than during the weekend periods (night and day). This is a result of the disproportionate decrease in casualty crash risk during the weekend night periods since 1988 (section 3.2.3). Both males and females have the same high risk of 3.03 casualty crashes per million kilometres during the weekday night but female first year licence holders have an even higher risk (3.65) during the weekday days. Table 3.15: Casualty Crash Risks for Melbourne ~yr< ~yr<2 5~yr< All 2~yr< ~ < yr perience for FEMALE DRIVERS DRIVERS Years of Driving Experience for Driving Experience by LowlHigh Alcohol Times Drivers with 2 to 3 years' driving experience and those with at least 8 years' driving experience had similar casualty crash risks during low and high alcohol times (Table 3.16). All other 'driving experience' groups were more at risk of being involved in a casualty crash when driving during low alcohol times than during high alcohol times. Table 3.16: Casualty Crash Risk for Low/High Alcohol Function of Driving Experience, Melbourne Alcohol Years of Drivine: Exuerience Times as a Time <1 yr 1~yr<2 2~yr<3 3~yr<5 5~yr<8 ~8 yr All Low High All CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 27

38 3.4 RISK ESTIMATES BY LICENCETYPE In addition to driver age and years of driving experience, the type of licence held by a driver may be used to discriminate between driver groups with regard to crash risk. Table 3.17 gives the risk estimates, casualty crash involvement and exposure estimates according to the type of licence held by the driver. Only five drivers were surveyed who were unlicensed so the relatively large casualty crash risk for this group should be treated with caution. Novice and learner drivers were significantly more at risk of casualty crash involvement than drivers with full licences. However, the learner driver risk is likely to be unreliable because of the mis-coding of some non-learner driver casualty crashes as crashes involving learner drivers. Greater detail of the estimated risks, exposure estimates and 95% lower and upper limits can be found in Appendix A (Tables A3.1 to A3.8). Table 3.17: Casual Crash Risks b Driver Licence T e Melbourne Licence Casualty Number of Exposure Casualty Type Crash Drivers Estimate Crash Involvement Surveyed (million km) Risk Learner* Probationary StandardlFull 296 5,353 31, , , , Unlicensed Total 39,172 6,345 47, *Leamer driver risk is likely to be unreliable since casualty crash involvement may have been overestimated Licence Type by Driver Gender For both probationary and full licence holders, female drivers were statistically significantly more at risk of casualty crash involvement than males (Table 3.18). Details of 95% confidence limits for the risk estimates as a function of licence type are given in Tables A3.2 and A3.3 (Appendix A). Gender Table 3.18: Casualty Crash Risk as a Function of Licence Type and Driver Gender. Melbourne Driver Probationary Standard/Full Driver 1.77 All Licence Tvpe 4 Individual cells may not add to total due to the presence of drivers with unknown licence status. 28 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

39 3.4.2 Licence Type by Vehicle Age For both probationary and full licence holders, the casualty crash risks generally decreased with decreasing vehicle age (Table 3.19). However there were two exceptions. Drivers with full licences who drove pre-1970 vehicles were less likely to be involved in a casualty crash if they were driving a pre-1970 vehicle than if they were driving a or vehicle. Secondly, probationary licence holders driving a vehicle had a relatively large risk of 2 casualty crashes per million kilometres travelled. Table 3.19: Casualty Crash Risk as a Function of Licence Type and Year of Vehicle Manufacture, Melbourne Vehicle Driver Licence Type Age Probationa Standard/Full All Pre All Licence Type by Number of Passengers For both standard and probationary licence holders, casualty crash risks were estimated to be greatest for drivers who were travelling with at least two passengers (Table 3.20). However, for standard licence holders the risks were similar with no statistically significant differences between them (and relatively low compared with probationary drivers irrespective of the number of passengers in the vehicle). Conversely, probationary licence holders had a significantly greater risk when driving with at least 2 passengers than with fewer than 2 passengers, and this difference was statistically significant (Figure 3.14). s Table 3.20: Casualty Crash Risk as a Function of Licence Type and Number of Passen!!ers in Vehicle. Melb No. of Probationary StandardlFull All 1.92 Driver Licence Type CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 29

40 < 3 El No passengers II1II1passenger r;:j 2 or more passengers o Probationary ~.' <... : :':':-:-:':-:':':-: ;, :::::::::::::::::::: , ;...-..: ,... Driver Licence Type Standard/Full Licence Type by DaylNight Driving The risk of casualty crash involvement was greater during the day than at night irrespective of whether the driver held a full or probationary licence (Table 3.21), although the day risk was statistically significantly greater than the night risk for probationary drivers only. Day vs. Night Day Night All Table 3.21: Casualty Crash Risks for Day/Night Driving as a Function of Licence Type, Melbourne Driver Licence Tvne Probationa StandardlFull Licence Type by Time of Week All For drivers with probationary licences the greatest risks occurred during weekday days (estimated risk of 1.93) and weekday nights (estimated risk of 1.56). These weekday risks were statistically significantly greater than the weekend risks (for both day and night). Standard licence holders had relatively low risks during all times of the week, with the risk estimated for weekend day driving being significantly smaller than the weekday risks (both day and night). Table 3.22 gives the estimated casualty crash risks as a function of licence type and time of week. 30 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

41 Time Table 3.22: Casualty Crash Risk as a Function of Licence Type and _-.~ Probationary StandardlFull All Driver Licence Type Licence Type by Low/High Alcohol Times Table 3.23 shows the casualty crash risks during low and high alcohol times of the week by the type of licence held by the driver. Both probationary and standard licence holders were more at risk during low alcohol times than during high alcohol times. These results correspond to the day and night casualty crash risks presented in Table Table 3.23: Casualty Crash Risks for Low/High Alcohol Times as a Function of Licence Type, Melbourne Alcohol Time Low High All Probationa Driver Licence Tvne StandardlFull All RISKS DURING POTENTIAL NIGHT DRIVING CURFEW PERIODS As functions of driver age and driving experience, three possible night curfew periods have been considered: 8 pm to 6 am, 10 pm to 6 am, and midnight to 6 am As a comparison, the day (6 am to 6 pm) casualty crash risks have also been presented in the appropriate tables. The detailed results, including 95% lower and upper confidence limits, can be found in Appendix A (Tables A4.1 and A4.2) Driver Age by Potential Night Curfew Periods Table 3.25 displays the casualty crash risks for the range of possible night-time driving curfew periods as a function of driver age. As in Sections 3.1 and 3.2, in determining the risk estimates as a function of driver age, the distance travelled in an average week for drivers whose age was not known has been distributed proportionately amongst the known driver ages. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 31

42 For the youngest age group, amongst the potential night curfew periods, the greatest risk of casualty crash involvement occurred when driving between midnight and 6 am. This risk, however, was marginally smaller than the day risk, although no statistically significant differences were found. For drivers aged the midnight 6am risks were greater than the day risks which contrasts with the findings in Section 3.14 that presented the day/night crash risks. For each driver age group under 60 years, the risk of casualty crash involvement was greater during the day than at night (6 pm to 6 am). Thus driving during the early hours of the morning appears to be the most hazardous time of night for casualty crash involvement. It should be noted that for most of the younger age groups, the crash risks do not increase monotonically with the reduced duration of the night curfew periods. Generally a greater average risk occurs when driving between 8 pm and 6 am than between 10 pm and 6 am, whereas after midnight the average risk increases again. However, only for drivers aged was there a statistically greater risk when driving between midnight and 6 am than between 10 pm and 6 am (Figure 3.15). Drivers aged have a relatively low risk of 0.17 casualty crashes per million kilometres travelled during midnight to 6 am. This risk is significantly lower than the day risk as well as the risks occurring during other night curfew times. The highest average 'night' risk for this age group occurred during the longer night curfew period of 8 pm to 6 am. The night curfew risks for the oldest age group (75+) should be treated with caution because of the relatively wide confidence limits in Figure 3.14 placed on the risk estimates. There is no error bar placed on the "midnight - 6 am" risk, because there were fewer than ten drivers involved in casualty crashes who were aged 75 years and above, thus violating one of the assumptions in section 2.3. Table 3.24: Casualty Crash Riskfor Function Night Curfew Time vs. Da 8pm- 6am 10pm- 6am Midnight - 6am 6am - 60m (Day) Potential Night Curfew Periods as a of Driver Age, Melbourne Driver A~e CYears) All MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

43 Figure 3.15: Casualty Crash Riskfor Potential Night Curfew Periods as a Function of Driver A!!e for. Melbourne 2.5 E 2..I<: C ~ 'E CD 1.5 a...i<: rjl a:.s::: rjl!.l! ().2: (ij ::J rjl ltl () 0.5 1D8pm - 6am D10pm - 6am IlIIIMidnight - 6am ~ Day 6am - 6pm o Driver Age (Years) Driving Experience by Potential Night Curfew Periods ay Night Curfew There was considerable variation in the night curfew casualty crash risks as a function of years of driving experience (Table 3.26). Amongst the night curfew periods, inexperienced drivers «1 year) were most at risk of crash involvement from midnight to 6 am, although this risk was not as large as the day crash risk of 2.5 casualty crashes per million kilometres travelled. Furthermore, there were no statistically significant differences between the various night curfew periods and day risks for inexperienced drivers (Figure 3.16). Conversely, the most experienced drivers (those with more than 5 years driving experience) had smaller day crash risks than those occurring during the period of midnight to 6 am. For the driving experience groups, "5 $ years < 8" and " ~ 8 years", the risk of casualty crash involvement was greater from midnight to 6 am than at other times of the day. Table All $yr<2 5$yr<8 3.26: $yr< ~8yr $yr<5 Casualty <1 yr Crash RiskYears for Potential of Driving Night Experience Curfew Periods as a Function 0 Drivin Ex erience Melbourne CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 33

44 Figure 3.16: Casualty Crash Risk for Potential Night Curfew Periods as a Function of Drivinf! EXDerience. Melbourne 4 [l8prn 6arn D10prn 6arn _Midnight 6arn ~ Day 6arn - 6prn o <Iyr I<=yr < 2 2<= yr < 3 3<=yr< 5 5< = yr < 8 >=8yr Driving Experience Comparison of Night Curfew Times for 1988 and 1994 Decreases in casualty crash risks have occurred since 1988 for each possible night curfew period (Figure 3.17). However, the patterns in 1994 are similar to what they were in 1988 with the largest risk estimated to occur for the period midnight to 6 am. The day risk (approximately 0.8 casualty crashes per million kilometres) was similar to the risk for the period 10 pm to 6 am in However, in 1988 the day risk (1.27) was smaller than the risk estimates for each potential night curfew period (approximately 1.5 casualty crashes per million kilometres) E... " ~ E li Q.... VI a:.c: VI. 13 ~ 1ii ::l VI. U 0.5 Cl1988 _1994 o 8pm - 6am 10 pm - 6am Midnight - 6am NIGHT CURFEW PERIOD Day 6am - 6pm 34 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

45 3.6 RISKS OF SERIOUS CASUALTY AND FATAL CRASHES Driver Age and Gender in Serious Casualty and Fatal Crashes The following section examines whether the same age and gender trends that were found for drivers involved in casualty crashes (Table 3.2 and Figure 3.1) also exist for drivers involved in serious casualty and fatal crashes. Detailed results can be found in Appendix A (Tables A4.3 and A4.4). Serious Casualty Crash Risks Drivers involved in serious casualty crashes (crashes resulting in a death or serious injury) were selected for this analysis. Table 3.27 presents the serious casualty crash involvement, 'weighted' exposure estimates and crash risks per million kilometres travelled as a function of driver age. Generally, the same age trend emerged for drivers in serious casualty crashes as those in all casualty crashes, with the youngest and oldest age groups being most at risk of serious casualty crash involvement Driver 10, , , , , ,358.2 Casualty 47,780.9 (million 11, Estimate* Surveyed Number Exposure 439 1,040 6, ,696 9,207 1,854 1,048 1, ,263 Drivers Crash Serious Risk km) Crash of *Exposure for drivers whose age was unknown was distributed proportionately amongst the known driver ages. Table 3.28 presents the female and male serious casualty crash risks as a function of driver age. With the exception of drivers aged and 75+ (where female risks were significantly greater than male risks), the female serious casualty crash risks are only marginally different from the male risks (Figure 3.19). A more noticeable difference was evident for all casualty crashes (Table 3.2) where female casualty crash risks were greater than male risks across all ages. It should be noted that the relative risk of male to female drivers in the 75+ age group may be somewhat unreliable due to the relatively wide error bar placed on the female and male risks in Figure Individual cells may not add to total due to the presence of drivers of unknown age and drivers aged 17 years and under. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 35

46 Table 3.28: Serious Casualty Crash Riskfor Male and Female Drivers as a Driver All Driver40-49 A2e (Years) Figure 3.19: Serious Casualty Crash Riskfor Male and Female Drivers as a Function of Driver A1?'e.Melbourne 1.6 E 1.4.:.t: c: o 1.2 E Q; a..:.t: en a:.s:: l:l 0.8 (j ~ ili 0.6 => en '" () ~ 0.4 o.~ en 0.2 I [] MALE IiIIFEMALE I o Driver Age (Years) Fatal Crash Risks Table 3.29 presents the fatal crash involvement, exposure estimates and fatal crash risks as a function of driver age for Melbourne. Exposure is expressed in 100 million kilometres travelled, due to the relatively small number of drivers involved in fatal crashes in each age group. Similar to the risks of serious casualty and all casualty crash involvement, the risk of being involved in a fatal crash decreased with increasing age until age 50, before increasing again. Greater detail of the fatal crash risks as a function of age can be found in Appendix A. 36 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

47 Involvement Number (100 Crash 1,040 Drivers 6, Estimate* Fatal Surveyed Exposure Crash Fatal million Risk** of km) *Exposure for drivers whose age was unknown was distributed proportionately **per 100 million kilometres travelled. amongst the known driver ages. With the exception of older drivers (aged 50 years and above), male fatal crash risks were greater than female risks for all ages, although not significantly greater (Table 3.30 and Figure 3.20). This is opposite to the gender trends presented in Table 3.2 for all casualty crashes. Females were more at risk of casualty crash involvement than males across all age groups. Hence, male risks appear to increase with the increasing severity of the crash. It should be noted that no error bars could be placed on the female risks for the and 75+ age groups, since there were fewer than ten female drivers involved in fatal crashes in these age groups. Table 3.30: Fatal Crash Risk *for Male and Female Drivers as a Function Driver A1?'e.Melb All Driver Aee (Years) of *per 100 million kilometres travelled 6 Individual cells may not add to total due to the presence of drivers of unknown age and drivers aged 17 years and under. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 37

48 Figure 3.20: Fatal Crash Risk as a Function Melbourne of Driver Age and Driver Gender, 10 9 E 8 ~ c,g 7 E g 6,... I [] MALE II1IIFEMALE I ~ 5 ~ Ul ii 4.c Ul ~ () 3 (ij ~ 2 o Driver Age (Years) Driver Experience in Serious Casualty and Fatal Crashes As in the driver age analysis (Section 3.6.1), driver serious casualty and fatal crash involvement is presented as a function of driving experience in the following section. Serious Casualty Crash Risk Unlike the casualty crash risks presented in Table 3.11 where female risks were greater than male risks across all driving experience levels, there was more variation in the male/female risks as a function of driving experience for serious casualty crashes (Table 3.31). Females driving for 1-2 years, were significantly more at risk of being involved in a serious casualty crash than males, whereas males with 3-5 years' driving experience were more at risk than female drivers (Figure 3.21). All other driving experience levels had similar male and female serious casualty crash risk estimates. er Table 3.31: Serious Casualty Crash Risk as a Function of Driving Experience and Driver Gender. Melbourne Driver All ~yr<2<1 5~yr< ~yr<3 ~8yr ~yr< yr Years of Driving Experience 38 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

49 Figure 3.21: Serious Casualty Crash Risk as a Function and Driver Gender. Melbourne of Driving Experience 0.7 E 0.6 ~ c: ~ E 0.5 Qi a. ~ Cl) cc 0.4.c: Cl) l'll o ~ 0.3 a; :J Cl) l'll () 0.2 Cl) :J.Q (j; Cl) Under 1 year 1 to under 2 2 to under under unders >=Syears Years of Driving Experience Fatal Crash Risk Table 3.32 displays the fatal crash risks for male and female drivers as a function of years of driving experience. For all drivers, males are more at risk of fatal crash involvement than females (1.17 and 0.91 fatal crashes per 100 million kilometres travelled respectively). This was also found for drivers with the least «1 year) and the most (::2:3years) driving experience. However, female drivers with 1-3 years' driving experience were more at risk of fatal crash involvement than males with the same years of driving experience. As shown by the 95% error bars placed on the risk estimates in Figure 3.22, there were no statistically significant differences between the male and female fatal crash risks as a function of years of driving experience. er Table 3.32: Fatal Crash Risk* as a Function of Driving Experience and Driver Gender. Melbourne Driver All l~yr< ~yr< ~yr< ::2:8yr ~yr< < yr Years of Driving Experience *per 100 million kilometres travelled CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 39

50 Figure 3.22: Fatal Crash Risk as a Function of Driving Experience and Driver Gender. Melbourne 5 E 4.,.; c: ~ "E g 3... Q) Co.,.; f/l ii.s:: 2 f/l ~ () 19 ctl LL o Under I year I to under 2 2 to under 3 3 to under 5 5 to under 8 >=8yrs Years of Driving Experien.:e 40 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

51 4. DRIVER CRASH RISKS FOR PROVINCIAL TOWNS IN VICTORIA A total of 4,448 drivers were involved in casualty crashes that occurred on arterial and main roads in the sampled provincial Victorian towns during The total number of casualty crashes occurring in the sampled towns during the five year period was 11,755. Thus 62% of the casualty crashes did not occur on main or arterial roads. Extensive details of all casualty crash numbers, number of drivers surveyed, exposure estimates, casualty crash risk estimates and 95% confidence limits on the risk estimates can be found in Appendix B by driver age, driving experience and driver licence type. 4.1 RISK ESTIMATES BY DRIVER AGE Since almost half of the estimated distance travelled in an average week in rural towns was for drivers whose full date of birth was not known, it was decided to distribute this 'unknown age' exposure proportionately amongst the driver age groups with known exposure. Methodological issues regarding the accuracy and bias of the 'weighted' exposure estimates as a function of driver age are discussed further in Chapter 6. The 'weighted' exposure estimates are presented in Table B 1.8 (Appendix B) as a function of driver age and other variables. Casualty crash risk estimates, together with the number of drivers involved in casualty crashes, 'weighted' aggregate exposure estimates (for a 185 non-holiday week period) and the number of drivers surveyed are presented in Table 4.1 as a function of driver age. Risk estimates for drivers aged years are presented for individual ages in this group. Table 4.1: Casualty Crash Risk as a Function of Driver Age for Provincial Towns in Victoria Driver Involvement CrashDrivers Casualty , Number 3, Surveyed (million Estimate , Casualty Exposure 1.21 Crash Risk of km) * *Exposure for drivers whose age was unknown was distributed proportionately amongst the known driver ages. 7 Individual cells may not add to total due to the presence of drivers of unknown age and drivers aged 17 years and under. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 41

52 The high risk age groups in provincial towns are young and old drivers. Drivers aged 75+ years are most at risk of casualty crash involvement (estimated risk of 9.34 casualty crashes per million kilometres travelled). However, this estimate may be somewhat unreliable, as depicted by the relatively wide 95% error bar in Figure 4.0. Nevertheless, drivers aged 75 years and above were significantly more at risk of casualty crash involvement than drivers of all other ages. Young drivers (aged 18 to 21 years) had similar high casualty crash risk involvements (over 2 casualty crashes per million kilometres travelled). Amongst these young drivers, those aged 19 had the highest risk, however no statistically significant differences were found. Generally the risks decrease with increasing driver age until about 50 years. The risks begin to increase again for older drivers. For all drivers in provincial towns the risk was estimated to be 1.21 casualty crashes (95% confidence limits ranging from 0.90 to 1.85) per million kilometres travelled which was larger than the corresponding Melbourne estimate of 0.82 (95% confidence limits ranging from 0.76 to 0.89). 16 Figure 4.0: Casualty Crash Riskfor Victorian Provincial Towns as a Function of Driver A1?'e E c ~ 14..le: 12 "E 10 Q; Co..le: if 8.s::: Ul!! o 6 ~ (ij :J ~ 4 o 2 o Driver Age (Years) The following risk estimates as a function of driver age and other variables, combine drivers aged 18, 19, 20 and 21 years into a single age group because of sample size constraints for individual ages in this group of drivers Driver Age by Driver Gender As for the Melbourne sample, female drivers are more at risk of casualty crash involvement than males in rural towns (Table 4.2 and Figure 4.1). This was also true across all age groups, however the female estimated risks were significantly greater than the male risks only for drivers aged and MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

53 It should be noted that since only 5 female drivers aged 75+ were surveyed in provincial towns, their risk estimate is likely to be unreliable, so has not been presented in Table 4.2 or Figure 4.1. Table 4.2: Casualty Crash Risk for Male and Female Drivers as a Function of Driver Driver A~e60-74 CYears) All 75+ Figure 4.1: Casualty Crash Risk for Male and Female Drivers as a Function Driver Aee for Provincial Towns in Victoria of if ::J 14..ll: 6B..ll: e ()~ '" ell c-.c a; 4 DMALE IIIIIFEMALE I ~ 10 2 "E Q; 12 0 I Driver Age by Vehicle Age For all drivers in rural towns, the risk of casualty crash involvement increased as the age of the vehicle driven in the crash increased (Table 4.3). This trend did not occur for the younger drivers, where the risk of casualty crash involvement was relatively large when driving a vehicle manufactured in (Figure 4.2). Drivers aged were significantly more at risk of casualty crash involvement when driving a vehicle than a vehicle manufactured in The highest risk group driving pre-1970 vehicles was drivers aged years (estimated risk of 4.05 casualty crashes per million kilometres). However, this risk estimate was not significantly greater than the risks estimated when driving vehicles manufactured after 1970 for this age group. Drivers aged years, had the lowest risk amongst those driving vehicles. For this age group the risk of casualty crash involvement was significantly less when driving a vehicle than one manufactured earlier. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 43

54 It should be noted that recently manufactured vehicles may not have had the full period of exposure to crash, so have been weighted accordingly as in section Table 4.3: Casualty Crash Risks as a Function of Driver age and Year of Vehicle All nla Driver A2e (Years) Figure 4.2: Casualty Crash Risk as a Function of Driver Age and Vehicle Age for Provincial Towns in Victoria 20 E 16 -" c: I [] BEFORE !:l J;;;J I ~ 'E ~ 12 -" (fj a:.c (fj ell <:5 8 ~ ai ::3 (fj ell () 4 o Driver Age (Years) Driver Age by Number of Passengers Table 4.4 and Figure 4.3 display the risk estimates for rural towns as a function of driver age and by the number of passengers travelling with the driver. Overall, the greatest risk of casualty crash involvement occurs with at least two passengers in the vehicle, and the lowest risk when there is one passenger with the driver. Young drivers (aged years), however, have similar non-statistically significant risks whether driving alone, with one passenger or with 2 or more passengers (approximately 2.4 casualty crashes per million kilometres). The risk for drivers aged increased with increasing number of passengers, however no statistically significant differences were found. Conversely drivers aged had a significantly higher casualty crash risk of 2.68 when travelling with at least two passengers than with fewer than two. Drivers aged 75+ years had a relatively large risk 44 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

55 (approximately 13 casualty crashes per million kilometres travelled) when driving alone. This risk was significantly greater than the risk estimated when travelling with one passenger. Table 4.4: Casualty Crash Risk as a Function of Driver Age and Driver Age (Years) 50-59All Figure 4.3: Casualty Crash Risk as a Function of Driver Age and Number of Passenflers in Vehicle for Provincial Towns in Victoria E.><: c: 18 o ~ 16 E ~ 14.><: ~ 12.t:: ~ 10 (,) ~ 8 iii ~ ~ 6 (,) 4 c:l No passengers l1li1 passenger E:l2 or more passengers 2 o Driver Age (Years) Driver Age by Day/Night Driving For all drivers, the risk of casualty crash involvement for rural towns was greater when driving during the day than at night (Table 4.5). This was also found for drivers aged 22-49, although only drivers aged had a significantly larger risk when driving at night than during the day (Figure 4.4). Conversely, night casualty crash risks were larger than the estimated day risks for young drivers aged and older drivers aged 50+, however no statistically significant differences were found. It should be noted that the estimated risk of 19.7 casualty crashes per million kilometres travelled at night for drivers aged 75+ may be unreliable since only three drivers in this age group were surveyed at night. No error bar has been placed on this risk estimate due to the violation of one of the assumptions in Section 2.3. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 45

56 Table 4.5: Casualty Crash Risk for Day/Night Driving as a Function Driver Age (Years) All of Driver Age (j >:: c: 14 Vl a. () as 86..>:: oc 4.J::. Cii 'E ~ Q; E ::3 Vl Figure 4.4: Casualty Crash Risk for Day/Night Driving as a Function Driver Aee for Provincial Towns in Victoria of [ E:lDAY IiIINIGHT ] Driver Age (Years) Driver Age by Time of Week For all drivers in rural Victorian towns, the highest risk of casualty crash involvement occurs when driving on weekend days and the least on weekday nights (Table 4.6). Statistically, the weekend day risk was significantly greater than the weekday night risk. This result also occurred for drivers in most age groups (Figure 4.5). In contrast, Melbourne drivers were found to be most at risk during weekday days and weekday nights. Young drivers in rural towns aged years had the highest weekend day casualty crash risks amongst all age groups (almost 8 casualty crashes per million kilometres). The weekend day risk was significantly greater than the risks occurring on weekday days and weekend nights. For drivers in the age group 26-59, the weekend day casualty crash risk was significantly greater than the risks estimated for all other times of the week. It should be noted that only the weekday day and weekday night risks could be estimated for the 75+ age group, so risks for weekend travel for this age group are not presented in Figure 4.4. Details of each risk estimate, together with the corresponding 95% lower and upper limits are given in Appendix B. 46 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

57 ime Table 4.6: Casualty Crash Risk as a Function of Driver Age and Time of Week - ~ ~ All Driver 50-59Aj!e(Years) Figure 4.5: Casualty Crash Risk as a Function of Driver Age and Time of Week 'or Provincial Towns in Victoria ~ c:.2 ~ 8 CD a. ~ mcc 6 IIlWEEKDAY OWEEKEND IliIWEEKDAY i:ilweekend DAY DAY NIGHT NIGHT.c m <ll U 4 <ll :J m <ll () 2 o Driver Age (Years) Driver Age by Low/High Alcohol Times The low and high alcohol times used in the Melbourne analysis (Section 3.1.6) were also used for provincial towns since these alcohol times are estimates for all of Victoria. For most age groups driving during low alcohol times posed a greater risk of casualty crash involvement than during high alcohol times, although statistically significant differences were found only for drivers aged and (Table 4.7 and Appendix B). Older drivers, aged 75+, had a considerably large risk of crashes per million kilometres travelled during high alcohol times of the week, however this estimate should be treated with caution due to the large width of its 95% confidence interval (in the range of 8.02 to 57.18). CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 47

58 Alcohol Table 4.7: Casualty Crash Risk for Low/High Alcohol Times as a Function All Driver 50-59A~e (Vears) J of Seat Belt Usage Only 1.1% (or 40 drivers) of all drivers surveyed in Victorian rural towns and only 0.5% of drivers involved in casualty crashes on main roads were not wearing a seat belt. Risk estimates for seat belt usage, therefore, are not presented, although Appendix B contains the number of drivers surveyed, number of casualty crashes and exposure estimates by seat belt usage. 4.2 RISK ESTIMATES BY YEARS OF DRIVING EXPERIENCE Similarly to the Melbourne sample, only drivers aged years were used to estimate casualty crash risks as a function of driving experience because the crash database did not provide the years of driving experience for drivers over 40 years of age. Furthermore, the driving experience variable was not available for the 1994 rural town crash data, so only casualty crashes occurring during non-holiday periods for are used in the following analyses. Applying these restrictions, a total of 2,340 casualty crashes occurred in the selected rural towns with the number surveyed restricted to 2,015 drivers born during (a proxy group for drivers aged years). Table 4.9 and Figure 4.6 display the casualty crash risk estimates as a function of driving experience for rural towns in Victoria. Generally, the risk of casualty crash involvement decreased with increasing years of driving experience. Drivers who had held their licence for less than one year were estimated to have almost 4 casualty crashes per million kilometres travelled. However the least experienced group was not significantly more at risk of crash involvement than drivers with 1 to 5 years of driving experience. 48 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

59 of Table 4.9: Casualty Crash Risk as a Function of Driving Experience ~ (million Number Drivers 1, CasualtyCrash 1,483.2 Estimate , CrashInvolvement Exposure Surveyed Risk 2,340 km) of *Learner driver risk is likely to be unreliable since casualty crash involvement may have been overestimated. Figure 4.6: Casualty Crash Risk as a Function of Driving Experience or Provincial Towns in Victoria 6 5 o < 1 YEAR 1 <= YEAR< 2 2 <= YEAR< 3 3 <= YEAR< 5 5 <= YEAR< 8 Driving Experience >=8 YEAR Driving Experience by Driver Gender Female drivers were more at risk of casualty crash involvement than males in provincial Victorian towns across all driving experience levels (Table 4.10). However, the female estimated risk was only significantly greater than the male risk for drivers with 3 to 5 years of driving experience (Tables B2.2 and B2.3 in Appendix 8 Individual cells may not add to total due to the presence of drivers with unknown years of driving experience. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 49

60 B). Casualty crash risks for learner drivers are not presented in Table 4.10 (and in subsequent tables in this section) due to there being only 10 learner drivers surveyed in rural Victorian towns and the possible overestimation of their casualty crash involvement. der Driver Table 4.10: Crash Risk as a Function of Driving Experience and Driver Gender All ::;yr< ::;yr<5 5::;yr< ~8 1::;yr< <1 5.10yr Years of Driving Experience Driving Experience by Vehicle Age The estimated risks involving driving experience for rural towns have been calculated from casualty crashes occurring during (a 148 non-holiday week period). Vehicles manufactured from 1990 onwards would therefore have not had the full period of exposure to crash. To allow for this possible bias, weekly exposure estimates involving vehicles manufactured between were weighted by a factor or weeks and those manufactured between by a factor of 24.7 weeks instead of 148 non-holiday weeks. For all drivers, the estimated casualty crash risks decreased with decreasing vehicle age. This was also generally the case for the most experienced drivers, those with at least 8 or those with 5 to 8 years of driving experience. For drivers with less than one year's driving experience, however, the estimated casualty crash risk increased with decreasing vehicle age, with a relatively high risk estimate of 15.3 occurring when driving a vehicle. Drivers who had held their licences for between 1 to 2 years, 2 to 3 years and 3 to 5 years also exhibited high casualty crash risks in vehicles manufactured in (Table 4.11). Table 4.11: Casualty Crash Risk as a Function of Driving Experience and n ::;yr< ::;yr< ~8 2::;yr< All 1::;yr< < Years of Driving Experience Driving Experience by Number of Passengers In rural Victorian towns, drivers with less than one year's driving experience were most at risk of casualty crash involvement when driving alone - although a similar risk occurred with one passenger in the vehicle. No statistically significant differences existed between the estimated risks for this driving experience group. 50 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

61 Drivers in their second year of driving were significantly less at risk of casualty crash involvement when travelling with at least two passengers than when driving alone or with one passenger. For the other driving experience levels, the estimated risks were greatest when the driver was travelling with at least two passengers, although not significantly greater than the other risks (Table 4.12 and Figure 4.7). Table 4.12: Casualty Crash Risk as a Function of Driving Experience and Number o Passen ers in Vehicle or Provincial Towns in Victoria No. of Years of Drivin Ex erience Passen ers <1 yr 1$yr<2 2$yr<3 3$yr<5 5$yr<8 None One ~ Two All ~8 yr All [J No passengers 111 passenger E:l2 or more passengers 2 o <1 year 1<=year<2 2<=year<3 3<=year<5 Driving Experience 5<=year<8 >=8 years Driving Experience by DaylNight Driving Similar results to the Melbourne sample were obtained for the day/night driving casualty crash risks in rural towns as a function of driving experience (Table 4.13). The risk of casualty crash involvement was higher during the day than at night for all driving experience levels, with drivers in their second year of driving having a significantly greater day crash risk than night risk. Tables B2.2 and B2.3 in Appendix B give the 95% confidence limits for the risk estimates. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 51

62 t Table 4.13: Casualty Crash Risk for Day/Night Driving as a Function --~ All ~yr< ~8 3~yr<5 l~yr<2 2~yr< < yr - - Years of Driving Experience of Driving Driving Experience by Time of Week Across all levels of driving experience, driving during weekend days in rural towns produced the greatest casualty crash risks (Table 4.14). Weekday night driving also posed a relatively large risk of 4 casualty crashes per million kilometres travelled for drivers with less than one year's driving experience. Estimates of casualty crash risks together with the corresponding 95% lower and upper confidence limits by driving experience, time block and driver gender can be found in Tables B2.9, B2.1 0 and B2.11 (Appendix B). Table 4.14: Casualty Crash Risk as a Function of Driving Experience and " ~yr< ~yr< l~yr< ~yr< All ~ <1 yr Years of Drivin2 Experience Driving Experience by LowlHigh Alcohol Times Generally the low alcohol time and high alcohol time risks correspond to the day and night driving risks of Table All drivers of varying driving experience levels in Victorian rural towns have higher casualty crash risks during low alcohol times than at high alcohol times (Table 4.15). However, the low alcohol time risks were not significantly greater than high alcohol time risks for any driving experience group (Tables B2.2 and B2.3 in Appendix B). Table 4.15: Casualty Crash Risksfor Low/High Alcohol Times as a Function of Driving Experience for Provincial Towns in Victoria Alcohol Years of Driving Experience Time <1 yr l~yr<2 2~yr<3 3~yr<5 5~yr<8 ~8 yr All Low High All MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

63 4.3 RISK ESTIMATES BY LICENCE TYPE Risk estimates (expressed as number of casualty crashes per million kilometres travelled) for provincial towns in Victoria are presented in Table 4.16 as a function of driver licence type. Learner and probationary licence holders were most at risk of casualty crash involvement in rural towns with estimates of 3.25 and 2.11 casualty crashes per million kilometres respectively, although the learner risk is likely to be unreliable because the learner driver casualty crash involvement may have been overestimated. The risks between probationary and full licence holders, however, were not significantly different (Tables B3.2 and B3.3 in Appendix B). Table 4.16: Casualty Crash Riskfor Victorian Provincial Towns by Driver Licence T' Licence (million , ,661.9 Estimate 3.25 Number Drivers Casualty Exposure , , Crash Surveyed 4,448 Risk40 km) 3, of *Leamer driver risk is likely to be unreliable since casualty crash involvementmay have been overestimated, In the following sub-sections, learner driver casualty crash risks will not be presented due to the few number of learner drivers sampled and the unreliability of their crash involvement in Victorian provincial towns. Details of learner driver risk estimates, however, can be found in Appendix B Licence Type by Driver Gender Female drivers were more at risk of casualty crash involvement in Victorian provincial towns than males whether they were novice drivers or held full licences (Table 4.17), although no statistically significant differences were found. Table 4.17: Casualty Crash Risk as a Function Probationary StandardlFull All Driver Licence Tvpe of Licence Type and Driver Gender 9 Individual cells may not add to total due to the presence of unlicensed drivers and drivers with unknown licence type. CRASH RISKS OF ROAD USER GROUPS IN VICfORIA 53

64 4.3.2 Licence Type by Vehicle Age Table 4.18 presents the risk of casualty crash involvement for drivers in rural towns as a function of their licence type and vehicle age. Drivers with standard licences were less at risk of being in a casualty crash if they were driving recently manufactured vehicles. In contrast, probationary licence holders were estimated to have the highest casualty crash risk when driving vehicles, a risk of 15.1 casualty crashes per million kilometres travelled (although only nine probationary licence holders were surveyed in rural towns). This risk was significantly larger than the risks involving vehicles manufactured prior to Table 4.18: Casualty Crash Risk as a Function Vehicle Age Pre All of Licence Type and Year of Vehicle Manufacture for Provincial Towns in Victoria. Driver Licence Type Probationa Standard/Full All Licence Type by Number of Passengers s No. of Probationary licence holders in rural towns had similar risks of approximately 2 casualty crashes per million kilometres travelled whether they were driving alone or travelling with passengers (Table 4.19). Drivers with standard licences were most at risk when there were at least two passengers in the vehicle. Table 4.19: Casualty Crash Risk as a Function of Licence Type and Number of rassenf! ers m vemcle Standard/Full Probationary 2.13.for 2.12 All rrovmczal Driver 1'Ownsm Licence VICtorza Type Licence Type by DaylNight Driving Probationary and standard licence holders had larger estimated casualty crash risks for day rather than night driving in Victorian rural towns (Table 4.20). However, statistically, for both driver types the day/night crash risks were not significantly different. 54 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

65 Table 4.20: Casualty Crash Riskfor Day/Night Driving as a Function of Licence Type for Provincial Towns in Victoria Day vs. Ni~ht Day Night All Probationa Driver Licence Type Standard/Full All Licence Type by Time of Week Drivers with standard licences were most at risk of casualty crash involvement when driving on weekend days in rural towns and least at risk on weekday nights (Table 4.21). Probationary licence holders were also most at risk during weekend days, although they had similar casualty crash risks during other times. Their weekend day risk was significantly larger than the risk estimates for other time blocks. Table 4.21: Casualty Crash Risk as a Function of Licence Type and Time of Week 'OrProvincial Towns in Victoria Time Block Weekday Day Weekend Day Weekday Night Weekend Nil:!:ht All Probation Licence Type by LowlHigh Alcohol Times Driver Licence Tyne Standard/Full All The risk of casualty crash involvement was greater when driving during low alcohol times of the week than during high alcohol times for both novice drivers and for drivers with full licences (Table 4.22). However, the differences in the risk estimates were not statistically significant (Appendix B, Tables A3.2 and A3.3). These trends are similar to the ones found for the Melbourne sample of drivers. Table 4.22: Casualty Crash Risks for Low/High Alcohol Times as a Function 0.( Licence Type for Provincial Towns in Victoria Alcohol Time Low High All Probationa Driver Licence Type Standard/Full All CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 55

66 56 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

67 5. DRIVER CRASH RISKS FOR RURAL HIGHWAYS IN VICTORIA The total number of casualty crashes that occurred during on the selected highway sections of rural Victoria was 1,794. This amount includes the casualty crashes that happened in towns along the highways, since without these crashes the accident frequency would have been reduced by more than half to 870 crashes. The number of drivers surveyed on the selected highways was 5,981. Appendix C gives details of casualty crash frequencies, number of drivers, exposure estimates, risk estimates and 95% confidence limits for the risk estimates as functions of driver age, driving experience and driver licence type. 5.1 RISK ESTIMATES BY DRIVER AGE As had occurred for the Melbourne and rural towns samples, the distance travelled in an average on rural highways by drivers whose full date of birth (and hence age) was not known (76% of total rural highway exposure) was distributed proportionately amongst the driver age groups with known exposure. Table C1.8 (Appendix C) gives the 'weighted' exposure estimates as a function of driver age and other variables. Methodological issues relating to the accuracy and bias of driver age are discussed further in Chapter 6. Table 5.1 presents estimated risks (casualty crashes per million kilometres), number of drivers involved in casualty crashes, number of drivers surveyed and 'weighted' aggregate exposure estimates for the selected rural highways of Victoria. The groups at most risk of casualty crash involvement were drivers aged 75+ and 21 years (estimated risks of 3.21 and 1.32 respectively). This was also true for the Melbourne and provincial towns samples. As can be seen in Figure 5.0, the estimated casualty crash risks generally decreased with increasing age until age 50 when the risks increased again. Drivers aged 18, 19, and 21 years were significantly more at risk of casualty crash involvement than older drivers. However, drivers aged 20 years, had a significantly smaller risk estimate than drivers aged 18,19, or 21 years. The oldest age group of drivers, those aged 75 years and above, were significantly more at risk of casualty crash involvement than drivers of other ages. However the risk estimate for the oldest age group of drivers, may be somewhat unreliable due to its relatively wide 95% confidence interval. Overall for drivers on rural highways, the casualty crash risk was lower (0.46 casualty crashes per million kilometres travelled) than the corresponding Melbourne and provincial town risks (0.82 and 1.21 respectively). However, if the crashes that occurred in towns along the selected highways were not included, then an even lower (and possibly more accurate) estimate of 0.22 casualty crashes per million kilometres of rural highway travel is produced. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 57

68 Driver Table 5.1: Casualty Crash Risk as a Function Rural Hif!hwavs in Vict,. InvolvementCrash Casualty CrashDrivers , , Number 5, Surveyed (million Estimate* 3, Casualty Exposure Risk of km) of Driver Age for *Exposure for drivers whose age was unknown was distributed proportionately Figure 5.0: Casualty Crash Risk as a Function Rural Hif!hwavs in Victoria amongst the known driver ages, of Driver Age for 5 E 4..>:: c ~ 'E Qi 3 c...>:: Ul a:.s:: Ul () ~ 2 ~ (ij :J Ul <tl () 1 o Driver Age (Years) Driver Age by Driver Gender Female drivers on rural highways in Victoria had a significantly higher risk of casualty crash involvement than males (Table 5.2). This finding agrees with the Melbourne and rural town results presented in the earlier sections. The female risks were also greater than the male risks for each age group. Statistically significant 10 Individual cells may not add to total due to the presence of drivers of unknown age and drivers aged 17 years and under. 58 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

69 differences between the male/female risks were found for drivers aged 18-21, 22-25, and (Figure 5.1). The largest risk occurred for older female drivers aged 75+ (4.36 casualty crashes per million kilometres), although this risk was not significantly greater than the corresponding male risk. Table 5.2: Casualty Crash Riskfor Male and Female Drivers as a Function of All Driver 50-59Age (Years) - Figure 5.1: Casualty Crash Riskfor Male and Female Drivers as a Function of Driver Aee for Rural Hiehwavs in Victoria 9 8 I El MALE II1II FEMALE I o Driver Age (Years) Driver Age by Vehide Age With the exception of vehicles, the risk estimates generally decreased with decreasing vehicle age for most age groups. Similar findings occurred for the Melbourne and rural town samples. The risk estimates involving vehicles are particularly large for the youngest (18-21, 22-25, 26-29) and the oldest (75+) age groups (Table 5.3). Drivers aged and were significantly less likely to be involved in a casualty crash if driving a vehicle, whereas drivers aged and had a significantly larger risk when driving a vehicle than any other type of vehicle. Appendix C (Tables C1.2 and C1.3) give details of 95% lower and upper limits for each estimated risk. Age risk estimates involving pre-1970 vehicles are likely to be unreliable due to the relatively small number of drivers surveyed, and the unavailability of confidence limits for the corresponding risks because of the violation of the assumptions in CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 59

70 Section 2.3. Similarly, no confidence limits have been estimated for the risks involving vehicles for drivers aged and 60-74, thus placing doubt on the reliability of these risks. Table 5.3: Casualty Crash Risk as a Function of Driver Age and Year of Driver 50-59A~e (Years) All Driver Age by Number of Passengers For all drivers on rural highways, casualty crash involvement was greatest when driving alone, whereas for Melbourne and provincial towns the highest risk occurred with one passenger. However, drivers aged years driving on rural highways were most at risk when travelling with one passenger (Table 5.4). Statistically, however, this risk estimate was not significantly larger than the risks occurring when driving alone or with at least two passengers for this road user group. Conversely, drivers aged were statistically less at risk of casualty crash involvement when travelling with one passenger than when driving alone or with two or more passengers. It should be noted that the risk estimates for drivers aged 75+ may be unreliable due to their relatively wide 95% confidence limits (Appendix C, Tables C1.2 and C1.3). Table 5.4: Casualty Crash Risk as a Function of Driver Age and Number 0 Passen ers in Vehicle or Rural Hi hwavs in Victoria No. of Driver A e (Years Passen ers All None One ~ Two All Driver Age by Day/Night Driving Table 5.5 and Figure 5.2 display the casualty crash risks for day and night driving as a function of driver age on rural highways. The risk of casualty crash involvement for all drivers was greater at night than during the day when driving on rural highways in Victoria, although not significantly greater. However, for young drivers aged and older drivers aged 50+, day risks were greater than night risks, although only for drivers aged was the day risk 60 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

71 (1.36 casualty crashes per million kilometres) significantly greater than the night risk (0.85 casualty crashes per million kilometres). Conversely drivers aged were significantly more at risk of casualty crash involvement when driving at night than during the day. The largest night risk occurred for drivers aged 75+ (1.59 casualty crashes per million kilometres travelled at night). However, this risk estimate is likely to be unreliable as fewer than ten drivers aged 75+ were involved in casualty crashes on rural highways at night, and hence no error bar could be placed on this risk estimate. The opposite trends were found for Victorian provincial towns, where day time driving posed a greater risk than night time driving for most age groups with the exception of drivers aged and 50+, whilst for the Melbourne sample, day risks were generally greater than night risks for drivers of all ages. Table 5.5: Casualty Crash Risk for Day/Night Driving as a Function or Rural Highways in Victoria Day vs. Driver Age CYears) Night Day Night All of Driver Age All Figure 5.2: Casualty Crash Risk for Day/Night Driving as a Function Driver Aee for Rural Hiehwavs in Victoria of 6 5 E -'<: c:.q ~ 4 ~ C. -'<: Cl) a: 3.s::: Cl) l!! () 2 ltl ::J Cl) ltl () [ ClDAY IIIINIGHT ] o Driver Age (Years) CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 61

72 5.1.5 Driver Age by Time of Week In contrast to Melbourne and Victorian provincial towns, weekend night driving poses the greatest risk of casualty crash involvement on rural highways for all drivers (Table 5.6). Weekend night driving posed a significantly greater risk than either weekday day or weekday night driving. However, young drivers (aged years), although having a relatively high weekend night risk, were significantly most likely to be involved in a casualty crash on weekend days. This was also true for drivers aged years on Victorian rural highways, although no statistically significant differences were found between any time of the week risk estimates for this age group. Only two drivers aged 75+ years were surveyed during weekend nights on rural highways, so the high risk of almost 9 casualty crashes per million kilometres travelled should be treated with caution. No confidence limits have been calculated for this risk estimate due to there being fewer than 10 drivers involved in crashes (violating one of the assumptions in Section 2.3). Time Table 5.6: Casualty Crash Risk as a Function of Driver Age and Time Block Driver A~e (Years) All Driver Age by LowlHigh Alcohol Time Similar results to the day/night casualty crash risks were obtained for low/high alcohol time crash risks. The risk of casualty crash involvement was greater when driving during high alcohol times than at other times on rural highways for all drivers (Table 5.7). This trend held for drivers of most ages except for younger drivers aged and drivers aged and 75+, whose risks were greater when driving during low alcohol times. However, as depicted by the 95% confidence limits in Figure 5.3, statistically significant differences existed between the low and high alcohol time risks only for drivers aged In contrast, the Melbourne and Victorian rural town samples produced larger casualty crash risks during low alcohol times than during high alcohol times. Alcohol All 75+ Time 62 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

73 Figure 5.3: Casualty Crash Riskfor Low/High Alcohol Times as a Function of Driver A1?'efor Rural Hi1?'hwavsin Victoria 5 El LOW ALCOHOL TIME 11 HIGH ALCOHOL TIME o Driver Age (Years) Seat Belt Usage Only 0.7% of drivers surveyed and 0.8% of drivers involved in casualty crashes were not wearing seat belts on rural highways in Victoria. Appendix C displays the risk estimates and exposure estimates for seat belt and non-seat belt wearers for rural highways. 5.2 RISK ESTIMATES BY YEARS OF DRIVING EXPERIENCE As had occurred for the Victorian provincial town and Melbourne samples, only casualty crashes involving drivers aged years were used for the following 'driving experience' analysis because of the unavailability of the years of driving experience for drivers aged over 40 years. The driving experience variable was also unavailable for all drivers involved in crashes during 1994 on rural highways. The rural highway data has, therefore, been reduced to 891 casualty crashes and 2,830 surveyed drivers born during (a proxy group for drivers aged at the time of the survey). Table 5.8 presents the number of drivers involved in casualty crashes, number of surveyed drivers, aggregate exposure estimates and casualty crash risks as a function of driving experience for rural highways. The risk of casualty crash involvement on rural highways, generally decreased with increasing years of driving experience (Figure 5.4). The most inexperienced drivers (those who had their licence for less than one year) were significantly more at risk of casualty crash involvement than other drivers. However, there were no statistically significant differences between the risks for drivers with 1 to 2 years, 2 to 3 years and 3 to 5 years of driving experience. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 63

74 Since only four learner drivers were sampled, the relatively high risk of should be treated with caution, and is not presented in Figure 5.4. Furthermore, the casualty crash involvement for learner drivers is likely to be an overestimate because of the mis-coding of some non-learner driver crashes as learner driver crashes. Therefore, in the following sections where driving experience is presented with other variables for rural highways, learner driver casualty crash risks will not be shown in tables and charts, but can be found in Appendix C. Table 5.8: Casualty Crash Risk as a Function of Driving Experience Years of 3.64 (million Number Estimate 1, , Casualty Drivers , Crash Exposure 4Surveyed Risk116 Casualty km) of Crash 30 *Learner driver risk is likely to be unreliable since casualty crash involvement may have been overestimated. Figure 5.4: Casualty Crash Risk as a Function of Driving Experience for Rural Hif!hwavs in Victoria (/) CIl 'E (/)..li:: ~ a: Co ~ t: ::3 2.s::: ~()~a; CD 0.5 o < 1 YEAR 1 <= YEAR< 2 2 <= YEAR< 3 3 <= YEAR< 5 5 <= YEAR< 8 Driving Experience >=8 YEAR 11 Individual cells may not add to total due to the presence of drivers with unknown years of driving experience. 64 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

75 5.2.1 Driving Experience by Driver Gender Across all driving experience levels, female drivers were more at risk of casualty crash involvement than males (Table 5.9). The highest risk group was female drivers with less than one year's driving experience (estimated risk of 5.36). Statistically, this estimate was significantly larger than all other female and male risks. Tables C2.2 and C2.3 in Appendix C provide the 95% lower and upper confidence limits for the appropriate risks. Table 5.9: Crash Risk as a Function of Driving Experience and Driver Gender for Rural Hif!hwavs in Victl - - ~ All l~yr< ~yr< ~8 5~yr< ~yr< <1 yr Years of Drivin~ Experience Driving Experience by Vehicle Age Table 5.10 presents casualty crash risks as a function of vehicle age and driving experience for Victorian rural highways. For most driving experience levels, the risk of casualty crash involvement decreases with decreasing vehicle age for vehicles manufactured before However, the risks involving vehicles manufactured in are relatively large, particularly for the least experienced drivers. It should be noted that these risks may be somewhat unreliable since it was not possible to estimate the corresponding 95% confidence limits due to violation of the assumptions in section 2.3. This was also the case for risk estimates involving pre-1970 vehicles. Table 5.10: Casualty Crash Risk as a Function of Driving Experience and Year 0 Vehicle Manu acture or Rural Hi hwa s in Victoria VehicIe Years of Drivin Ex erience Age <1 yr l~yr<2 2~yr<3 3~yr<5 5~yr<8 Pre All Driving Experience by Number of Passengers ~8 yr All Drivers with less than one year's driving experience (as well as the most experienced drivers) are most at risk of casualty crash involvement when driving alone (Table 5.11). However, drivers who have had their licence for between 1 to 3 years have the largest risk estimates when travelling with one passenger. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 65

76 Table 5.11: Casualty Crash Risk as a Function of Driving Experience and Number o Passen ers in Vehicle or Rural Hi hwa s in Victoria No. of Years of Drivin Ex erience passengers <1 yr 1~yr<2 2~yr<3 3~yr<5 5~yr<8 None One ~ Two All ;;::8 yr All Driving Experience by DaylNight Driving Table 5.12 and Figure 5.5 depict the estimated casualty crash risks for day/night driving by years of driving experience on rural highways. Inexperienced drivers «1 year's driving experience) are more at risk of casualty crash involvement during the day than at night as are drivers with two to three years' driving experience. Other driving experience groups have higher (or similar) casualty crash risks when driving at night. These results contrast with the Melbourne and Victorian provincial town figures, where day time driving posed a greater risk than night driving across all driving experience groups. However, no statistically significant differences were found between the day and night risk estimates for any driving experience group on rural highways of Victoria. 4 I [] DAY IIINIGHT I o <1 year 1<=year<2 2<=year<3 3<=year<S >=8 year Driving Experience 66 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

77 Table 5.12: Casualty Crash Riskfor Day/Night Driving as a Function of Driving EXDerience for Rural Highways in Victoria Day vs. Years of Driving Experience Night <1 yr l:s;yr<2 2:S;yrd 3:S;yr<5 5:S;yr<8 ;:::8 yr All Day Night All Driving Experience by Time of Week For drivers aged years driving on rural highways, weekend night time driving poses the greatest risk of casualty crash involvement (Table 5.13). However, this was not the case for the least experienced drivers. This group was most at risk on weekend days, although their weekend night risk was the largest amongst all driver groups of varying years of driving experience. Table 5.13: Casualty Crash Risk as a Function of Driving Experience and Time of Week for Rural Highways in Victoria Time Years of Drivin~ Experience Block <1 yr l:s;yr<2 2:S;yr<3 3:S;yr<5 5:S;yr<8 ;:::8 yr Weekday Day Weekend Day Weekday Night Weekend Nillht All Driving Experience by LowlHigh Alcohol Times All Drivers with less than one year's driving experience and those who had been driving for two to five years, were estimated to have higher casualty crash involvement risks during low alcohol times - other driving experience groups were more at risk when driving during high alcohol times (Table 5.14). Table 5.14: Casualty Crash Risk for Low/High Alcohol Times as a Function of Driv~ng Experience for Rural Highways in Victoria Alcohol Years of Drivin~ Experience Time <1 yr 1:S;yr<2 2:S;yr<3 3:S;yr<5 5:S;yr<8 ;:::8 yr All Low High All CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 67

78 5.3 RISK ESTIMATES BY LICENCE TYPE The number of casualty crashes per million kilometres travelled are presented in Table 5.15 as a function of driver licence type for Victorian rural highways. The results are similar to the Melbourne sample with probationary drivers being significantly more at risk of casualty crash involvement on rural highways than fully licensed drivers. The learner driver risk should be treated with caution since some non-learner drivers involved in casualty crashes may have been mis-coded as learner drivers. Because of this, the following sections will not present the learner driver risk estimates. However, the risks for learner drivers can be found in Appendix C. Table 5.15: Casualty Crash Riskfor Victorian Rural Highways by Driver Licence T Licence Casualty Crash of 1,794 5,981 5, Number Drivers 1.03 Casualty 0.99 (million Surveyed 3, , ,459 Exposure Crash 279 Estimate Risk km) *Leamer driver risk is likely to be unreliable since casualty crash involvement may have been overestimated Licence Type by Driver Gender On rural highways, female drivers were significantly more at risk of casualty crash involvement than male drivers whether they held a probationary or a standard licence (Table 5.16). Table 5.16: Casualty Crash Risk as a Function of Licence Type and Driver Gender 'orrural Highways in Victoria Driver Driver Licence Type Gender Probationa Standard/Full All Male Female All Licence Type by Vehicle Age Table 5.17 gives the casualty crash risks for probationary and full licence holders as a function of vehicle age. For probationary licence holders, a statistically significant lower risk was estimated to occur when driving a vehicle than one manufactured in or It should be noted that confidence limits could not be estimated for the risks involving pre-1970 and vehicles due to the violation of the conditions in Section Individual cells may not add to total due to the presence of drivers with unknown licence type. 68 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

79 Drivers with full licences had significantly greater casualty crash risks when driving a vehicle manufactured prior to Appendix C (Tables C3.2 and C3.3) gives details of the lower and upper confidence limits placed on each risk estimate. Table 5.17: Casualty Crash Risk as a Function of Licence Type and Year of Vehicle Manu acture or Rural Hi hwa s in Victoria Vehicle Driver Licence T e A e Probationa StandardlFull All Pre All Licence Type by Number of Passengers For rural highways in Victoria, drivers with probationary licences were more at risk of casualty crash involvement when driving with one passenger than when driving alone (Table 5.18). Conversely, drivers with standard licences, had a higher estimated risk when travelling with no passengers. Table 5.18: Casualty Crash Risk as a Function of Licence Type and Number 0 Passen ers in Vehicle or Rural Hi hwa s in Victoria No. of Driver Licence T e Passen ers Probationa StandardlFull All None One ~ Two ~ O~ O~ Licence Type by Day/Night Driving The risk of casualty crash involvement was significantly greater at night than during the day for novice drivers on rural highways in Victoria (Table 5.19). For full licence holders similar day and night casualty crash risks were estimated. The opposite finding occurred for drivers in Melbourne and Victorian provincial Towns, with day risks generally being larger than night risks. Table 5.19: Casualty Crash Riskfor Day/Night Driving as a Function of Licence Tvve for Rural Highways in Victoria Day vs. Night Day Night All Probationa Driver Licence TYDe StandardlFull All CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 69

80 5.3.5 Licence Type by Time of Week Driving during weekend nights on rural highways poses the greatest casualty crash risk for drivers with either probationary or full licences (Table 5.20). For probationary licence holders, the weekend night risks were significantly greater than the weekday day risks, and for drivers with standard licences, the weekend risks (day and night) were significantly larger than the weekday risks. These figures contrast with the Melbourne results in which weekend night driving did not pose as great a risk as driving during other times of the day and week. Table 5.20: Casualty Crash Risk as a Function of Licence Type and Time of Week or Rural Highways in Victoria Driver Licence TVDe. StandardlFull All Time Block Weekday Day Weekend Day Weekday Night Weekend Ni2:ht All Probationa Licence Type by Low/High Alcohol Times Driving during high alcohol times on rural highways presents a higher casualty crash risk than driving during low alcohol times for probationary licence holders, although not a significantly greater risk (Table 5.21). Drivers with full licences had similar casualty crash risks of approximately 0.4 casualty crashes per million kilometres during low and high alcohol times. Table 5.21: Casualty Crash Riskfor Low/High Alcohol Times as a Function of Licence Tvoe for Rural Highways in Victoria Driver Licence TVDe StandardlFull All Alcohol Time Low High All Probation MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

81 6. ACCURACY AND BIAS OF DRIVER AGE The interviewer conducting the exposure survey obtained the date of birth from the driver to determine his/her age at the time of the survey. However, the response rate for the driver's full date of birth was low for all three samples. For the Melbourne sample of drivers, 33% did not give their full date of birth, hence their exact age at the time of the survey could not be determined. This group of drivers of unknown age, contributed approximately 26% towards the total exposure (distance travelled in an average week on Melbourne arterial roads). The "date of birth" response rates were smaller for the rural Victorian samples than for the Melbourne sample. Approximately 43% of drivers surveyed in the selected rural towns did not give their full date of birth, contributing almost 50% towards the total distance travelled in an average week on main roads in provincial towns. The unknown "date of birth" proportion was even higher for drivers surveyed on rural highways -73% of drivers did not give their full date of birth to the interviewer. This resulted in a contribution of 76% towards the total distance travelled on rural highways from drivers whose age was not known. Rather than discard these relatively large proportions of total exposure for the three locations, the authors were requested by VicRoads to distribute the distance travelled in an average week for drivers whose age was not known proportionately amongst the driver age groups with known exposure to estimate the casualty crash risks as functions of driver age. However, this method could result in a potential bias in the distribution of the 'unknown driver age' exposure towards certain age groups. To investigate whether such a bias exists, the driver's year of birth was used as a proxy for the driver's age at the time of the survey. Although thefull date of birth was not responded to very highly in the survey, the year of birth had a much higher response rate for all three locations. In Melbourne 17% of drivers did not state their year of birth (compared with 33% not stating their date of birth); in the selected Victorian rural towns 13% of drivers did not give their year of birth (compared with a 50% unknown "date of birth" rate) and on rural highways 16% of drivers failed to give their year of birth (compared with 73% not giving their date of birth). The corresponding contributions towards the total exposure from the drivers with unknown years of birth ranged from 15% to 16% for the three samples. Table 6.1 gives the total distance travelled in an average week on Melbourne arterial roads for drivers born before 1977 according to whether their full date of birth was known or not known. The single years of birth and the year groups correspond (approximately) to the single year of age and age groups used in the previous analysis (eg. Table 3.1). The driver's birthday would need to be during the survey (ie July 1994) for there to be exact correspondence between year of birth and age, ie. a driver born in 1976 during July would be almost exactly 18 years old at the time of the survey. As can be seen in Table 6.1 there does appear to be a bias towards older drivers having a higher proportion of the unknown dates of birth. A chi-square test of CRASH RISKS OF ROAD USER GROUPS IN VICfORIA 71

82 association showed there was a statistically significant difference in the distribution of unknown and known dates of birth by the year of birth of the driver (p<o.oool). For example, 96.1 % of the distance travelled by drivers born in 1976 was for those with a known date of birth (and therefore a known age when surveyed) whereas only 71.7% of the distance travelled by drivers born before 1920 was for those with a known full date of birth. The unknown "date of birth" proportions tended to increase with increasing driver age. The implications of this on the Melbourne casualty crash risks presented in Chapter 3 is that the estimates for younger drivers (particularly those aged under 22) are likely to be more accurate than those for older drivers (aged over 60 years). r Table 6.1: The distance travelled in an average week on Melbourne arterial roads during 1994 as afunction of Year of Birth of Driver and , ,164 6,537 4,897 2,443 6,882 3,730 (3.9%)* (5.6%)6,270 (3.2%)1,315 (10.8%) UnknownKnown (12.9%) 530 Total (8.2%) (10.7%) (11.7%) (14.1%) (13.7%) (29.5%) 1,264 6,020 5,278 1,347 49,561 41,869 20,344 11,724 24, ,782 1,877 5,914 22,787 26,333 48,751 27,228 6,639 6,218 16,622 56, ,343 (28.3%) 23,498 (96.1%) (96.8%) (94.4%) (89.2%) (71.7%) (89.3%) (88.3%) (85.9%) (70.5%) (91.8%) (87.1%) (86.3%) Date of Birth of Driver *figures in brackets refer to the proportion of the distance travelled by drivers born in each year group with unknown and known dates of birth. Tables 6.2 and 6.3 give the corresponding exposure estimates to Table 6.1 for the Victorian rural towns and rural highways samples, respectively. For travel in rural towns, the distribution of known "date of birth" proportions tends to be higher for younger drivers, however the unknown proportion did not increase uniformly with increasing driver age. In fact, a chi-square test of association did not show a significant difference between the known and unknown dates of birth with respect to the driver's year of birth (p=0.3487). Conversely, there was a statistically significant association between the driver's year of birth and whether or not his/her date of birth was known for drivers surveyed on rural highways (p<o.oool). The unknown proportions were less of a problem for young drivers than for older drivers. 72 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

83 r Table 6.2: The distance travelled in an average week in Victorian Rural Towns during 1994 as a function of Year of Birth of Driver and 67,035 2,428 2,806 1, ,281 15,554 5,759 6,208 8,915 3,566 (15.5%)*2,319 (45.6%)2,896 (45.4%)1,604 (47.3%)3,121 Unknown (40.3%) 759 TotalKnown (32.5%) (31.2%) (36.6%) (45.3%) (43.4%) (41.5%) 1,429 15, ,192 11,418 2,938 (53.1%) 5,927 5,324 44,381 2,743 19,671 19,117 37,495 12,909 10,757 25,100 21,941 7,852 9,988 99, (46.9%) (84.5%) (54.6%) (68.8%) (63.4%) (52.7%) (54.4%) (54.7%) (67.5%) (58.5%) (56.6%) (59.7%) Date of Birth of Driver - *figures in brackets refer to the proportion of the distance travelled by drivers born in each year group with unknown and known dates of birth. r Year of Birth Table 6.3: The distance travelled in an average week on Victorian Rural Highways during 1994 as a function of Year of Birth of Driver and ,185 2,561 2,559 2, ,020 31,261 11,991 19,616 10,765 11,638 (50.7%)* (59.8%)1,725 (73.8%) (66.2%)1,181 Unknown (72.5%) 663 TotalKnown (68.6%) (73.8%) (75.4%) (70.1%) (73.2%) (74.8%) 3,467 1,567 26,029 15,549 17,490 15, ,584 4,286 48,799 42,695 3, (78.8%) 12,780 11,435 49,399 4,597 5,499 3,911 6, (21.2%) (49.3%) (40.2%) (33.8%) (31.4%) (29.9%) (25.2%) (24.6%) (26.2%) (26.8%) (27.5%) Date of Birth of Driver *figures in brackets refer to the proportion of the distance travelled by drivers born in each year group with unknown and known dates of birth. Figure 6.1 displays the proportion of exposure in which the driver's was known for the three locations by driver "year of birth". "date of birth" The proportion of the exposure where the driver's date of birth was known was greater for the Melbourne sample of drivers than for the rural Victoria samples across all ages. Thus, the casualty crash risk estimates for Melbourne are likely to be more accurate than the risk estimates for travel in rural towns or rural highways. Further, the Melbourne risks for young drivers would be more reliable than the older driver risks because of the increasing proportion of the distance travelled by older drivers CRASH RISKS OF ROAD USER GROUPS IN VICfORIA 73

84 whose full date of birth was not known. This trend was also found for rural highway travel, but not for travel in rural towns in which the known proportions did not generally decrease with increasing driver age. Figure 6.1: Percentage of Distance Travelled by Drivers with KNOWN Date of Birth as a function of Driver Year of Birth and Location t:: :i: ;; 1ii t, 0 >< ':i w ; Cl a.. id Z 0Zl<: > '"E.~Q) '"8. 40 (j;.r: (j; '0 2l r>:lrural Towns DRural Highways Driver Y.ar of Birth PRE MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

85 7. SUMMARY The estimates of casualty crash risk per million kilometres were statistically significantly greater on Melbourne arterial roads (0.82) and in Victorian rural towns (1.21) than on rural highways (0.46). Driver age The driver's age at the time of the survey was not asked directly by the interviewer. Instead the interviewer asked the driver's date of birth and obtained at least the year of birth in 83-87% of cases. However, the year of birth was not adequate to accurately determine the driver's age as, for example, a driver born in 1976 could be aged from 17 years 7 months to 18 years 7 months at the time of the survey in July August The age of the driver was calculated for those for whom their full date of birth was known, and the exposure of those with unknown age was distributed in proportion to those with known age. However, the percentage of drivers whose full date of birth was known was considerably lower than those with year of birth known, so this method could have resulted in considerable errors of estimation. The proportion of the exposure where the driver's date of birth was known was greater for the Melbourne sample of drivers than for the rural Victoria samples across all ages. Thus, the casualty crash risk estimates for Melbourne are likely to be more accurate than the risk estimates for travel in rural towns or rural highways. Further, the Melbourne risks for young drivers would be more reliable than the older driver risks because of the increasing proportion of the distance travelled by older drivers whose full date of birth was not known. This trend was also found for rural highway travel, but not for travel in rural towns in which the known proportions did not generally decrease with increasing driver age. For each location, the risk of casualty crash involvement apparently decreased with increasing driver age up to about age 50 years. After 50 years of age the estimated risk apparently increased again. The highest risk driver groups were young and older drivers, aged and 75+. On Melbourne arterial roads and in Victorian rural towns, the risk estimate for drivers aged 18, 19 and 21 were significantly greater than the risks for drivers aged 20 and those aged between 22 and 74, whilst on rural highways, the risks for drivers aged 18, 19 and 20 years did not differ significantly. The casualty crash risk for drivers aged 75+ was significantly larger than for younger drivers in all three locations. Driver experience and licence type The Melbourne, Victorian rural town and rural highway samples also showed similar casualty crash risk patterns when the risk was estimated as a function of driving experience or driver licence type. For the three samples, drivers with the least years' experience (namely those with less than one year's driving experience) were significantly more likely to be involved in a casualty crash than more experienced drivers, however statistically significant differences were only found for Melbourne and rural highways. Further, probationary and learner drivers had significantly larger CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 75

86 risk estimates than drivers with full licences on Melbourne arterial roads and on the selected rural highways of Victoria. Driver gender Except for drivers aged and female drivers in Melbourne were significantly more at risk of being involved in a casualty crash than male drivers of the same age. Further, females in their second year of driving had a significantly greater risk of casualty crash involvement than males with the same driving experience in Melbourne. For the rural town sample of drivers, females aged and had significantly larger casualty crash risk estimates than males of the same age, whereas females with between 3 to 5 years of driving experience were significantly more at risk of being involved in a casualty crash than males with the same level of driving experience. On rural highways, females aged 18-21,22-25, and had significantly larger crash risk estimates than males. Further, females in their first year of driving were estimated to have a significantly larger casualty crash risk than males. There was evidence that on Melbourne arterial roads, for more severe crashes the differences between the male and female risks tended to disappear, particularly for younger drivers. Female serious casualty crash risks were only marginally different from the corresponding male risks, with the exception of drivers aged and 75+ where female risks were significantly larger. Male drivers aged involved in fatal crashes had larger risks than females, however, no statistically significant differences between the female/male risks were found for fatal crashes across all age groups and all driving experience levels. Vehicle age Generally the risk of casualty crash involvement increased with increasing vehicle age for Melbourne, Victorian provincial towns and Victorian rural highways for all drivers. When considering crash involvement as a function of driver age, however, young drivers aged had relatively larger risks if driving a vehicle manufactured in for all three samples. Accompanying passengers Drivers aged and drivers in their second year of driving, who were travelling with at least two passengers on Melbourne arterial roads, were significantly more at risk of casualty crash involvement than when driving with one or no passengers. Drivers aged in rural towns were also significantly more likely to be involved in casualty crash when travelling with at least two passengers than with fewer than two. Further, the high risk group (drivers aged 18-21) appeared to be more at risk of casualty crash involvement when travelling with two or more passengers than when driving alone on Melbourne arterial roads, although this risk was not significantly larger than the risk produced when travelling with fewer than two passengers. Conversely, this road user group appeared to be most at risk when driving with one passenger on rural highways, however again no statistically significant differences 76 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

87 were found. For rural towns, drivers aged had similar risks irrespective of the number of accompanying passengers. Time of day and day of week When expressing the casualty crash risks as functions of time of day and/or time of week, the findings for the Melbourne and Victorian provincial towns during 1994 were contrary to previous studies. Drummond and Yeo (1992) had found higher crash risks during night hours than during the day in Melbourne during For drivers on Melbourne arterial roads and in Victorian rural towns, no statistically significant differences were found between the day and night risk estimates (though the day risks were marginally higher). However, drivers aged had a significantly larger weekday day risk than weekday night risk in Melbourne and in rural towns whilst those aged had a significantly larger weekday night risk only on Melbourne arterial roads. Younger drivers, aged 18-21, were significantly more likely to be involved in a casualty crash on weekdays than on weekends in Melbourne (during either day or night times of day), whilst the oldest group of drivers (aged 75+) were significantly more likely to be involved in a casualty crash on weekends than on weekdays. On rural highways, the day risks were significantly greater than the night risks for drivers aged but for drivers aged the night risks were larger. Changes since 1988 Casualty crash risks have decreased for all drivers on Melbourne arterial roads since 1988, from 1.36 to 0.82 casualty crashes per million kilometres travelled in The largest decrease in risk between 1988 and 1994 was for weekend night crashes. In particular, the greatest reduction apparently occurred for young drivers aged years on weekends during both the night and day. For metropolitan Melbourne and Victorian rural towns, driving during the day was associated with a higher risk of casualty crash involvement in 1994 than driving during the night, although no statistically significant differences were found. This is opposite to what was found on rural highways, which suggests that the large decrease in casualty crash risk during night hours, since 1988, has not occurred to the same extent on rural highways as in urban areas. Potential night driving curfew periods On Melbourne arterial roads, the risk of casualty crash involvement when driving during potential night curfew periods of 8pm-6am or midnight-6am was greater than the day risk, though a monotonic decrease in risk did not occur with the reduced duration of the night curfew periods. A statistically significant greater risk occurred when driving during midnight-6am than during lopm-6am for drivers aged 22-25, whereas drivers aged had a significantly lower risk of casualty crash involvement during midnight-6am than at any other time of the day. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 77

88 78 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

89 8. DISCUSSION This report presents the results of an analysis of two different data sets which were provided to MUARC. The first of these was data collected during an exposure survey conducted by Amp Transportation Planning during 1994 on behalf of VicRoads. It was presumed by MUARC that the data were accurate and that the survey measured what it purported to measure. The second data set was a file of accident report data, originally collected by the Victoria Police during , and subsequently enhanced by VicRoads. Because the MUARC analysis made use of data prepared by others, it should be emphasised that the results depend on the validity of the original data collections. They also depend on a range of assumptions made in the analysis, which include the assumption that accident patterns during the years were relatively stable and hence could be compared with 1994 exposure estimates to measure casualty accident risks during that year, and the assumption that drivers with unknown age have their exposure distributed in proportion with those of known age. Subject to the accuracy of the data provided and the assumptions on which the analysis was based, this report shows some remarkable findings regarding casualty accident risks in Victoria during The first of these was the generally higher casualty accident rate per kilometre driven by female drivers compared to males. There has been a belief that male accident risks are generally higher (notwithstanding the fact that previous research had also found higher casualty accident risks for females in 1988). The relative difference between the sexes may reduce and even reverse when the risks of more severe accidents are considered, thus explaining the prior beliefs. In this study it was found for more severe crashes, the differences between the male and female risks tended to disappear, however there was still no statistically significant difference between the fatal accident risks of the two sexes. The second was the apparently high casualty accident rate per kilometre driven by drivers aged 75 and above. There is a belief that drivers of this age do not have high accident rates and this may be the case on a per licence holder basis. However, when distance driven is taken into account, the effect of advanced age on casualty accident risks becomes apparent. It is likely that drivers of this age drive relatively few kilometres per annum on average, thus explaining their not unusual accident risks per licence holder. The third remarkable finding was the absence of statistically significant differences between casualty accident rates per kilometre when driving at night compared with the rate when driving during the day. There were even suggestions of lower accident risks during the night than the day in urban areas during Previous research of the same type had found higher casualty accident rates at night in Melbourne during 1988, compared with the rates during the day. CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 79

90 The fourth noteworthy set of findings was that there have been substantial reductions in accident risk per kilometre driven in Melbourne for some types of drivers and circumstances between 1988 and The largest reductions appear to have been experienced by young drivers, particularly on weekends, and during both the night and day-time hours. These reductions in accident risk are to be expected, given the dramatic success of the random breath testing and speed camera programs (plus supporting mass media publicity) in reducing road trauma in Victoria during (Cameron et al 1995) and the fact that recreational driving by young drivers was frequently the focus of these programs. There was probably also a contribution from the downturn in the economy during this period. These new findings suggest that a new pattern of casualty accident risks has taken shape in Victoria during the 1990' s compared with earlier years. If this is the case, the implications for countermeasure development in Victoria are significant. Because of these implications, it is important that research be conducted to be sure that the changes are real. This could be determined, to a greater extent than the current work, by repeating the 1994 exposure survey during 1996, and combining the results with the 1994 exposure data and accident data, to verify the new pattern of accident risks apparently present during the 1990's. 80 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

91 9. CONCLUSION A survey of motorised vehicle travel conducted under contract for VicRoads in Melbourne, selected country towns and on rural highways during 1994 has provided estimates of the distances travelled by drivers in Victoria. When combined with data on casualty accident involvements during , this information has allowed casualty accident risks per kilometre travelled to be estimated for a broad range of circumstances in each of these three environments. It has been possible to add confidence limits to the risk estimates so that real differences in risk can be seen. However the estimates are only as good as the original data collections, which were beyond MUARC's control, and the assumptions made in the analysis. A comparison with similar risk estimates derived for Melbourne during 1988 has found some specific reductions in risk, probably associated with the substantial improvement in road safety in Victoria during the 1990's. 10. RECOMMENDATIONS Because of the major implications of this type of research for countermeasure development in Victoria, it is recommended that the research be extended by repeating the 1994 exposure survey during 1996, and combining the results with the 1994 exposure data and accident data, to verify the new pattern of accident risks apparently present during the 1990' s. CRASH RISKS OF ROAD USER GROUPS IN VICfORIA 81

92 82 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

93 11. REFERENCES Arup Transportation Planning. (1995). The 1994 Crash Exposure Survey. Prepared for VicRoads, Melbourne. Cameron, M. (1992). Accident Data Analysis to Develop Target Groups for Countermeasures, Volume 1: Methods and Conclusions. Report No. 46, Monash University Accident Research Centre, Victoria. Cameron, M., Newstead, S. & Gantzer, S. (1995) Some principles learnt from evaluations of enforcement and supporting publicity programs in Victoria. Proceedings, 1995 Road Safety Conference, Institute of Transportation Engineers Australia, Melbourne, August Cameron, M. & Oxley, J. (1995). Investigation of Improved Exposure Datafor the Assessment of Road Safety. Report No. RIIP-6, Australian Institute of Health and Welfare, National Injury Surveillance Unit, South Australia. Diamantopoulou, K., Skalova, M. & Cameron, M. (1995). Crash Risks of Road User Groups: Interim Report on the Crash Risks of Young, Old and Inexperienced Drivers. Monash University Accident Research Centre, Victoria. Drummond, AE. & Yeo, E.Y. (1992). The Risk of Driver Crash Involvement as a Function of Driver Age. Report No. 49, Monash University Accident Research Centre, Victoria. Fildes, RN., Lane, J.C., Lenard, J. & Vu1can, AP. (1991). Passenger Cars and Occupant Injury. CR 95, Federal Office of Road Safety, Canberra. Harrison, W.A (1990). Update of Alcohol Times as a Surrogate Measure of Alcohol-Involvement in Accidents. Research Note, Monash University Accident Research Centre, Victoria. Kotz, S. & Johnson, N.L. (1986). Encyclopedia of Statistical Sciences, Volume 7. John Wiley & Sons, Inc. U.S.A CRASH RISKS OF ROAD USER GROUPS IN VICTORIA 83

94 84 MONASH UNIVERSITY ACCIDENT REsEARCH CENTRE

95 APPENDIX A CRASH RISK ESTIMATES AND CONFIDENCE LIMITS, EXPOSURE ESTIMATES AND STANDARD ERROR ESTIMATES, CRASH INVOLVEMENT AND NUMBER OF SURVEYED DRIVERS ON MELBOURNE ARTERIAL ROADS

96 TABLEA1.1 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: MELBOURNE, 1994,,4~ diljl!!jiliii;:':ili ~~!' Total MANUFACTURE MI DRIVER ONLY I S ' J SUNDAY ( O.7l O.9tHI

97 TABLE Al.2 95% LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: MELBOURNE, 1994 GENDER MANUFACTURE DRIVER ONLY SUNDAY ( HI O.7!H'

98 TABLE Al.3 95% UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: MELBOURNE, 1994 GENDER MANUFACTURE DRIVER ONLY SUNDAY 'Se?

99 TABLE Al.4 NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's DURING AS A FUNCTION OF DRIVER AGE GENDER l loo l I I I3 5272I3 226 I l HXl I loll I l I l Il Il 2 I I Il I 47 I MANUFACTURE SUNDAY DRIVER ONLY

100 TABLE A1.5 NUMBER OF DRIVERS SURVEYED ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's DURING 1994 AS A FUNCTION OF DRIVER AGE GENDER III II I I I I I I MANUFACTURE DRIVER WEEKDAY SUNDAY LOW Total MIDNIGHT-1:59AM ALCOHOL ONLY 6AM-6PM TIME 91

101 TABLEAl.6 DISTANCE TRA VELLED IN AN AVERAGE WEEK (in thousand KM) DURING) 994 ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's AS A FUNCTION OF DRIVER AGE I GENDER MIDNIGHT-1:59AM Total WEEKDAY DRIVER MANUFACTURE SUNDAY LOW ALCOHOL ONLY 6AM-6PM TIME

102 TABLE Al.7 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in thousand KM) ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's DURING 1994 AS A FUNCTION OF DRIVER AGE GENDER " ()(J lHl t MANUFACTURE DRIVER SUNDAY ONLY

103 TABLE Al.8 DISTANCE TRAVELLED DURING A 185 NON-HOLIDAY WEEK PERIOD (in million KM) ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's AS A FUNCTION OF DRIVER AGE, with distance travelled for drivers of unknown age distributed proportionately amongst driver age groups GENDER MANUFACTURE WEEKDAY SUNDAY DRIVER MIDNIGHT-1:59AM Total LOW ALCOHOL ONLY 6AM-6PM TIME

104 TABLEA2.1 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: MELBOURNE, Learner , licence Total and Under to over under nnder under NOTVIC. 1 year MANUFACTURE SUNDAY DRIVER 1976 ONLY 8 years

105 TABLEA2.2 95% LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: MELBOURNE, _ Total and Under to to over under NOTVIC. 1licence year 35 8Learner MANUFACfURE DRNERONLY SUNDAY years

106 TABLEA2.3 95% UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: MELBOURNE, N1A Total and Under to to over under NOTVIC. licence 1 year 35 8Learner DRIVER MANUFACTURE SUNDAY 1976 ONLY 8 years

107 TABLEA2.4 NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's DURING AS A FUNCTION OF DRIVING EXPERIENCE I Total licence Learner Under I I and to under over lyear NOTVIC years DRlVERONLY MANUFACfURE SUNDAY 1976

108 TABLEA2.5 NUMBER OF DRIVERS SURVEYED ON MELBOURNE ARTERIAL ROADS FOR SELECfED LGA's DURING 1994 AS A FUNCTION OF DRIVING EXPERIENCE Total licence 31 Learner Under I and 6 to under over year 13NOTVIC I I MANUFACTURE DRIVER MIDNIGHT-1:59 Total WEEKDAY LOW SUNDAY 1976 ALCOHOL ONLY 6AM-6PM TIME 8 years

109 TABLEA2.6 DISTANCE TRAVELLED IN AN AVERAGE WEEK (in thousand KM) DURING 1994 ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's AS A FUNCTION OF DRIVING EXPERIENCE Total licence Learner Il Il Under Il Il and to under over lyear NOTVIC years Total LOW WEEKDAY MIDNIGHT-1:59 MANUFACTURE DRlVERONLY SUNDAY 1976 ALCOHOL 6AM-6PM TIME

110 TABLEA2.7 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in thousand KM) DURING 1994 ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's AS A FUNCTION OF DRIVING EXPERIENCE licence Learner Q and Under Total to to under over NOTVIC. 1 year MANUFACTURE DRIVER SUNDAY 1976 ONLY 8 years

111 TABLEA2.8 DISTANCE TRAVELLED DURING A 185 NON-HOLIDAY WEEK PERIOD (in million KM) ON MELBOURNE ARTERIAL ROADS FOR SELECfED LGA's AS A FUNCTION OF DRIVING EXPERIENCE licence Learner Total Under and to under2 over NOTVIC. Iyear 8 32 to under MANUFACTURE SUNDAY DRNERONLY years

112 TABLEA2.9 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: MELBOURNE, WEEKEND NIGHT PM-6AM IWEEKDA Y NIGHT 6PM-6AM 8 I and over Learner licence I Total 8 years I INOT VIe BLOCK to under 5 I5 to under 8 Iand over Learner licence I Total 8 years I INOT VIe.

113 TABLEA % LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: MELBOURNE, 1994 '-." Total and Under to to over under NOTVIC. licence 1 year 38 WEEKEND 5Learner 2 DAY AM-6PM WEEKDAY NIGHT 6PM-6AM 8 years Total and to over under NOTVIC. licence 38 5Learner 1 to under years

114 TABLEA % UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: MELBOURNE, Total and Under 1 to to under over NOTVIC. licence 1 year 35 8Learner WEEKEND WEEKDAY 2 DAY NIGHT AM-6PM 6PM-6AM ut 8 years Total and Under to to under over NOTVIC. licence 1 year 35 8Learner years

115 TABLEA2.12 NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's DURING , AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER () Total licence Learner and to WEEKDAY under overnotvic to NIGHT under PM-6AM 2 WEEKEND NIGHT 6PM-6AM 1 8 years Total licence 1Learner and 964 to under over 14NOTVIC to under years

116 TABLEA2.13 NUMBER OF DRIVERS SURVEYED ON MELBOURNE ARTERIAL ROADS DURING 1994 FOR SELECTED LGA's, AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER Total I72 licence Learner and to under overnotvic to under c:::, WEEKDAY WEEKEND 6 NIGHT 47 NIGHT PM-6AM 6PM-6AM WEEKDAY 6AM-6PM 8 years Total licence Learner and to under over 184NOTVIC to under WEEKDAY 6AM-6PM 8 years

117 TABLEA2.14 DISTANCE TRAVELLED IN AN AVERAGE WEEK (in thousand KM) ON MELBOURNE ARTERIAL ROADS DURING 1994 FOR SELECTED LGA's, AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER Total licence Learner and to WEEKEND WEEKDAY under over NOTVIC NIGHT to under1630 6PM-6AM Cf) WEEKDAY 6AM-6PM 8 years Total 195 licence Learner and 273 to under over 6972 NOTVIC to under WEEKDAY 6AM-6PM 8 years

118 TABLEA2.15 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in thousand KM) DURING 1994 ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's, AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER licence Learner and Under Total 1 to to under over NOTVIC. WEEKDAY WEEKEND 1 year NIGHT PM-6AM -0 8 years licence Learner and Total Under to to under over NOTVIC. lyear years

119 TABLEA2.16 DISTANCE TRAVELLED DURING A 185 NON-HOLIDAY WEEK PERIOD (in million KM) ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's, AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER WEEKEND DAY 6AM-6PM licence Learner Total and to under over NOTVIC. WEEKDAY 8 NIGHT PM-6AM 3 to under 5 WEEKEND NIGHT 6PM-6AM 8 years licence Learner Total and to under over NOTVIC to under 5 8 years

120 TABLEA3.1 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF LICENCE TYPE: MELBOURNE, 1994 UNLICENSED PROBATION LEARNER FULL Total MANUFACTURE DRIVER SUNDAY ONLY I!!

121 TABLEA3.2 95% LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF LICENCE TYPE: MELBOURNE, 1994 GENDER N1A MANUFACTURE DRIVER SUNDAY ONLY

122 TABLEA3.3 95% UPPER BOUNDS FORCASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF LICENCE TYPE: MELBOURNE, 1994 GENDER D MANUFACTURE SUNDAY DRIVER ONLY ( 13

123 UNLICENSED Total TABLEA3.4 NUMBER FULL OF DRIVERS INVOLVED 3 IN CASUALTY CRASHES '.~:',,}0",,?,r j~"~ t<?:,?: ',. ".~ ",:~ ;-,)~:,~;~-,,~ ~--J:K~:'" :~t,. L~>;'8r;.j::'-" ):~';9;,f ARTERIAL ROADS FOR MANUFACTURE DRIVER SUNDAY ONLY SELECTED AS LEARNERA LOA's FUNCTION DURlNO PROBATION MELBOURNE OF LICENCE TYPE ~ ~ _.ww,... t lif

124 TABLEA UNLICENSED I Total FUlL 2,,",,.""0,,. "" ":1,._,,.":' LEARNER. _,_, '.,_,_ PROBATION """".,"~.,.... :,,_,_..,,' v'@,.,-"",, 00,8'<;. X"". "'...,,"'..,im._...,.,.","~~w,@ _..Am MANUFACTURE SUNDAY LOW MIDNIGHT-1:59 Total DRIVER WEEKDAY ON MELBOURNE ALCOHOL ONLY 6AM-6PM TIME SELECTED ARTERIAL LGA's ASROADS A FUNCTION DURINGOF1994 LICENCE FOR NUMBERTYPE OF DRIVERS SURVEYED ~ JIII""".JII8W.]1p11"i )15"

125 TABLEA3.6 DISTANCE TRA YELLED IN AN AVERAGE WEEK (in thousand KM) DURING 1994 ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's AS A FUNCTION OF LICENCE TYPE GENDER DRIVER MANUFACTURE Total MIDNIGHT-1:59 SUNDAY WEEKDAY LOW ALCOHOL ONLY 6AM-6PM TIME 32 /(10

126 TABLEA3.7 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in thousand KM) DURING 1994 ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's AS A FUNCTION OF LICENCE TYPE GENDER DRIVER MANUFACTURE SUNDAY ONLY

127 TABLEA3.8 DISTANCE TRAVELLED DURING A 185 NON-HOLIDAY WEEK PERIOD (in million KM) ON MELBOURNE ARTERIAL ROADS FOR SELECTED LGA's AS A FUNCTION OF LICENCE TYPE GENDER MANUFACTURE SUNDAY Total MIDNIGHT-I:59 WEEKDAY LOW DRIVER ALCOHOL ONLY 6AM-6PM TIME li~

128 TABLEA4.1 POTENTIAL NIGHT DRIVING CURFEW PERIODS BY AGE: MELBOURNE, pm-6am CASUALTY "'"- mtal CRASH RISK BY AGE BY NIGHT CURFEW TIME ~ 95% LOWER BOUNDS FOR CASUALTY CRASH RISKS 8pm-6am TOTAL pm-6am 95% UPPER BOUNDS FOR CASUALTY CRASH RISKS TOTAL " 8pm-6am NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES DURING TOTAL BY AGE BY NIGHT CURFEW TIME " 8pm-6am NUMBER OF DRIVERS SURVEYED DURING 1994 BY AGE BY NIGHT CURFEW TIME TOTAL :;* I!ltIJl~1'l: DISTANCE TRAVELLED IN AN AVERAGE WEEK (IN 1000 KM) DURING 1994 BY AGE BY NIGHT CURFEW TIME 8pm-6am TOTAL pm-6am STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (IN 1000 KM) DURING 1994 BY AGE BY NIGHT CURFEW TIME, TOTAL DISTANCE TRAVELLED DURING A 185 NON-HOLIDAY WEEK PERIOD (in million km) BY AGE BY NIGHT CURFEW TIME, with distance travelled for drivers of unknown age distributed proportionately amongst driver age groups y'.- mtal atJ! 8pm-6am 11,

129 TABLEA4.2 POTENTIAL NIGHT DRIVING CURFEW PERIODS BY DRIVING EXPERIENCE: MELBOURNE, 1994 CASUALTY CRASH RISK BY DRIVING EXPERIENCE BY NIGHT CURFEW TIME 8pm-6am licence Learner Total l.l1l and l.l to Under to over under under NOTVIC vear yean; 95% LOWER BOUNDS FOR CASUALTY CRASH RISKS 8pm-6am licence l.l18 Learner Total and l.l to I1.262 Under to over under NOTVIC year yean; 95% UPPER BOUNDS FOR CASUALTY CRASH RISKS 8pm-6am licence Learner l.l76 Total t l and l l to Under to over under NOTVIC year years NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES DURING BY DRIVING EXPERIENCE BY NIGHT CURFEW TIME 8pm-6am licence Learner 41 Total and to to IUnder tounderz unders over NOTVIC. 13 8vear years 8pm- 6am NUMBER OF DRIVERS SURVEYED DURING 1994 BY DRIVING EXPERIENCE BY NIGHT CURFEW TIME Total to under 35 I!! licence Learner and to Under tounderz over NOTVIC. 18year yean; DISTANCE TRAVELLED IN AN AVERAGE WEEK (IN 1000 KM) DURING 1994 BY DRIVING EXPERIENCE BY NIGHT CURFEW TIME 8pm-6am licence Learner Total and to 1806 IUnder to over underz NOTVIC year years 8pm licence Learner Total and lounder to Under to over underz NOTVIC year am 8 yean; STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (IN 1000 KM) BY DRIVING EXPERIENCE BY NIGHT CURFEW TIME L 8pm- DISTANCE TRAVELLED DURING A 185 NON-HOLIDAY WEEK PERIOD (in million km) BY DRIVING EXPERIENCE BY NIGHT CURFEW TIME licence Learner Total to Under to under 8 years under NOTVIC. I35 year 2 and over am I 5 launders 11

130 TABLEA4.3 SERIOUS CASUAL TV CRASH RISKS: MELBOURNE 1994 SERIOUS CASUALTY CRASH RISKS BY AGE BY GENDER TOTAL TOTAL tl % LOWER BOUNDS FOR SERIOUS CASUALTY CRASH RISKS BY AGE BY GENDER " TOTAL o.m % UPPER BOUNDS FOR SERIOUS CASUALTY CRASH RISKS BY AGE by- GENDER NUMBER OF DRIVERS INVOLVED IN SERIOUS CASUALTY CRASHES DURING IN MELBOURNE BY AGE BY GENDER TOTAL , m licence Learner Total and to I Under to underj over NOTVIC. 15 8year SERIOUS CASUALTY CRASH RISKS BY DRIVING EXPERIENCE BY GENDER,>" " 8 years 95% LOWER BOUNDS FOR SERIOUS CASUALTY CRASH RISKS BY DRIVING EXPERIENCE BY GENDER Learner Total and to Under to over under under NOTVIC year $\l1iim!n("b_t~ ;' licence years 95% UPPER BOUNDS FOR RISK SERIOUS CASUALTY CRASH RISKS BY DRIVING EXPERIENCE BY GENDER licence Learner Total and to Under to over under under NOTVIC. I53 8year years x_,j NUMBER OF DRIVERS INVOLVED IN SERIOUS CASUALTY CRASHES DURING IN MELBOURNE BY DRIVING EXPERIENCE BY GENDER licence Learner Total lounderS and 624 5to to 1Under to over under NOTVIC. I3 8year years

131 TABLEA4.4 FATAL CRASH RISKS: MELBOURNE, 1994 TOTAL I OJ)(J() " FATAL CRASH RISKS BY AGE BY GENDER., 95% LOWER BOUNDS FOR FATAL CRASH RISKS BY AGE BY GENDER TOTAL % UPPER BOUNDS FOR FATAL CRASH RISKS BY AGE BY GENDER TOTAL , NUMBER OF DRIVERS INVOLVED IN FATAL CRASHES DURING IN MELBOURNE BY AGE BY GENDER TOTAL rs licence Learner Total and tu to 1 Under tounderz over NOTVIC year 0 '".~.' FATAL CRASH RISKS BY DRIVING EXPERIENCE BY GENDER 95% LOWER BOUNDS FOR FATAL CRASH RISKS BY DRIVING EXPERIENCE BY GENDER, licence Learner Total and to Under to over under under NOTVIC vear 2 8 years 95% UPPER BOUNDS FOR FATAL CRASH RISKS BY DRIVING EXPERIENCE BY GENDER i'ti'~ licence Learner Total and to 1 Under to under over uoder8 under NOTVIC. 13 5year 2 8 years NUMBER OF DRIVERS INVOLVED IN FATAL CRASHES DURING BY DRIVING EXPERIENCE BY GENDER licence Learner Total and to 161 6Under to over under NOTVIC year years

132 APPENDIXB CASUALTY CRASH RISK ESTIMATES AND CONFIDENCE LIMITS, EXPOSURE ESTIMATES AND STANDARD ERROR ESTIMATES, CASUALTY CRASH INVOLVEMENT AND NUMBER OF SURVEYED DRIVERS IN VICTORIAN PROVINCIAL TOWNS 1;(3

133 TABLE Bl.l CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: VICTORIAN PROVINCIAL TOWNS, 1994 "" Total MANUFACTURE l.l DRIVER ONLY l.l !.l SUNDAY l.l l.l Ul

134 TABLE Bl.2 95% LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: VICTORIAN PROVINCIAL TOWNS, 1994 GENDER l.l MANUFACTURE l.l DRIVER ONLY U l.l SUNDAY l.l

135 TABLE Bl.3 95% UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: VICTORIAN PROVINCIAL TOWNS, 1994 GENDER MANUFACTURE DRIVER ONLY SUNDAY la UK)

136 TABLE Bl.4 NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES ON ARTERIAL ROADS IN VlcrORIAN PROVINCIAL TOWNS DURING AS A FUNcrION OF DRIVER AGE I GENDER loo I4 I I I I I I I SUNDAY MANUFACTURE DRIVER ONLY

137 TABLE Bl.5 NUMBER OF DRIVERS SURVEYED ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS DURING 1994 AS A FUNCTION OF DRIVER AGE I GENDER Jl Jl Jl LOW WEEKDAY Total DRIVER MIDNIGHT-1:59 MANUFACTURE SUNDAY ALCOHOL ONLY 6AM-6PM TIME

138 TABLEB1.6 DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS AS A FUNCTION OF DRIVER AGE GENDER \ I LOW WEEKDAY MIDN1GHT-1 Total DRIVER MANUFACTURE SUNDAY ALCOHOL ONLY 6AM-6PM :59 TIME 130

139 TABLE Bl.7 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS DURING 1994 AS A FUNCTION OF DRIVER AGE GENDER MANUFACTURE DRIVER ONLY SUNDAY D (HJ

140 TABLEB1.8 DISTANCE TRAVELLED DURING A 185NON-HOLIDAY WEEK PERIOD (in million KM) ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS AS A FUNCTION OF DRIVER AGE with distance travelled for drivers of unknown age distributed proportionately amongst driver age groups GENDER )() O.()()() O.(X)() 5.()( ,()(J (0) (l.()()() O.IXX) O.IX)() O.(lOO O.llOO !.I O.(XX) O.O()() O.O()(J IXJ O.(lOO (l.ooo O.llOO O.()()() LOW MIDNIGHT-I:59 MANUFACTURE WEEKDAY SUNDAY DRIVER ALCOHOL ONLY 6AM-6PM TIME \32

141 TABLEB2.1 CASUAL TV CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: VICTORIAN PROVINCIAL TOWNS, O.!KJO l Learner A l.lkj licence oJ l.lkj ( and Total Under to over under NOTVIC. Ivear MANUFACTURE DRlVERONLY SUNDAY years 133

142 TABLEB2.2 95% LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: VICTORIAN PROVINCIAL TOWNS, Learner licence and Total Under to uoder over under NOTVIC. 1year MANUFACTURE SUNDAY DRIVER 1976 ONLY 8 years

143 TABLEB2.3 95% UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: VICTORIAN PROVINCIAL TOWNS, Learner NJA licence and Under to Total under over NOTVIC. Iyear MANUFACTURE DRIVER 1976 SUNDAYONLY 8 years

144 TABLEB2.4 NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES ON ARTERIAL ROADS FOR SELECTED PROVINCIAL TOWNS DURING AS A FUNCTION OF DRIVING EXPERIENCE Total (JO and I Under 1 to over under NOTVIC licence 1year Learner years 15 MANUFACTURE MALE DRIVER SUNDAY 1976 ONLY

145 TABLEB2.5 NUMBER OF DRIVERS SURVEYED ON ARTERIAL ROADS DURING 1994 IN VICTORIAN PROVINCIAL TOWNS AS A FUNCfION OF DRIVING EXPERIENCE Total licence 710 Learner Under IHHJ and 2 to under over lyear NOTVIC. 2 3 I I 3 I 10 I Total SUNDAY MIDNIGHT-1:59 MANUFACTURE LOW WEEKDAY DRIVER 1976 ALCOHOL ONLY 6AM-6PM TIME 8 years 24

146 TABLEB2.6 DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS AS A FUNCTION OF DRIVING EXPERIENCE III Total licence Learner Under I Ito lKl and 472 to under over year NOTVIC LOW Total SUNDAY WEEKDAY MIDNIGHT-l:59 MANUFACTURE DRlVERONLY 1976 ALCOHOL 6AM-6PM TIME years

147 TABLEB2.7 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) ON ARTERIAL ROADS DURING I994 IN VICTORIAN PROVINCIAL TOWNS AS A FUNCTION OF DRIVING EXPERIENCE NIA licence Learner Total and NIA Under to tojl79 under over NOTVIC. I year NIA MANUFACTURE SUNDAY DRIVER 1976 ONLY 8 years

148 TABLEB2.8 DISTANCE TRAVELLED DURING A 148 NON-HOLIDAY WEEK PERIOD (in million KM) ON ARTERIAL ROADS FOR VICTORIAN PROVINCIAL TOWNS AS A FUNCTION OF DRIVING EXPERIENCE O,()(X) (0) licence Learner O,(X){) O,(){)() H){) Total O.(XXJ HXJ O.()(XJ O.()(X) (0) Under and to under over NOTVIC. 1year 8 1 In under O,(lOO (0) totounder 53 MANUFACTURE DRIVER SUNDAY 1976 ONLY 8 years

149 TABLEB2.9 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) DURING 1994 AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: VICTORIAN PROVINCIAL TOWNS, Total WEEKDA Y NIGHT PM-6AM 1 and over 1 Learner 1 licence 1 8 years der 5 I5 to under 8 Iand over Learner licence I Total 8 years I INOT VIe.

150 TABLEB % LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: VICTORIAN PROVINCIAL TOWNS: Learner licence and Total to under over NOTVIC to under WEEKDAY NIGHT 6PM-6AM WEEKEND NIGHT 6PM-6AM 8 years Learner licence and Total to under over NOTVIC to under years

151 TABLEB % UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: VICTORIAN PROVINCIAL TOWNS: Learner 1.543WEEKEND licence and to Total under over NOTVIC NIGHT to under6pm-6am years Learner licence and to Total under over NOTVIC to under years

152 TABLEB2.12 NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS DURING , AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER Total and to under over NOTVIC. 6licence 35 8Learner 1 to 4 under WEEKEND Total NIGHT 6PM-6AM 1 8 years Total and to over under NOTVIC. 3licence Learner 1 to under years

153 TABLEB2.13 NUMBER OF DRIVERS SURVEYED ON ARTERIAL ROADS DURING 1994 IN VICTORIAN PROVINCIAL TOWNS, AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER Total licence Learner Under Total and to under over 1227 year NOTVIC WEEKDAY 6AM-6PM 8 years Total licence Learner and to under over 158 NOTVIC to under WEEKDAY 6AM-6PM 8 years

154 TABLEB2.14 DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS, AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER WEEKEND NIGHT 6PM-6AM Total licence Learner Total and to under over 4052 NOTVIC to under WEEKDAY 6AM-6PM 8 years Total licence Learner and to under over 5704 NOTVIC to under WEEKDA Y 6AM-6PM 8 years

155 TABLEB2.15 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS, AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER licence Learner Total and Under to to under over WEEKDAY WEEKEND NOTVIC. 1 year DAY 6AM-6PM 6AM-6PM years licence Learner Total and Under to to under over NOTVIC. 1 year WEEKDAY 6AM-6PM 8 years

156 TABLEB2.16 DISTANCE TRAVELLED DURING A 148 NON-HOLIDAY WEEK PERIOD (in million KM) ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS, AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER licence Learner Total and to under over WEEKEND NOTVIC. 58 DAY AM-6PM to under 3 WEEKDAY NIGHT 6PM-6AM 8 years licence Learner Total and to under over NOTVIC to under 5 8 years

157 TABLEB3.1 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF LICENCE TYPE: VICTORIAN PROVINCIAL TOWNS, 1994 UNLICENSED PROBATION LEARNER Total FULL MANUFACTURE DRIVER SUNDAY ONLY? /4.'1 I

158 TABLEB LEARNER Total FULL MANUFACTURE DRIVER SUNDAY ONLY ESTIMATES VICTORIAN (number AS PROVINCIAL A FUNCTION of casualtytowns, OF crashes LICENCE 95% per 1994 LOWER million TYPE: KM BOUNDS travelled) FOR CASUALTY CRASH RISK Is-a

159 ~.~.. TABLEB t UNLICENSED PROBATION os FUlL Total SUNDAY DRlVERONLY MANUFACTURE ESTIMATES VICTORIAN LEARNER (number AS PROVINCIAL A FUNCTION of casualty TOWNS, crashes OF LICENCE 95% per 1994 million UPPER TYPE; KM BOUNDS travelled) FOR CASUALTY CRASH RISK 15/

160 UNLICENSED lOO lOO I Total TABLEB3.4 NUMBER I OF DRIVERS 1 INVOLVED IN3 CASUALTY CRASHES DRIVER MANUFACTURE SUNDAY ONLY IN VICTORIAN PROVINCIALAS ATOWNS FUNCTION DURING ARTERIAL OF LICENCE ROADSTYPE FULL 1'52-

161 TABLEB3.5 UNLICENSED I I I I Total FULL PROBATION I NUMBER OF DRIVERS SURVEYED WEEKDAY MANUFACTURE Total DRIVER ONLY ON ARTERIAL MIDNIGHT-l:59 LOW SUNDAY ROADS ALCOHOL DURING 6AM-6PM VICTORIAN TIME 1994 AS A FUNCTION PROVINCIAL OF LICENCE TOWNS TYPE /53

162 TABLEB3.6 UNLICENSED FULL Total DISTANCE TRAVELLED DURING 1994IN ONAN ARTERIAL AVERAGEROADS WEEK IN(in hundred KM) VICTORIAN PROVINCIAL MANUFACTURE MIDNIGHT-l:59 SUNDAY DRIVER LOW WEEKDAY Total ALCOHOL ONLY 6AM-6PM TIME TOWNS AS A FUNCTION TYPE OF LICENCE PROBATION /5.

163 TABLEB3.7 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON ARTERIAL ROADS IN VICTORIAN P~OVINCIAL TOWNS AS A FUNCTION OF LICENCE TYPE UNLICENSEDTotal DRIVER MANUFACTURE SUNDAY ONLY FUlL 155

164 TABLEB3.8 DISTANCE TRAVELLED DURING A 185 NON-HOLIDAY WEEK PERIOD (in million KM) ON ARTERIAL ROADS IN VICTORIAN PROVINCIAL TOWNS AS A FUNCTION OF LICENCE TYPE MANUFACTURE DRIVER SUNDAY ONLY Total 1510

165 APPENDIXC CASUALTY CRASH RISK ESTIMATES AND CONFIDENCE LIMITS, EXPOSURE ESTIMATES AND STANDARD ERROR ESTIMATES, CASUALTY CRASH INVOLVEMENT AND NUMBER OF SURVEYED DRIVERS ON VICTORIAN RURAL mghw AYS

166 TABLECl.l CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: VICTORIAN RURAL HIGHWAYS, Total OA49 OA MANUFACTURE OJl DRIVER ONLY l l l SUNDAY I.() l l8 I.() /5 'I

167 TABLECl.2 95% LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: VICTORIAN RURAL IDGHW AYS, 1994 GENDER MANUFACTURE DRIVER ONLY SUNDAY (

168 TABLECl.3 95% UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVER AGE: VICTORIAN RURAL IDGHW AYS, 1994 GENDER MANUFACTURE DRIVER ONLY SUNDAY l:5 I

169 TABLE CI.4 NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES ON VICTORIAN RURAL HIGHWAYS DURING AS A FUNCflON OF DRIVER AGE GENDER I I I I I I I I I I 4 2I DRIVER Total MANUFACTURE MALE SUNDAY ONLY Ib2

170 TABLE C1.5 NUMBER OF DRIVERS SURVEYED ON VICTORIAN RURAL HIGHWAYS DURING 1994, AS A FUNCTION OF DRIVER AGE III GENDER I IO I Total LOW MANUFACTURE SUNDAY WEEKDAY DRIVER MIDNIGHT-1:59 ALCOHOL ONLY 6AM-6PM TIME

171 TABLECl.6 DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON VICTORIAN RURAL HIGHWAYS AS A FUNCTION OF DRIVER AGE GENDER oo S lool S MIDNIGHT-1:59 DRIVER MANUFACTURE Total SUNDAY WEEKDAY LOW ALCOHOL ONLY 6AM-6PM TIME 164

172 TABLE CI.7 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON VICTORIAN RURAL IDGHWAYS AS A FUNCTION OF DRIVER AGE GENDER Ul MANUFACTURE DRIVER ONLY / SUNDAY

173 TABLE C1.8 DISTANCE TRAVELLED DURING A 185 NON-HOLIDAY WEEK PERIOD (in million KM) ON VICTORIAN RURAL HIGHW AVS AS A FUNCTION OF DRIVER AGE, with distance travelled for drivers of unknown age distributed proportionately amongst driver age groups GENDER (Kl ll l.l O.lX){) O.(lOO O.IX){) O.(XX) O.(K){) l.l O.lK){) (WOO (KX) IXl O.2(K) O.<KXl O.lXX) O.<KX) lux){) (){) O.<KX lK l.l O.(KXl Im l.l X> o.om llO Ol.l l.l <KX O.llOO (){) O.<KX) O.llOO O.(K){) WEEKDAY LOW MANUFACTURE DRIVER MIDNIGHT-1:59AM SUNDAY ALCOHOL ONLY 6AM-6PM TIME (lob

174 TABLEC2.1 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: VICTORIAN RURAL IDGHW AYS, OAS Learner licence I Total OAS and D Under to over under NOTVIC. lvear MANUFACTURE DRIVER SUNDAY 1976 ONLY 8 years

175 TABLEC2.2 95% LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: VICTORIAN RURAL ffighw AVS, U Learner licence 2.l Total and l.ll99 Under toonders under over NOTVIC. 1 year MANUFACTURE DRIVER SUNDAY 1976 ONLY LICENCE TYPE 8 years

176 TABLEC2.3 95% UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE: VICTORIAN RURAL IDGHW AYS, Learner licence Total OM and Under to under over NOTVIC. tyear MANUFACTURE SUNDAY DRIVER 1976 ONLY 8 year.;

177 TABLEC2.4 NUMBER OF DRIVERS INVOLVED IN CASUAL IT CRASHES ON VICTORIAN RURAL HIGHWAYS DURING AS A FUNCTION OF DRIVING EXPERIENCE III I I Learner I Total II 1I licence I J4J I I I tounder IS I4I I and I I5I Under over NOTVIC I lyear I 15 1I I years 11 Total MANUFACTURE SUNDAY MALE DRIVER 1976 ONLY /10

178 TABLEC2.5 NUMBER OF DRIVERS SURVEYED ON VICTORIAN RURAL IDGHW AYS DURING 1994 AS A FUNCTION OF DRIVING EXPERIENCE Total IQI II Ilicence Learner I II Under to II III and to under over lyear 9 NOTVIC I1 I5 I I DRIVER SUNDAY WEEKDAY Total MIDNlGHT-I:59 MANUFACTURE LOW 1976 ALCOHOL ONLY 6AM-6PM TIME 8 yem!;q I

179 TABLEC2.6 DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON VICTORIAN RURAL IDGHW AYS, AS A FUNCTION OF DRIVING EXPERffiNCE Total licence Learner Under ] and tu to under over lyear NOTVIC years MlDNIGHT-l Total DRlVERONLY MANUFACTURE WEEKDAY LOW SUNDAY 1976 ALCOHOL 6AM-6PM :59 TIME

180 TABLEC2.7 STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON VICTORIAN RURAL IDGHW AVS. AS A FUNCTION OF DRIVING EXPERIENCE licence Learner Under ToW to to and under over I year 3NOTVIC MANUFACTURE DR1VERONLY SUNDAY years

181 TABLEC2.8 DISTANCE TRAVELLED DURING A 148 NON-HOLIDAY WEEK PERIOD (in million KM) ON VICTORIAN RURAL HIGHWAYS, AS A FUNCTION OF DRIVING EXPERIENCE licence Learner Total Under and to under over NOTVIC. lyear to under DRlVERONLY MANUFACTURE SUNDAY years

182 TABLE C2.9 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: VICTORIAN RURAL HIGHWAYS, Total Iand over Learner IWEEKEND licence I NIGHT 6PM-6AM Total 8 years I INOT VIe under 5 I5 to under 8 Iand over Learner licence I Total 8 years I INOT VIe.

183 TABLE C % LOWER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: VICTORIAN RURAL HIGHWAYS, Learner licence Total and Under 1 to to over under NOTVIC. WEEKDAY WEEKEND 1 year DAY NIGHT AM-6PM 6PM-6AM 8 years Learner licence Total and Under 1 to to over under NOTVIC. 1 year years

184 TABLE C2.tt 95% UPPER BOUNDS FOR CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER: VICTORIAN RURAL IDGHWAYS, Learner licence Total and Under 1 to to under over WEEKEND NOTVIC. 1 year DAY AM-6PM WEEKDAY NIGHT 6PM-6AM 8 years Learner licence Total and to under over NOTVIC to under 2 8 years

185 TABLE C2.12 NUMBER OF DRIVERS INVOLVED IN CASUALTY CRASHES ON VICTORIAN RURAL HIGHWAYS DURING AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER Learner 328 Total 12 licence and Under 1 to to under over NOTVIC. 16 year WEEKDAY WEEKEND NIGHT 3 6PM-6AM 1 8 years Total and to under over NOTVIC. 21 licence 38 5Learner 1 to under years

186 TABLE C2.t3 NUMBER OF DRIVERS SURVEYED ON VICTORIAN RURAL HIGHWAYS DURING 1994 AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER Total licence Learner Under and WEEKEND WEEKDAY to under over 1 year NOTVIC NIGHT NIGHT 389 6PM-6AM 6PM-6AM Total WEEKDAY 6AM-6PM 8 years Total 2licence Learner and to under over 172NOTVIC to under WEEKDA Y 6AM-6PM 8 years

187 TABLE C2.14 DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON VICTORIAN RURAL HIGHWAYS AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER Total licence Learner WEEKEND and to under overnotvic DAY 1 to6am-6pm WEEKDAY NIGHT under PM-6AM WEEKDAY 6AM-6PM 8 years Total licence Learner and to under over 4712 NOTVIC to under WEEKDAY 6AM-6PM 8 years

188 TABLE C2.tS STANDARD ERRORS FOR DISTANCE TRAVELLED IN AN AVERAGE WEEK (in hundred KM) DURING 1994 ON VICTORIAN RURAL HIGHWAYS AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER WEEKDAY NIGHT 6PM-6AM licence Learner Under Total to to and WEEKEND under over 1 year NOTVIC NIGHT PM-6AM years licence Learner Under Total to to and under over 1 year NOTVIC years

189 TABLE C2.16 DISTANCE TRAVELLED DURING A 148 NON-HOLIDAY WEEK PERIOD (in million KM) ON VICTORIAN RURAL HIGHWAYS AS A FUNCTION OF DRIVING EXPERIENCE BY TIME BLOCK AND GENDER licence Learner Total and to under over WEEKEND WEEKDAY NOTVIC. 85 NIGHT PM-6AM 2 to under 3 ~0 8 years licence Learner Total and to under over NOTVIC to under 5 8 years

190 TABLEC3.1 CASUALTY CRASH RISK ESTIMATES (number of casualty crashes per million KM travelled) AS A FUNCTION OF LICENCE TYPE: VICTORIAN RURAL HIGHWAYS, 1994 UNLICENSED PROBATION LEARNER l.0l D Total FUll SUNDAY DRIVER MANUFACTURE ONLY