Age and contributing factors to unlicensed teen motorcycling
Tsu-Hurng Yeh
a,*, Hsin-Li Chang
b,1aResearcher of Institute of Transportation, Ministry of Transportation and Communications, 7F, 240, Tunhwa North Road, Taipei 105, Taiwan, ROC bDepartment of Transportation Technology and Management, National Chiao-Tung University, 1001, Ta Hsueh Road, Hsinchu 300, Taiwan, ROC
Received 20 March 2007; received in revised form 17 February 2008; accepted 19 February 2008
Abstract
This study aims to examine the initial age when teenagers engage in unlicensed motorcycling in Taiwan and the factors contributing to their unlicensed behavior. Data were collected from the retrospective experience of a cohort of senior high school students and a Cox regression model was applied. The results indicated that an estimated 63.4% of the students had experienced riding before the legal age of 18 years. Contributing factors such as whether the students were in a vocational senior high school, male, and in households with higher motorcycle ownership rates led to an earlier riding age; in contrast, a higher degree of parental monitoring, fewer motorcycles in the household, living in Greater Taipei, and an increase in the bus density in the district delayed the ages of beginning motorcycling. The higher minimum licensing age of 18 in Taiwan has caused a debate about its effectiveness because of the high prevalence of unlicensed teenage riding. Measures such as parental monitoring, power- or speed-limited mopeds, training programs, and no duo-passengers should be carefully examined if a policy is to be made on lowering the age limit of motorcycling.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Initial age of motorcycling; Minimum licensing age; Unlicensed riding; Cox regression method
1. Introduction
Motorcycle riders have an increased likelihood of acci-dents compared with other motor vehicle drivers ( Manner-ing and Grodsky, 1995; Plase`ncia et al., 1995). In particular, younger groups have a disproportionately higher accident risk when riding a motorcycle than when driving a car. For example, moped riders and car drivers aged 18–19 years experienced 1.58 and 0.12 casualties per million kilometers traveled, respectively, during 1995– 1997 in the Netherlands (see Appendix 2 from Schoon, 2004), showing that young moped riders had 13 times the casualty risk encountered by car drivers of the same age. Adolescent riders exhibit higher risk behavior (Chesham et al., 1993; Rutter and Quine, 1996) and have less experi-ence (Chang and Yeh, 2007), so the minimum licensing age
and measures regulated by the licensing system may influ-ence the accident risk experiinflu-enced by these young riders.
The motorcycle licensing system in Taiwan classes motorcycles according to engine capacity: mopeds (550 cc), light motorcycles (>50 to 5250 cc) and heavy motorcy-cles (>250 cc). The minimum licensing age for mopeds and light motorcycles in Taiwan is 18 years, whereas for heavy motorcycles the minimum age is 20 years (Ministry of
Transportation and Communications, 2002). Mopeds and light motorcycles (especially those with engine capacities less than 150 cc) are popular in Taiwan because of their easy operation and low cost for short trips, even though they are prohibited from the expressways and freeways. At the end of 2003, the total number of motorcycles com-prised 67% of all motor vehicles (1.8 motorcycles per household) (MOTC, 2004).
Disqualified motorcyclists under the legal age limit of 18 years are prevalent in Taiwan.Chang (1996)reported that around 38.8% of Taiwan’s junior high school students and as many as 68.9% of senior high school students rode a motorcycle without a license at some stage. However, 0925-7535/$ - see front matterÓ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssci.2008.02.002
*
Corresponding author. Tel.: +886 2 23496856; fax: +886 2 25450429. E-mail addresses: barryyeh7@hotmail.com (T.-H. Yeh),
hlchang@cc.nctu.edu.tw(H.-L. Chang).
1 Tel.: +886 3 5731908; fax: +886 3 5712365
www.elsevier.com/locate/ssci Safety Science 47 (2009) 125–130
Taiwan appears not to be a special case around the world. Even though a learner license can be granted at age 15 in New Zealand, Reeder et al. (1995) pointed out that most licensed riders (86%) had ridden on public roads before being licensed.
Most western countries allow adolescents aged around 16 years or less to use speed-limited mopeds (Schoon, 2004) or restrict their riding to lower risk circumstances such as having curfews or no carriage of passengers, and a stricter blood alcohol concentration (BAC) limit for the learner’s stage (Haworth and Mulvihill, 2005). On the other hand, Taiwan has set a higher entry age of 18 years to postpone adolescents from gaining access to motorcycle use. However, a higher age limit does not necessarily pre-vent adolescents who do not meet the required age from riding a motorcycle unlicensed.
These unlicensed young riders may have a higher likeli-hood of being involved in an accident because of their lack of experience or immature state of mind. This raises con-cerns over not only the safety problems of these very young groups but also the effectiveness of the motorcycle licensing system to set a stricter licensing age. It is valuable to inves-tigate the age of beginning unlicensed riding, and contrib-uting factors, to develop effective countermeasures to mitigate this problem. Therefore, this study aimed to exam-ine the initial unlicensed age, associated with related fac-tors, using retrospective self-report data from senior high school students. Differences in the initial unlicensed experi-ence contributed by various factors were examined by a Cox regression method.
2. Methods 2.1. Data
We applied a retrospective self-report survey on the atti-tudes of senior high school students to motorcycle use, which was administered by Taiwan’s Institute of Transpor-tation (IOT), to collect the students’ experience in motor-cycle use and opinions on lowering the minimum motorcycle licensing age (IOT, 2006). A stratified random sampling method was employed, based on the population distribution of the 16–18 year cohort by districts. The
pre-designated number of samples (overall sampling error was controlled within ±3%) was then distributed to the stu-dents affiliated with two randomly sampled schools in each of 23 distinct administrative districts, since most of them fall into this age range. A total of 1747 effective samples were collected.
Some answers in this study were retrieved from the ques-tionnaire (Table 1). The core questions were whether the sampled students had experienced unlicensed motorcycle riding and, if so, what was their initial age and schooling stage. To explore the regression relationship between the beginning age of unlicensed riding and contributing fac-tors, some variables such as school type, gender and family factors (i.e., living with parents, parents’ attitude and household’s motorcycle size) were also introduced.
The basic statistics of the sampled data are listed in
Table 2. The mean age of the students was 17.2 years. Of these, 45.4% were vocational high school students and 54.6% ordinary high school students; 58.8% of them were male and 41.2% female. Most of the sampled students lived with both of their parents (85.3%). Those living with a sin-gle parent or with neither parent, however, comprised a smaller proportion. Of the sampled students, 40.6% per-ceived that their parents had a ‘‘neutral” attitude towards
their motorcycling, 33.9% felt that their parents
‘‘approved” and 25.5% perceived that their parents ‘‘disap-proved”. Most students expressed that their household had one (34.0%) or two motorcycles (37.1%) and only 9.3% did not have any motorcycles in their household.
More than half of the sampled students (53.1%) had experienced unauthorized motorcycling and their initial age of unlicensed riding was around 16 years. Of those with unlicensed experience, most experienced riding a motor-cycle at the junior high stage (49.4%), but 13.1% of stu-dents had experienced motorcycling at a very early stage in elementary school.
2.2. Cox regression and model specification
The initial unlicensed motorcycling age is a time-to-event process with two kinds of possible outcomes in this study: an event or censored states. An event observation for a sampled student indicates the initial age of experienc-Table 1
Research related questions 1. Gender: h male; h female.
2. School type: h vocational senior high; h ordinary senior high. 3. Age: ___ years ___ months.
4. What is your parents’ (or senior members in your family if parents are not available) attitude towards your riding a motorcycle? h approval; h neutral; h disapproval.
5. Whom do you live with? h with both parents; h with father only; h with mother only; h with neither parent.
6. How many motorcycles in your household? h none; h one motorcycle; h two motorcycles; h three motorcycles or more. 7. Have you ever experienced unlicensed motorcycle riding? h yes (continue no. 8); h no;
8. If answered ‘‘yes” in no. 7 question, continue answering the following question:
8-1. What was the schooling stage for your initial trial of unlicensed motorcycle riding? h elementary school; h junior high school; h senior high school.
ing unlicensed motorcycling. A censored observation, how-ever, indicates a sampled student has not yet experienced motorcycling and his/her age at the time of the survey was the censored time.
The Cox regression model is comprised of two parts: the baseline hazard and an exponential component. Since the baseline hazard part is unspecified, the Cox regression belongs to the semi-parametric class of models. Its func-tional form is expressed as (Kleinbaum, 1996):
hðtjX ; bÞ ¼ h0ðtÞ eXb ð1Þ
where h0(t) is the baseline hazard function and X and b are
vectors of the independent variables and their correspond-ing parameters.
The estimated parameters ^bare obtained by maximizing the joint probability of the partial likelihood function L(b): LðbÞ ¼Y k j¼1 eXðjÞb X l2Rj eXlb , ! ð2Þ where k is the number of events (unlicensed riding) during the observation period; Rjis the set of students at risk (the
risk set) at the jth rank of event time (5k), which represents the set of students that have not had any unlicensed
expe-rience before the jthrank of time; X(j)is the vector of
inde-pendent variables corresponding to the event occurring at the jth rank of time (the number j in parentheses represents the rank order of time); and Xlis the vector of independent
variables corresponding to the lth set of students in Rj.
The estimated parameters ^bdrawn from the above func-tion, while not fully statistically efficient, are consistent and in large samples also asymptotically normal and unbiased. When comparing the relative hazard contribution of two distinct independent variable vectors, it is conventional to use the hazard ratio. This is expressed by Eq. (3), which is constant since h0(t) cancels out the equation (the
propor-tional hazard assumption). HR¼ ^h tjX1; ^b . ^ h tjX2; ^b ¼ eðX1X2Þ^b ð3Þ
where X1and X2stand for the two values of the
indepen-dent variable vector considered. If HR is significantly greater than 1, then the conditions described by X1 are
more likely to experience unlicensed motorcycling than the conditions given by X2. In other words, the age of
expe-riencing unlicensed riding is earlier under X1. For
conve-nience of comparison, we usually use one unit increase for each variable, one at a time, to estimate the hazard ra-tio, and thus HR in Eq. (3) reduces to e^bi for the ith
variable.
Since the independent variables gathered in this study mostly reflect the status of the students at the time of the survey, these one-wave investigation data assume the inde-pendent variables are time-invariant. However,
time-vary-ing independent variables obtained by multi-waves
investigation may be more accurate for some predictors that change over time. The limitation of the time-invariant assumption in our Cox regression model should be released in future studies.
Table 3 indicates the independent variables introduced in the Cox regression model and their expected associations with the hazard of the occurrence of unlicensed riding. Since students in a vocational senior high school have less pressure to advance to a higher education, they are more expected to take part in extracurricular activities (Chang, 1996) and thus have a higher hazard in unlicensed motor-cycling (or earlier experience). Male students are assumed to have a higher likelihood of exhibiting risky behavior than females (Harre´ et al., 1996; Lam, 2003) and this may lead to an earlier unlicensed experience. In addition, students living with both parents or with parents showing a disapproving attitude towards their motorcycling may represent a stricter monitoring from their parents com-pared with the reference groups, respectively (i.e., living with neither parent or with parents showing an approving attitude). According to Hartos et al. (2000), the higher degree of monitoring function is assumed to have a nega-tive association with the unlicensed hazard. Students living in Greater Taipei have more convenient public transit sys-tems and this is speculated to delay their average age of unlicensed riding. Also, according to Reeder et al. (1995), Table 2
Basic statistics of the sampled data
Variable Category Frequency
(percentage)/ mean (S.D.)
Age Mean (S.D.) 17.18 (0.95)
School Type Vocational senior high 793 (45.4%) Ordinary senior high 954 (54.6%)
Gender Male 939 (58.8%)
Female 658 (41.2%)
Living with parents With both parents 1476 (85.3%) With father only 64 (3.7%) With mother only 119 (6.9%) With neither parent 72 (4.2%)
Parents’ attitude Approval 544 (33.9%)
Neutral 653 (40.6%)
Disapproval 409 (25.5%) Household’s motorcycle size None 162 (9.3%)
One 592 (34.0%)
Two 646 (37.1%)
Three or more 339 (19.5%)
Living district Greater Taipei 586 (33.5%)
Other districts 1161 (66.5%) Unlicensed riding experience Yes 868 (53.1%)
No 766 (46.9%)
Schooling stage of the first unlicensed experience (if any)
Elementary stage (age 7–12)
111 (13.1%) Junior high stage
(age 13–15)
418 (49.4%) Senior high stage
(age 16–18)
317 (37.5%)
none or fewer motorcycles owned in a household may reduce motorcycle availability for the students and decrease the unlicensed hazard.
In addition, different district traffic compositions mea-sured by passenger car, motorcycle and city bus ownership rate represent the diverse mobility patterns where the stu-dents live. Therefore, the substitute modes such as a greater number of passenger cars and city buses owned in a district reduce the amount of students’ unlicensed motorcycling and delay their initial riding age. In contrast, in districts with a greater number of motorcycles owned, the students’ hazard in experiencing unlicensed riding was higher. 3. Results
The age of beginning unlicensed motorcycling for a cohort of senior high school students was analyzed using a Cox regression method. Some associated factors affecting the age of beginning motorcycling were examined and the estimated survival curve offered the difference of initial age experienced by different groups.
3.1. Cox estimated results
Table 4 shows the Cox estimated results. The censored rate was 47.8%, which means over half of the sampled stu-dents had experienced unlicensed riding. As expected, the speculated associations contributed by the independent variables almost all agreed with our assumptions. The haz-ard ratio between two levels of an independent variable was expressed as eb, one at a time while controlling for other variables. Students affiliated with a vocational senior high school had 36.8% (=100% (1.368 1)) more hazard in experiencing unlicensed riding compared with those in an ordinary senior high school, indicating an earlier start at motorcycling. Males experienced 61.1% more hazard than females.
Students living with both of their parents had a hazard reduced by 37.9% compared with those living with neither parent. Living with single parent also had a negative
asso-ciation with the hazard but was not statistically significant. Parents perceived by their children as showing a disapprov-ing or neutral attitude also decreased the hazard by 26.8% and 16.5%, respectively, compared with children of parents with an approving attitude. Fewer motorcycles in a house-hold also reduced the likelihood of unlicensed riding. Com-pared with households with three motorcycles or more; none, one and two motorcycles reduced the hazard by 49%, 45.1% and 15.8%, respectively. In addition, students living in Greater Taipei had the unlicensed hazard decreased by 31.4%.
From a regional perspective, district traffic composi-tions, except passenger car ownership rate, also showed a mild association with the unlicensed riding of the district’s students. An increase in one motorcycle owned for 100 res-idents raised the district’s unlicensed hazard by 0.7%, while an increase of one city bus for 10,000 residents reduced the district’s hazard by 2.4%.
3.2. Initial age of unlicensed motorcycling by factors To compare the different initial ages of unlicensed riding experience by groups, we further applied the adjusted sur-vival curve based on the previous Cox estimated results. The adjusted curve means we compared different categories of each variable one at a time while controlling for other variables as mean values. The cumulative probability before specific ages and the median age for unlicensed motorcycling by different groups are displayed inTable 5. The median age, which stands for the age when half of the students had experienced unlicensed motorcycling, for the overall samples was 17.03 years. Stratified by school type, the median value for vocational school students (16.74 years) was 0.64 years earlier than ordinary school students (17.38 years). Male students (16.55 years) started motorcycling 1.43 years earlier than females (17.98 years). Students living with both parents obviously engaged in motorcycle riding later compared with students living with neither parent (17.22 vs. 15.87 years). Students living with a single father had a slightly higher median age than those Table 3
Independent variable notation and expected association with hazard
Variable notation Variable description Expected association
School type 1 if the school type is a vocational senior high school, otherwise 0 +
Sex 1 if the student is male, 0 if female +
Living (1) 1 if the student lives with both parents, otherwise 0 –
Living (2) 1 if the student lives with his/her father only, otherwise 0 –
Living (3) 1 if the student lives with his/her mother only, otherwise 0 –
Parents’ attitude (1) 1 if the parents have a disapproving attitude towards motorcycling, otherwise 0 – Parents’ attitude (2) 1 if the parents have neutral attitude towards motorcycling, otherwise 0 – Greater Taipei 1 if the student lives in Taipei city or country, 0 if other districts – Household motorcycle size (1) 1 if the household of the student has no motorcycle, otherwise 0 – Household motorcycle size (2) 1 if the household of the student has one motorcycle, otherwise 0 – Household motorcycle size (3) 1 if the household of the student has two motorcycles, otherwise 0 – District’s passenger car density Passenger cars owned per hundred residents in the residence district of the student in 2005 – District’s motorcycle density Motorcycles owned per hundred residents in the residence district of the student in 2005 + District’s bus density City buses owned per 10,000 residents in the residence district of the student in 2005 –
living with a single mother (16.27 vs. 16.15 years). The unli-censed median age increased with the parents’ disapproving attitude. Students experienced unlicensed riding 0.71 years
later if their parents had a disapproving attitude compared with the students who gained their parents’ approval. In both households with no motorcycle or only motorcycle, students had a median age greater than the minimum licensing age of 18. However, a household with three motorcycles or more reduced the median age to 16.13 years. Students living in Greater Taipei also showed a delayed age of riding (17.94 years) when compared with other districts (16.92 years).
In addition, students undergoing unlicensed riding revealed different distributions of cumulative probability before specific years of age by groups. The overall proba-bilities of unlicensed experience before the ages of 14, 16, and 18 were 12.5%, 36.2%, and 63.4%, respectively. Similar to previous results, students in a vocational senior high school, male students, students living with both parents, students perceiving a disapproving attitude in parents, stu-dents in households with no motorcycles, and living in Greater Taipei had a lower probability of unlicensed riding compared with their counterpart groups. For example, 73.3% of male students had experienced unlicensed riding before the age of 18, while only 51.4% of females had the same experience; 79.9% of the students living in a house-hold that held more than three motorcycles had unlicensed experience before the age of 18, while only 46.9% of the students in a household owning no motorcycles had this experience.
4. Discussion
The Cox regression results indicated that students affili-ated with a vocational senior high school, male students, and students in districts with a higher motorcycle owner-ship rate had a greater chance of experiencing unlicensed riding and thus had an earlier riding age. In contrast, the beginning age of motorcycling was delayed in students hav-ing a higher degree of monitorhav-ing by their parents (livhav-ing with both parents or perceiving that their parents had a dis-approving attitude), living in households with fewer motor-cycles, living in Greater Taipei, and living in districts with a greater bus density.
The onset of motorcycle riding at an earlier age raises concerns over a disproportionately higher likelihood and severity of an accident produced by these young children. The over-represented risk of accident for the youths has been identified by many studies (Lin et al., 1998; Schoon, 2004; IOT, 2006). The increased crash risk of these young riders may be contributed to by the factors of less experi-ence and immaturity. The youngsters’ immaturity appears to receive more attention than their inexperience. Studies indicated that the immaturity may result from their weaker cognitive ability (Eby and Molnar, 1998) and lead to a stronger propensity towards risky behavior (Zuckerman and Neeb, 1980; Russo et al., 1993; Rutter and Quine, 1996; Mullin et al., 2000; Lin et al., 2003).
The minimum licensing age for motorcycling in most countries is 16 years or lower but Taiwan has set a higher Table 5
Cumulative probability and median age of unlicensed experience by groups
Factor Group 5Age
14 (%) 5Age 16 (%) 5Age 18 (%) Median age (years) Overall – 12.5 36.2 63.4 17.03
School type Vocational senior high 16.4 42.0 69.6 16.74 Ordinary senior high 9.4 31.6 58.1 17.38 Sex Male 14.1 41.2 73.3 16.55 Female 11.0 31.5 51.4 17.98 Living with parents With both parents 11.4 34.5 62.7 17.22 With father only 21.0 45.5 70.1a 16.27 With mother only 16.1 44.1 68.9a 16.15 With neither parent 25.2 53.2 64.1b 15.87
Parents’ attitude Approval 10.7 32.8 57.4 16.75
Neutral 12.1 35.0 61.3 17.10 Disapproval 14.6 40.5 70.7 17.46 Household motorcycle size None 10.8 30.2 46.9 18.32 One 10.0 30.8 47.7 18.26 Two 14.3 39.4 71.3 16.81 Three or more 16.4 47.5 79.9 16.13 Living district Greater Taipei 8.6 30.3 54.6 17.94 Other districts 14.7 39.6 67.9 16.92 Notes: ‘‘a” and ‘‘b” denote that the cumulative probability is less than 17.25 and 17.17 years of age, respectively.
Table 4
Estimated results for the Cox regression model
Independent variable b (S.E.) eb
School type 0.314a (0.073) 1.368 Sex 0.477a (0.074) 1.611 Living (1) 0.477a (0.161) 0.621 Living (2) 0.096 (0.228) 0.908 Living (3) 0.106 (0.198) 0.899 Parents’ attitude (1) 0.313a (0.095) 0.732 Parents’ attitude (2) 0.181a (0.080) 0.835
Household motorcycle size (1) 0.674a (0.161) 0.510
Household motorcycle size (2) 0.600a (0.098) 0.549
Household motorcycle size (3) 0.171b (0.089) 0.842
Greater Taipei 0.377a (0.098) 0.686
District’s passenger car density 0.002 (0.010) 1.002 District’s motorcycle density 0.007b (0.004) 1.007 District’s bus density 0.025b (0.014) 0.976 Number of observations 1594
Censored observations (rate) 762 (47.8%)
LL(b) 5645.6
LL(0) 5755.9
Degrees of freedom 14
Notes: ‘‘a” and ‘‘b” denote that b is significantly different from 0 at the a= 0.05 and a = 0.1 levels, respectively.
licensing age of 18, aiming to postpone the time that ado-lescents ride a motorcycle. This delaying strategy appears ineffective because unlicensed riding experience is prevalent and the possibility of starting motorcycling is significant from about 14 years of age. The high prevalence and low age of unlicensed behavior in Taiwan may have caused many legal or safety problems resulting from the immatu-rity and inexperience of young people. In addition, mopeds with no speed limitation, combined with the immaturity of the young riders in Taiwan, may raise the likelihood and severity of an accident.
The prevalence of authorized riding appears hard to suppress if the mobility needs, especially for their weekend activities, cannot be satisfied by other suitable alternative modes of transport (Yeh et al., 2007). However, a certain number of accidents may be controlled by the overall reduction in riding exposure if the current licensing age is maintained (IOT, 2006). Therefore, whether the current minimum age for motorcycle riding should be lowered depends on the establishment of effectively matched countermeasures.
Some measures combined with a lower licensing age in some advanced countries may be considered to control the risk of teen motorcycling accidents. Specifically, either speed-limited mopeds (i.e., mopeds having speeds lower than 45 kph or light-mopeds lower than 25 kph) (Schoon, 2004) or a lower risk riding situation such as mandatory training and instruction, no carriage of passengers, a maxi-mum speed restriction, and a stricter BAC based on a grad-uated licensing system is requested to regulate the behavior of young novice riders (Haworth and Mulvihill, 2005).
In addition, as indicated by Hartos et al. (2000), low parental monitoring and control were associated with higher risk behavior, violations and crashes among teenag-ers driving an automobile. Promoting parental manage-ment of teenage driving at the stage of the learner’s permit or provisional license has been suggested because of the significant limits on teenage driving in the first few months (Simons-Morton et al., 2002, 2003). Our results also showed that a higher degree of parental monitoring delays the unlicensed age of motorcycling. Therefore, much attention should be paid to countermeasures of parental monitoring, power- or speed-limited mopeds, training pro-grams, and no carriage of passengers if the government
considers lowering the minimum licensing age of
motorcycling. References
Chang, H.L. 1996. A survey on the transportation demand and traffic risk perception by adolescent students in Taiwan. The Culture and Education Foundation of Tong-Yang Wu-Suon, Taipei, Taiwan. Chang, H.L., Yeh, T.H., 2007. Motorcyclist accident involvement by age,
gender, and risky behaviors in Taipei, Taiwan. Transportation Research Part F: Traffic Psychology and Behaviour 10, 109–122. Chesham, D.J., Rutter, D.R., Quine, L., 1993. Motorcycling safety
research: A review of the social and behavioral literature. Social Science and Medicine 37 (August), 419–429.
Eby, D.W., Molnar, L.J. 1998. Matching Traffic Safety Strategies to Youth Characteristics: A Literature Review of Cognitive Develop-ment. Commissioned by US Department of Transportation National Highway Traffic Safety Administration.
Harre´, N., Field, J., Kirkwood, B., 1996. Gender differences and areas of common concern in the driving behaviors and attitudes of adolescents. Journal of Safety Research 27, 163–173.
Hartos, J.L., Eitel, P., Haynie, D.L., Simons-Morton, B.G., 2000. Can I take the car? Relations among parenting practices and adolescent problem-driving practices. Journal of Adolescent Research 15, 352– 367.
Haworth, N., Mulvihill, C., 2005. Review of Motorcycle Licensing and Training. Monash University Accident Research Center, Victoria, Australia.
Institute of Transportation. 2006. A Feasibility Study on Lowering Motorcycle Licensing Age. Taipei, Taiwan.
Kleinbaum, D.G., 1996. Survival Analysis: A Self-learning Text. Springer Press, New York, USA.
Lam, L.T., 2003. A neglected risky behavior among children and adolescents: Underage driving and injury in New South Wales, Australia. Journal of Safety Research 34, 315–320.
Lin, M.R., Chang, S.H., Pai, L., Keyl, P.M., 2003. A longitudinal study of risk factors for motorcycle crashes among junior college students in Taiwan. Accident Analysis and Prevention 35, 243–252.
Lin, D.Y., Lin, F.F., Chang, K.K. 1998. In: Preliminary Discussion on Adolescent Accident Analysis and Safety Education, Traffic Safety, vol. II. Taipei, Taiwan.
Mannering, F.L., Grodsky, L.L., 1995. Statistical analysis of motorcyclists’ perceived accident risk. Accident Analysis and Prevention 27, 21–31. Ministry of Transportation and Communications. 2002. Taiwan’s Road
Traffic Regulations. Taipei, Taiwan.
Ministry of Transportation and Communications. 2004. Statistical Abstract of Transportation and Communications. Taipei, Taiwan. Mullin, B., Jackson, R., Langley, J., Norton, R., 2000. Increasing age and
experience: are both protective against motorcycle injury? A case control study. Injury Prevention 6, 32–35.
Plase`ncia, A., Borrell, C., Anto´, J.M., 1995. Emergency department and hospital admissions and deaths from traffic injuries in Barcelona, Spain: A one-year population-based study. Accident Analysis and Prevention 27, 591–600.
Reeder, A.I., Chalmers, D.J., Langley, J.D., 1995. Young on-road motorcyclists in New Zealand: age of licensure, unlicensed riding, and motorcycle borrowing. Injury Prevention 1, 103–108.
Russo, M.F., Stokes, G.S., Lakey, B.B., Christ, M.A.G., McBurnett, K., Loeber, R., Stouthhamer-Loeber, M., Green, S., 1993. A sensation seeking scale in children: further refinement and psychometric devel-opment. Journal of Psychopathology and Behavioral Assessment 15, 69–86.
Rutter, D.R., Quine, L., 1996. Age and experience in motorcycling safety. Accident Analysis and Prevention 28, 15–21.
Schoon, C. 2004. Traffic legislation and safety in Europe concerning the moped and the A1-category (125 cc) motorcycle. Literature and Questionnaire Study Commissioned by the Swedish National Road Administration. SWOV Institute for Road Safety Research, Leids-chendam, The Netherlands.
Simons-Morton, B.G., Hartos, J.L., Beck, K.H., 2003. Persistence of effects of a brief intervention on parental restrictions of teen driving privileges. Injury Prevention 9, 142–146.
Simons-Morton, B.G., Hartos, J.L., Leaf, W.A., 2002. Promoting parental management of teen driving. Injury Prevention 8 (Suppl. II), ii24–ii31.
Yeh, T.H., Chang, H.L., Chang, H.W. Initial age of unlicensed motor-cycling experience for a cohort of high school students. Accident Analysis and Prevention. doi:10.1016/j.aap.2007.08.005 (Available online 4 September 2007).
Zuckerman, M., Neeb, M., 1980. Demographic influences in sensation seeking and expressions of sensation seeking in religion, smoking, and driving habits. Personality and Individual Differences 1, 197–206.
