'3 TRANSPORT RESEARCH LABORATORY TITLE by Preliminary analysis of motorcycle accidents: short - term impacts of the running headlights campaign and regulation in Malaysia R S Radin Umar, G M Mackay and B L Hills Overseas Centre Transport Research Laboratory Crowthorne Berkshire United Kingdom $k &%1
, -72
--
( 1 1 1 ICSPJ
O[n

--K _ C. W). ---- -- -- L....L._ - UMAR, R S, G M MACKAY and B L HILLS (1 995). Preliminary analysis of motorcycle accidents: Short-term impacts of the running headlights campaign and regulation in Malaysia. Journal of Traffic Medicine. (1995) Vol 23, No 1. JTraffic Med (1995) Vol 23, No 1 Preliminary Analysis of Motorcycle Accidents: Short-Term Impacts of the Running Headlights Campaign and Regulation in Malaysia R. S. RADIN UIMAR, G. M. MACKAY, B. L. HILLS Accident Research Unit, Universiti Pertanian Malaysia, Accident Research Centre, Birmingham University and Overseas Centre, Transport Research Laboratory. United Kingdom. Radin Umar RS, Mackay GM, Hills BL. Preliminary Analysis of Motorcycle Accidents: Short-Term Impacts of the Running Headlights Campaign and Regulation in Malaysia. J Traffic Med 1995;23:17-28. A preliminary investigation of motorcycle fatalities showed that riding a motorcycle is 17 times more dangerous than driving a passenger car. About 50 per cent of motorcycle accidents in this country occur at junctions and 38 per cent of the incidents involve other vehicles crossing motorcycle's paths. In most cases, motorcycles are found to be moving straight ahead. Daytime accidents constitute about 73 per cent of the motorcycle accidents and about two thirds of the riders involved in multiple accidents are on their right-of-way. Based on this analysis, improved motorcycle conspicuity was proposed, and a nation-wide "1running headlight" Campaign and Regulation were implemented in July and September, 1992, respectively. Detailed analysis on the impact of running the headlight campaign and regulation in the districts of Seremban and Shah Alamn revealed that there had been a sizeable drop (6.9%) in multiple vehicle-day time motorcycle accidents in the study areas. The percentage of riders switching on their lights increased sharply just after the campaign and remained at about 82 per cent by the end of 1992. Conspicuity-related accidents while motorcycles are going straight ahead or turning on the right of way, MSTOX, were found to have dropped significantly immediately after the campaign by about 22 per cent. Statistical tests done before and after the campaign confirmed that the running headlight Campaign and Regulation had a significant impact (p <0.005) in reducing MSTOX accidents in the study areas. Keywords: motorcycle accidents, injury index, colliding mechanisms, conspicuity-related accidents. conspicuity, running headlight, impacts I NTRO DU CTIO0N personal transport in Malaysia. About 60 per cent of the registered vehicles in this country are Motorcycles are one of the major modes of motorcycles and their proportions on the road 1 7 JTraffic Med (1995) Vol ?2- No. 1 Figure 1. Registered Motorised Two-Wheelers in Ma- laysia [Royal Malaysian Police] Y-a Figure 2. Motorisation and Death Rates (1987) [1] Indonesi. PflipplneS .M.1.0U. .U K Sing~oe P---rc~,o ~nod varies between 30 to 60 per cent, according to locations. The number of the motorized two-wheelers (motorcycles and scooters) registered has been increasing tremendously in past years with an average increase of about 12 per cent per year in the last 10 years as shown in Figure 1. Consequently, casualties among motor- cyclists form a large portion of the traffic injury problem (Figure 2) and are placed among the top [1] in comparison to other countries. This is primarily because the vehicle itself offers little protection to the riders and pillion passengers in the event of collision with other vehicles. Thus, research on motorcycle safety is essential so that appropriate measures can be taken to reduce the casualty and mortality rates among motorcyclists. This paper presents the analysis of motorcycle accidents and the impact of the recent `running headlight Campaign and Regulation" in the pilot project areas, Seremban and Shah Alam, Malaysia. HYPOTHESIS especially those involving riders travelling straight ahead on their right-of-way, can be used to test the hypothesis that the majority of motorcyclists are victims of traffic accidents because they are not seen by other road users. Improvement in their conspicuity may result in reducing their exposure to accidents. Consequently, detailed analyses of collision mechanisms involving motorcycles were studied and the impacts of the national Ride-Bright Campaign and Regulation were carried out. DATA COLLECTION AND ANALYSIS This study was based primarily on a specially created police accident form POL27(Pin 1/9 1). The form was designed for easier completion and fully compatible with a customised version of TRLs Microcomputer Accident Analysis Package. MAAP[2]. National accident data from the mainframe computer system at the Police Headquarters, Bukit Aman, were downloaded using a specially designed transcription program MALMAINF [3] for the overall analysis of the accident patterns, while the up-to-date Pilot Project data in the districts of Seremban and Shah Alamn [4] was used for the evaluation of the campaign and regulation. In both cases, special working files containing all motorcycle accidents were created separately for each data set for further analysis. Detailed collision types involving motorcycle accidents were reclassified and recoded based on sketch diagrams and written police descriptions available on the last page of the form. A total of 193 collision mechanisms dividing into 37 colliding groups were identified as summarised in Figure 3. The colliding group was classified according to the road geometry, vehicle manoeuvres and the riders' right way with the notations S to represent all Straight directions on the right-of-way. T for either Turning left or right on the right-of-way and X for crossing manoeuvre on losing the right-of-way. Collision types were also identified by the impact directions with notations F, B and D for the front, rear and side collisions, respectively. These recoded records were then manually updated into MAAP. The characteristics of motorcycle accidents, 18 Traffic Med (1995) Vo: .3, No 1 Figure 3. The 37 Collision Groups Defined for Motorcycle Accidents 1Colliding Metain EXn m rtCS Notation Cotliuir rg Hrr Si 5 -.5 4 I Motorcyclc Straightl P'edestrian Cros., Motor'cYcte Turn Pedestrian Cr,,ss Motorcycle Strai ght Pedestriali Obscured Motorcycle Straight. Pedestrian Miid dle of Roand Motorcycle Straight Pedestrian Walk Along Road Motorcycle Straight Pedestrian Walkc Opposite Direction Motorcycle Straight and Ilit Objects llotor-cyole Straight arid I-lit Aniaiols Motorcycle Straight arid Iit Froan back Motorcycle Turn aind lit. From Back Motorcycle Cross arid lIlt Front Back, Motorcycle Cross and Ihit Froin Front Miotorcycle Cross antd Ifit From Sidte M o o c y l T ..r . mod lilt Fromt Sidte 7W11 -pF- 7Tr7 -i n ;- LIF Ill l tr r c le T ur nci antI1 ttrt From Front I Motorcycle Srrgi Otcher tVehticles Cross Sudb Collrsoror Otol~orcpele .51iraighrt tO ther eli es C ro s, Fr.ont ColliSion Mooryce ira~r a -1r [ 7i1 MBOTtI MSOSI) h SOS F MIJOSII MEOSIJ MuSOEB MISOED MSORF MIC'rllP MtYOFF MYON11 MQ0TF` MS i H O 1. OYI hisoSI3 j b otor,, 3cle t kr c MOther StraiS tr i h S d -l e ( : o ll r s i u, t t Motorcycle strurigilt Otlier strurigltt lited-0i, Cillisroir tultot cycle lBlckrirg Other straightt Rear-End Coltriton, Motorcycle Park/Stop Other Straighrt Rear-End Col lision Motorcycle Straight Other Park/Stop Rear-End Collision Motorcycle Straightt Otlier Park, Side Cottisiorr Mrotorcycle Straight Otlier Revero Front Collisioni Motorcycle Yaw/Fall rind lit Pedentrian, mrotorcycle a /F l Off the Road Motorcycle Yaw/Fall On thre Rood Motorcycle Qiuctr at. Stop Line Other Turn Hit Froms Front Motorcycles Straighrt Other Yaw Head-On Cillisio.i llotorcycle V-n/F.11 Other Straight [tea rt- 0 % C ot, i i Motoorcycle Straight Straight Otlier Yaw Rear End Collision Mtot or.cycle a /F l Oth e r tr ig t Rear-End Clrrr Eon mpIC le -i(~ 71~, -TI.- -zILL_ £iti It-I-  a r .0. MEOTIB Mot~orcycle Park -- ~ ~~~~~Otilrer T ..rr. - I~~~~~~~~~~~-1 19 . . . .-. . . . Notation MCSP~x hMCTIPX MSPOI3 MSPMR LISPIVA MSPIVO td~liOJ MCSIIA MCSIIR MCTHIB MCXI4B 1ACKttF IICXHO 1ICT1I1) MCTI1F 11SOXI) R15OXIF ---~l -- 4-;J_ .
,=
i 7E7
- .2---.
e- . Z-;, ;c-
- ~0
1 1 JTraffic Med (1995) Vol 23, No. 1 Table 1Injury Risk for Occupants of Motorised Vehicles in Malaysia Vehicle Injuries Vehicle Index (per 1000 vehicles) Relative Risk Type Fatal Serious Slight Involved Fatal Serious Slight Fatal Serious Slight Car 227 522 2057 63674 3.6 8.2 32.3 1.0 1.0 1.0 Van 100 187 541 11324 8.8 16.5 47.8 2.5 2.0 1.5 Motorcycle .1255 4153 8162 20810 60.3 199.6 392.2 16.9 24.3 12.1 lorry 109 169 520 13119 8.3 12.9 39.6 2.3 1.6 1.2 Bus 14 30 77 4837 2.9 6.2 1 5.9 0.8 0.8 0.5 R E S ULTS Injury Risk for Occupants of Motorised Vehicles in Malaysia The accidents and injuries involving motorised vehicles successfully downloaded from the mainframe computer throughout Malaysia in 1992 are shown in Table I. Motorcycle casualties constitute about 76 per cent (1 3,570 cases) of the total casualty injuries of motorised vehicles. This is followed by about 16 per cent (2806 injuries) passenger car casualties and about 5 per cent (828 cases) and 5 per cent (798 cases) for the passenger van and commercial vehicle casualties respectively. The relative risk of injuries for the occupants of different classes of vehicle involved in accidents were calculated from the ratio of the Figure 4. Collision Mechanisms With Motorcycles I KS i M F1. injury index per 1000 vehicles of a particular type of vehicle over the injury index of a passenger car. These confirmed further the fact that the chance of injury amongst the motorcyclists are much higher than other types of motorised vehicles. This is illustrated in the last column of Table I with the relative risk being 17 and 24 times greater than passenger cars for (i) motorcycle fatalities and (ii) serious injuries' respectively. This is followed by passenger vans and commercial lorries both with fatality indices of about 3 compared with passenger cars. Collision Mechanisms With M ot or c ycle s The severity of accidents against collision group as coded in Figure 3 is shown in Figure 4. Crossing accidents while motorcycles are travelling Er 20 JA L JTraffic Med (1995) Vol 23, No 1 Figure 5. Colliding Mechanismns and Casualty Injuries With Motorcycles Collision N mer of suahlies ~Mechanis Fatal Serious Sliqht Total 17 -91 208 - 316 1 2 4 7 0 2 5 7 4 9 15 28 0 8 19 27 0 0 1 1 6 8 21 35 5 8 19 32 30 100 205 335 2 15 34 51 9 55 111 175 1 4 61 90 165 14 79 137 230 1 15 31 47 0 3 5 8 31 211 460 702 5 25 20 50 5 34 81 120 0 23 50 73 4 29 85 118 44 90 123 257 12 102 200 314 1 0 3 4 4 19 22 45 0 5 1 4 19 2 6 7 15 1 4 2 7 37 47 89 173 10 1 5 35 60 0 4 4 8 2 9 10 21 5 20 18 43 2 1 3 6 0 2 4 6 1 0 2 3 0 0 0 0 0 0 0 0 269 1102 2137 3508 Number of Casualties 0 100 200 300 400 S00 600 U) 0 04r. MCspXMCTPXMS P0 1 MSPMR,MSP WA MSPWOMS H 0i McsHAMCSH B MCTIBMCXHBMCXHFMCXH D MCTHDMCTHFMSOXOMSOXFMSOXBMOOITBM50S0MsosPMSOSBMPOSBMOOPBOMS0MDMSORPFCY14PMYCOFFMY ONPRMOOTFM5OYFMYC00SFMOOYBMY OSBMPOTBMTOTDM5OT1B straight ahead and other vehicles cross their travelling paths (MSOXID) constitute about 22 per cent (699 cases) of accidents investigated. This is followed by about 12 per cent (367 cases) rear-end collisions with motorcycles [motorcycles moving straight ahead and struck from the back (MCSH B)]. Another 1 0 per cent (31 5 cases) of the accidents are rear-end collisions with other vehicles [motorcycle backing other straight (MB0SB13 and about 7 per cent (213 cases) is the pedestrian-motorcycle accidents while motor- cycles are moving. straight ahead (MCSPX). The analysis of the injuries to motorcyclists for the 37 collision groups is shown in Figure 5. Although the figures of casualties are generally higher than the number of accidents, the proportion of casualties according to the collision mechanisms are quite comparable to the proportion of accidents. Casualties resulted from crossing accidents while motorcycles are 21 700 800 MACTIPXMSPOBMVSPMVRIVSPWAMVSPWOMSHOJMCSHAMCSHBMCTHBMCXHBMVCXI-F'MVCXHIDMCTHDMCTHFMVSOXDIVSOXF'MVSOXMSOTBMSOSDMSOSFMSOSBMPOSBMSOPBMVSOPIDMSORFMVCYHPMYOFFMYONRMQOTF'MVSOYF'MYOSFMVSOYBMYOSB3MPOTBMVTOTI)MSOTBTOTAL 1 9 1 1 1 1 1 11in.9 1111 59 m in1 :
, 1 " ;., 1 7 ER CH9112aJ

1:1911m11 J1 Traffic Med (1995) Vol 23, No. 1 Table 11 Collision Group and Conspicuity-Related Accidents Collision Group with Motorcycles Priority Fatal Serious Slight Damage Total ConspicuityRelevant Motorcycles With Pedestrians and Other Vehicles 2502 Motorcycle Straight/turn other vehicles cross Motorcycles 38 254 421 238 951 Yes Motorcycle Straight/turn Pedestrians cross Motorcycles 22 100 133 6 261 Yes Motorcycle Straight/turn Struck from Back Motorcycles 28 85 179 129 421 No Motorcycle Cross Main Paths Others 30 156 196 94 476 No Motorcycle hit other vehicles from Back Others 12 105 154 122 393 No Single Motorcycle Accidents 431 Motorcycle yaw/skid/fall NA 46 69 1OS 63 283 No Motorcycle Hit Animal/Object NA 14 32 56 46 148 No Others NA 40 83 ill 64 298 NA Total 230 884 1355 762 3662 travelling straight ahead and other vehicles cross their travelling paths (MSOXID) still ranked the highest (701 cases) followed by rear-end casualties with motorcycles [motorcycles moving straight ahead and struck from the back (MCSH-B)] (335 cases) and the pedestrian casualties (316 cases). Conspicuity-Related Accidents To understand the collision characteristics further, the 37 collision groups were classified into two collision types known as 'motorcycle to pedestrians and other vehicles accidents', (MPVA) and 'single motorcycle accidents", SMA. These are shown in Table 11 and Figure 6. The former group was further classified into two categories known as `conspicuity-related' accidents and "non conspicuity-related` accidents. The `conspicuity- related accidents" consists of all accidents involving motorcycles moving straight or turning with the right of way when pedestrians and other vehicles cross their paths. The above classifications are necessary to differentiate between accident situations in which running headlights could have potentially improved safety and those in which they probably would have been irrelevant. 'Motorcycle with pedestrians and other vehicle" accidents, MPVA, constitute about 77 per cent of accidents investigated (2502 cases) while less than one quarter of the investigated Figure 6. Collision Groups With Motorcycles Mtooryck SWaitIUAw. Pd~i.- ~ (104%) Mot0~Y& tw.ibththa St-c± f-o B.dt (26.8%l) Motmqyd, SDr.igbft~we oh~ 'hid,.d ". (3&0%) Mo(omyd, bit ot5h- bi, fro B.,k I 57% Moto.cde cm'. Msin Pths (19tO%) 22 JTraflf Med (1995) Vol 23. No 1 accidents constitutes the single motorcycle accidents, SMA. A total of 951 cases (38 per cent) of the MPVA accidents occurred when motor- cycles were travelling straight or turning on their right-of-way and collided with other vehicles. An additional 261 cases (about 10 per cent) were motorcycle-pedestrian accidents and another 421 cases (17 per cent) involved rear-end collisions with motorcycles. Of the 2502 MPVA cases, a total of 1633 (about 65 per cent) involves motorcyclists who were on their right-of-way and colliding with either pedestrians or other vehicles. Out of those accidents, a total of 1212 (about 74 per cent) of motorcycles involved in the conspicuity-related accidents. It was noted also that one-third of the MPVA were the results of the direct riders' negligence crossing the paths of other vehicles or backing too close to the front vehicles in the rear-end situations. Daytime Running Headlight Campaign and Regulation There is a widespread belief that motorcycles are more difficult to detect in traffic than other motorised vehicles. Studies of individual collisions involving motorcycles [5], indicate that drivers who violate motorcyclists' right-of-way often claim not to have seen them before the collision ("looked but failed to see"). The conspicuity of motorcyclists has been of concern for many years Figure 7. Rate of Compliance for Daytime Running Headlight' 100 00'IE0 j 01 30 and has resulted in a number of `Running Headlights" campaigns in which motorcycle headlights are switched on in the daytime. These have been shown to reduce motorcycle collisions in selected countries 15,6.7,81. In view of the high proportion of motorcycles in Malaysia, it was anticipated that a campaign to improve the conspicuity of motorcycles would have a clearer impact than in developed countries when there is a smaller proportion of motor- cycles. Since the majority of the accidents occurred while motorcycles had the right-of-way and particularly while they were travelling straight ahead or turning, it can be concluded that in the majority of traffic accidents in which they are involved, motorcyclists tend to be the victims of errors made by other road users. With the above evidence that improving the conspicuity of motorcycles reduces accidents, a nation-wide "Daytime Running Headlight" campaign was proposed and this was launched in July 1992. This was followed by the compulsory use of headlights regulation in September 1992. Rate of Compliance for Running H ea d I ig9 ht s The percentage of motorcycles complying to the running headlights campaign during the day time in the districts of Seremban and Shah Alam is shown in Figure 7. Samples of at least 100 motor- cyclists, each at eight sampling points in Month. 23 0 J iroi fic Med (1 995) Vol 23, No. 1 Seremban and Shah Alam, were taken each month starting from June 1992. The percentage of motorcyclists riding with lights switched on increased sharply just after the campaign and by the end of 1992, it stood at 82 per cent. The proportion of all riders using their main beam or dipped light was consistently maintained at about 45 and 37 per cent, respectively, throug- hout the campaign and regulation periods. Multiple-Day and Night Accidents With Motorcycles The pattern of day and night time multiple accidents with motorcycles in Seremban and Shah Alam is as in Table Ill. The cumulative plot of accidents before and after the campaign and the respective projected mean on the basis of data six months before the campaign against time are shown in Figure 8. By comparing the predicted mean trend line with the actual cumulative number of the multiple-day time accidents, it can be seen that there is a distinct separation for the cumulative number of multiple-day accidents just after the campaign in July 1992. The upward or downward curvature of the cumulative plot indicates a corresponding short term trend and changes in slope indicate long term changes in the rate of occurrence of the accidents investigated [9]. In contrast, the number of all motorcycle accidents and multiple-niqht accidents lie close to the straight lines of corresponding cumulative means, implying that these types of accidents fluctuated very closely about a steady average over the same period. Table filt Multiple-Day and Night Time Accidents With Motorcycles Collision Types Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec All Motorcycle Accidents 141 124 181 130 130 1 56 135 149 122 152 129 164 CumulativeAll Motorcycle Accidents 141 265 446 576 706 862 997 1146 1268 1420 1549 1713 Mean All Motorcycle Accidentsl141 285 428 572 716 860 1003 1147 1291 1434 1578 1722 Multiple-Day Accidents 102 82 121 89 92 104 82 93 79 94 91 110 CumulativeMultiple-Day Accidents 102 184 305 394 486 590 672 765 844 938 1029 1139 Mean Multiple-Day Accidents 102 200 299 397 495 594 692 790 889 987 1085 1184 Multiple-Night Accidents 1 7 16 28 1 1 1 5 20 23 22 13 21 17 26 CumulativeMultiple-Night Accidents 1 7 33 61 72 87 107 130 1 52 165 186 203 229 Mean Multiple-Night Accidents 1 7 35 53 70 88 106 124 142 160 177 195 213 Figure 8. Cumulative Plotof Multiple-Day and Night Time Motorcycle Accidents Against Month of Year - -- ~~~~~~~~~~ ~ ~~~~C~~~~kM-C,- i Month of the Y'car (1992) 24 Traffic Med (1995) Vol 23, i~u 1 Table IV Right of Way Accidents With Motorcycles Collision Types Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec All Motorcycle Accidents 141 124 181 130 130 1 56 135 149 122 152 129 164 CumulativeAll Motorcycle Accidents 141 265 446 576 706 862 997 1146 1268 1420 1549 1713 Mean All Motorcycle Accidents 141 285 428 572 716 859 1003 1147 1290 1434 1578 1721 MSTOX Accidents 68 48 79 44 61 73 53 43 38 49 40 68 CumulativeMSTOX Accidents 68 116 195 239 300 373 426 469 507 556 596 664 Mean MSTOX Accidents 68 130 192 255 31 7 379 441 503 565 628 690 752 Other Types of Accidents 73 76 102 86 69 83 82 106 84 103 89 96 CumulativeOther Type of Accidents 73 149 251 337 406 489 571 677 761 864 953 1049 Mean Other Types Accidents 73 155 236 318 399 481 562 644 725 807 888 970 Figure 9. Cumulative Plot of MSTOX and OTHER Accidents Against Month of the Year E~~~~~~~~~~~~~~~~~~~~~~~~~~~~2 Conspicuity-Related Accidents (M STOX) Table IV and Figure 9 show the number and the cumulative plots of motorcycle accidents with respect to the riders' manoeuvres, collision types and the right-of-way respectively. The number of accidents involving motorcycles moving straight or turning when other vehicles or pedestrians cross their paths (MSTOX) dropped sharply with an overall drop of about 22 per cent (see Table V) six months before and after the campaign. A similar trend was also noticed on the cumulative plot. The clear downward diversion for the MSTOX accidents just after the campaign in July 1992, indicates the positive changes in the rate of occurrence of MSTOX accidents. Conversely, the number of other types of motor- cycle accidents were found to increase steadily after the campaign while the overall accidents lie close to the straight lines of corresponding cumulative means, inferring that the overall accidents fluctuated very closely about a steady average over the observation period. Since the number of multiple accidents while motorcycles are moving straight or turning on the right-of-way is decreasing and there has been an upward toll on the other types of collisions it can be deduced that the campaign and regulation resulted in a reduction to the MVSTOX accidents in the study areas. Bef ore a ndc Af ter Ana lysis5 The number of different types of motorcycle accidents and injuries six months before and after 25 JTraffic Med (1995) Vol 23, No. 1 Table V Before and After Analysis on Motorcycle Accidents Motorcycle Accidents Multiple-Day MSTOX Al Motorcycle injuries MSTOX All Before Period (January-June) 590 373 862 303 635 After Period (July-Decemnber) 549 291 851 239 676 Chi Square (df=l) 2.06 9.19 11.7 Probability p <0.2 p <0.005' p <0.001 the campaign and regulations are shown in Table V. The total motorcycle accidents and correspond- ing injuries within the same areas are also presented for the control data. Based on the Chi Square values, it can be concluded that there has been a significant effect on the reduction on mo- torcycle accidents and injuries involving MSTOX (motorcycles moving straight or turning and other vehicles or pedestrians crossing their paths). The computed Chi Square values for accidents and injuries for one degree of freedom are 9.19 (p <0.05) and 11.7 (p <0.001) respectively. The drop in multiple-day time accidents however was only indicative (p <0.2) at 0.2 level of significance. DIS 5C U SSIO0N It was anticipated that the running lights campaign for motorcycles could only improve multiple vehicle. accidents with motorcycles. No improvement was expected in the multiple-night time accidents since almost all motorcycles are using their headlights for the night riding. The biggest impact expected in this exercise however, was the change in the right-of-way accidents involving motorcycles. This is because about two thirds of multiple motorcycle accidents occurred while the motorcyclists had the right-of-way and collided with either pedestrians or other vehicles. The time series analysis showed a sizeable drop in the number of multiple-day. accidents with motorcycles (6.9%), comparing six months before and after the campaign. Nationally, this will result of saving about 13,177 motorcycle accidents in 1992. This reduction is indicated by the downward diversion of the cumulative curve. On the other hand, the number of night-time and overall accidents lies very close to the mean values of the computed accidents over the same period of observation. These results support the conclusion that the campaign and regulation reduced multiple motorcycle accidents during the day time. The biggest impact anticipated from this campaign and regulation was the reduction of conspicuity-related accidents, MSTOX. The over- all drop for this type of collision six months be- fore and after the campaign and regulation was about 22 per cent. Since the number of multiple accidents while motorcycles are moving straight or turning on the right-of-way is decreasing, while there has been a continuing upward trend of other types of collisions, it can be concluded that the camnpaign and regulation must have given a significant reduction to the motorcycle accidents in the study areas. This conclusion is further supported by the 'before' and 'after' analyses which showed highly significant reduction for MSTOX accidents. Although there has been some drop in the multiple vehicle-day time accidents it is still not conclusive that the campaign and regulation have changed these particular types of accidents, significantly. However, at 0.005 level of significance, it is certain that the reduction in the multiple accidents and the associated injuries, particularly for motorcycles travelling straight ahead or turning on the right--of-way and other vehicles or pedestrians crossing their travelling paths, does not occur by chance. Therefore, it is concluded that the "running headlight" campaign and regulation has been successful in improving motorcycle safety in this country. 26 Campaign & Regulation Traffic Med (1995) Vol 23. No 1 Whilst an overall benefit was predicted, it can also be anticipated that the campaign and regulation might reduce the safety of those remaining motorcyclists who continue to travel without their lights on. Pedestrians and drivers of other vehicles who become used to the running headlights may not be expecting motorcyclists travelling without their lights on. This could result in a higher risk of accident for those motorcycles not complying with the regulation. Motorcycles without their lights could be due to riders forgetting or choose not to switch on their lights or to faulty lighting systems or blown bulbs. To overcome these problems, it is recom- mended that passive measures, such as the use of headlights which automatically turn on when the engine is started, be made available by the manufacturers. A panel indicator to indicate any failure of lighting system and spare bulb facility should be considered for the future motorcycle design. In addition, riders should be encouraged to use their high beams during the daytime and should be informed about the advantages of this. Police enforcement should also be carried out regularly to improve further the rate of compli- ance. CONCLUSIONS The following conclusions can be drawn: i.Casualties among motorcyclists form the largest part of road traffic casualties in Malaysia. The rate for motorcycle fatalities was found to be 17 times higher than for passenger cars. Serious attention to this group is thus strongly justified. ii. Crossing accidents while motorcycles are travellings traight ahead and other vehicles cross their travelling paths (MSOXD) constitute the highest type of collision mechanisms investigated. Further grouping based on the riders' right-of-way revealed that about two thirds of motorcyclists involved in multiple accidents were on their right-of-way and their travelling paths were intercepted by other vehicles and pedestrians. Since the majority of the accidents happened while motorcyclists were following their right-of-way, it can be concluded that the majority of these motorcyclists are the victims of the situations and additional attention to this problem is required. ii.Detailed evaluation of the campaign and regulation revealed that there has been a decreasing trend on multiple- day time accidents in the study areas. A larger drop was found for multiple motorcycle accidents in which motorcycles are moving straight or turning with right-of- way, MSTOX. Although the statistical analysis of the multiple vehicle-daytime accidents is only indicative, a highly significant (p <0.005) reduction in MSTOX accidents was found and it is concluded that daylight use of headlights should be encouraged as one of the best ways to improve motorcycling safety in Malaysia. In view of the increasing number in motorcycles in developing cities, it is proposed that these positive findings be seriously taken into conside- ration in drawing up low-cost countermeasures to reduce motorcycle accidents and injuries in the developing countries. REFERENCES 1. Aminuddin A, Radin Umar RS. Analysis of Traffic Accidents in Malaysia. In: Proceedings of the National Road Safety Seminar, Kuala lumpur, 1991:1-144 to 1-1 54. 2. Radin Umar RS, Baguley C.J. The Prioritisation, Identification and Analysis of Accident Blackspot in Malaysia. Journal of Road Engineering Association for Asia and Australasia REAAA. 1994. 3. Baguley CJI. TRI- Support for MAAP Pilot Project in Malaysia. Crovwthorne, United Kingdom. 1992; Technical Paper OU/TP1188192, MKJR/TRL. 4. Radin Umar RS, Aminuddin A. A Pilot Project on the Accident Diagnostic System in Malaysia. National Road Safety Council, Malaysia. 1992; Research Report No 1, JK3P. 5. Olson PL. Motorcycle Conspicuity Revisited. Human Factors 1989;31(l):141-146. 6. Huebner ML--. Rosta's Motorcycle Visibility Campaign: its Effect on Use of Headlights and High Visible Clothing on Motorcycle Involvement and on Public Awareness. Linthicum USA: Motorcycle Safety Foundation. 1980:2:715-739. 7. Thomson GA. The Role Frontal Motorcycle Conspicuity has in Road Accidents. Acc Anal & Prey 1.980; 1 2(3):1 65-178. 27 J Traffic Med (1 995) Vol 23, No. 1 8. Williamns MJ. Motorcycle Conspicuity and Traffic Accidents. Acc Anal & Prey 1979; 1 1(3)209-224. 9. The Institution of Highways and Transportation. Guidelines for Urban Safety Management. 1990:112-113. Received Revised Accepted September 14, 1993 May 15. 1994 September 5. 1994 Correspondence Dr. G. M. Mackay Accident Research Centre School of Manufacturing and Mechanical Engineering University of Birmingham Edgbaston, Birmingham B1S 2TT England 28