Risk Factors associated with Stress-related Riding Behavior among Indonesian Motorcyclists

The risk of road accidents is influenced by various elements, particularly human factor. In Indonesia, motorcycles contribute significantly to accidents due to insufficient knowledge of traffic laws and a lack of awareness about safe riding practices. Stress-related riding behavior, which may lead to this incident, is closely linked to driving discipline. Therefore, this study aimed to examine stress-related behavior as the risk factor in road accidents among motorcyclists. To achieve this, a self-reported questionnaire was distributed to 300 motorcyclists in Banda Aceh City and surrounding areas in Aceh Province. Factor analysis was used to simplify the number of indicators included, after which they were prioritized in order of importance. The results showed the most influential indicators out of the 26 considered in this study. A total of 3 underlying factor were identified, comprising anxious riding, aggressive riding, and anger. By addressing the risk factor associated with stress-related riding behavior, road safety in Indonesia was enhanced.


Introduction
According to the World Health Organization, 2018 [1], the number of deaths associated with road traffic is rising globally, reaching 1.35 million in 2016.This makes road accident the 8th leading cause of mortality across all age groups, particularly children and young adults between 5 and 29 years.Over the last 15 years, the rate of road traffic deaths has remained relatively constant at around 18 per 100,000 people.Some progress has been made, as evidenced by a decrease from 135 per 100,000 vehicles in 2000 to approximately 64 in 2016.According to ITF, 2021 [2], data from 34 member countries of the IRTAD (International Traffic Safety Data and Analysis) Group, showed that the Covid-19 pandemic affected mobility patterns and the number of road fatalities in 2020.The number of road deaths decreased by 8.6% across all modes of transportation.However, the significant decrease in 2020 fell short of the target of 50% worldwide reduction in fatalities by 2020.This was set by the United Nations Decade of Action for Road Safety.Most of the data from member countries were obtained from Europe, Australia, and Japan.Meanwhile, information from motorcycle-dependent countries such as Indonesia remains scarce.
As reported by WHO (2018), the updated data from Indonesia showed that motorcycle users accounted for up to 74% of all road fatalities in the country.The excessive use of this transportation means, is most likely attributed to the high cost of a car, as well as the inadequacy of public transportation services and facilities [3].Understanding factor contributing to violations and traffic accidents is critical in formulating policies to implement road safety on urban roads.Despite the influence of factor such as vehicles and roads [4], the human factor remains the most dominant [5].Tasca (2002) defined aggressive riding behavior as an intentional act that increases the risk of traffic accidents.This behavior is motivated by impatience, annoyance, hostility, and/or an attempt to save time [16].Furthermore, it represents a dysfunctional pattern of social behavior that endangers public safety.The identification of specific aggressive driving behaviors includes tailgating, weaving in and out of traffic, improper passing (such as cutting in too close in front of the overtaken vehicle), passing on the road shoulder, making improper lane changes (failure to signal), failing to yield the right of way to other road users, preventing other drivers from passing, displaying unwillingness to extend cooperation to fellow motorists, struggling to merge or change lanes due to traffic conditions, and driving at speeds significantly exceeding the norm.This behavior often leads to frequent tailgating, abrupt lane changes, running stop signs, and running red lights.
Abojaradeh et al. (2014) reported that riding behavior is a significant factor in many traffic accidents, injuries, and fatalities [17].Stress is a significant factor influencing riding behavior and is closely linked to individual characteristics.Given that riding is a complex task demanding concentration, individuals under this condition encounter difficulties in making sound decisions and are more prone to succumb to road rage.Furthermore, the condition is associated with an increased likelihood of engagement in serious accidents [18].
Thwe et al., (2017) investigated the relationship between riding stress, riding behavior, and accident in Myanmar [19].The survey comprised 46 professional motorcyclists from certain companies in the country.In addition, it included both experimental and questionnaire surveys.The devices used were a heart rate sensor, a watch, and a video camera.Each participant attached a heart rate sensor around his/her chest, and a watch provided respiratory rate recording via its connection to the sensor.Valid data were subjected to heart rate variability (HRV) analysis.To represent stress levels, the difference between the average baseline condition and the most stressful segments was analyzed.The questionnaire survey included demographic information, unsafe riding behavior, the 26-item Driver Behavior Questionnaire (DBQ), and the 43-item riding behavior related to stress.The results showed that stress, as well as demographic factor, influence riding behavior and contribute to traffic accidents.
According to WHO (2018), citing data from the Indonesian National Police in 2016, a substantial 74% of the fatalities among road users in the country were comprised of two-wheeled and three-wheeled motorcyclists.Similar to other countries with a high population of motorcyclist, there are issues related to behavior on the road.Some studies have used the DBQ instrument to measure motorcyclemotorcyclists behavior.Suwarto et al., (2019) conducted a report on the fear of novice motorcyclists in Semarang City using DBQ.Based to the results, fear does not necessarily reduce lapses and errors while riding, but it decreases the intention of young motorcyclists to commit ordinary and aggressive violations [20].Lady et al. (2020) used the DBQ instrument to investigate the effect of age, riding experience, and accident rate on motorcyclists riding behavior.The study showed a significant disparity in violations committed by motorcyclists under the age of 35 in Cilegon City when compared to their older counterparts [21].Putranto and Alyandi (2019) explored deeper into this issue by utilizing the Indonesian Motorcycle Rider Behavior Questionnaire (IMRBQ), a tool derived from the DBQ.The results presented significant correlations between certain constructs in Indonesian Family Values, the Indonesian Driver Behavior Questionnaire (IDBQ), and IMBRQ.Consequently, it was concluded that the family plays a pivotal role in instilling essential values that contribute to better riding and motorcycle conduct on the road [22].In the context of Indonesia, report focusing on stress-related riding behavior has been limited.This study aims to investigate the risk factors influencing road accidents, with a specific emphasis on stress-related riding behavior of motorcyclists in Banda Aceh City, Indonesia.To assess this behavior, some indicators will be drawn from the study of Thwe et.al (2017).

Study Area
The investigation was conducted in Aceh Province, located in the northwestern region of Indonesia.Banda Aceh City, the provincial capital, was selected as the primary study area, with an inclusion of the surrounding areas, comprising Aceh Besar Regency.Based on Central Bureau Statistics data in 2022, the total population of Banda Aceh City and Aceh Besar Regency, were 257,635 and 414,490, respectively.The regions composed a total area of 55.85 km² and 2,882.83km², with average population densities of 4,612.98 people/km² and 143.78 people/km², respectively.

Respondent, Samples, and Demographic
Multivariate analysis necessitates a sample size ranging from 100 to 200 [23], which should ideally be 5-10 times the number of indicators [24], or at least comprise10-15 participants per indicator (Field, 2009).In this particular study, the latter criterion was met, ensuring a robust foundation.With 26 indicators multiplied by 10 participants, the total amounted to 260, which was reasonably approximated to 300 samples.This quantity of samples proved to be suitable for conducting factor analysis [25].Comrey and Lee (1992) further categorized factor analysis samples into three groups, namely 100 (indicating a poor sample size), 300 (indicating a good sample size), and 1000 (indicating an excellent sample size) [26].

Questionnaire Design
A self-reported questionnaire survey was conducted in the year 2021 among motorcyclists residing in Banda Aceh City and Aceh Besar Regency, Aceh Province, Indonesia.The questionnaire was designed to identify the risk factor associated with stress-related riding behavior, which may have contributed to a road accident.In Bahasa Indonesia, it was distributed to motorcyclists using a 4-point Likert scale, ranging from Never to Always [27].To prevent respondents from becoming confused about the direction of their responses [28], this scale was deliberately chosen without a central tendency measure.
To collect data on riding behavior, the DBQ was adopted based on previous studies conducted in Indonesia and other countries ([17]; [29]; [30]; [31]; and [19]).For this report, 26 indicators were developed based on scenarios and situations commonly encountered by motorcyclists in the country, with the majority being adapted from Thwe et al., 2017 [19].A five-point Likert scale (1=never, 2=rare, 3=sometimes, 4=often, and 5=always) was used by Thwe et al. (2017), while this study used a fourpoint Likert scale (1=never, 2=rare, 3=seldom, and 4=nearly all the time).Riding behavior of 42 professional motorcyclists in Myanmar was examined, focusing on 43 items of stress-related riding behavior.Given that this investigation centered on motorcyclists riding behavior, only a few indicators were applied, and others were selected based on their correlation with motorcycles and motorcyclists.Table 2 presents full details of the 26 indicators, along with the mean and standard deviation.It is important to note that these indicators exhibit some similarities, necessitating a new grouping based on exploratory factor analysis (EFA).

Exploratory Factor Analysis (EFA)
According to [23], EFA is a multivariate analytical method used to examine the underlying patterns or relationships among a large number of variables.It aims to determine whether the information can be condensed or summarized in a smaller set of factor or components.Hair (2019) stated that the main goal of factor analytic methods is to discover a means of condensing the amount of information included in a large number of original variables into a smaller group of new, composite dimensions (factor) with minimal information loss.This can be achieved through either an exploratory or a confirmatory factor analysis.Based to [24], those underlying variables or dimensions were known as factor (or latent variables).
In EFA process, each variable or indicator will be examined for feasibility, to determine whether it is possible to be incorporated [24].EFA method can be performed when the KMO MSA value is greater than 0.5 [32].Meanwhile, according to Field (2009), the variable of a value less than 0.5 should be disregarded.After the screening test process, the core process of EFA includes the extraction of a set of existing variables (indicators) to form one or more factor (factor also termed as dimensions, components, or latent variables).The most widely used extraction method was Principal component analysis (PCA).
PCA was used when the objective was to summarize most of the original information (variance) in a minimum number of factor for prediction purposes.In contrast, common factor analysis was applied primarily to identify underlying factor or dimensions that reflect what the variables share in common.The most direct comparison between the 2 methods is by their use of the explained versus unexplained variance.

Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO-MSA), Bartlett's Test of Sphericity, and Extraction Method of Principal Component Analysis (PCA)
The first stage of EFA is the screening of several variables to obtain those that meet the criteria for analysis.Table 3 shows the results of the KMO-MSA and Bartlett's test of sphericity as a suitability measure for EFA.KMO-MSA was 0.963 (> 0.5), and Bartlett's test of sphericity χ 2 (325) = 5759.859(p < 0.05).This indicated that the correlations between items were sufficiently large, adequate, and suitable for EFA.0.000

Factor Loadings
A Principal Component Analysis (PCA) was performed on the 26 items using orthogonal rotation (varimax).Initially, an analysis was conducted to acquire eigenvalues for each component in the dataset.Out of these, 3 components exhibited values greater than 1, collectively accounting for 50.982% of the variance, as presented in Table 3.According to the suggestions of Field (2009), eigenvalues serve as an approach to determining the optimal number of factors for inclusion in the analysis.Another method was the scree plot, which was also considered.Field (2009) contends that eigenvalues associated with a variate signify its substantive significance.Since not all factors are retained in the analysis, and there exists a debate over the criterion for determining the statistical importance of a factor, it is prudent to retain only those with substantial eigenvalues.Table 4 presents the factor loadings post-rotation alongside communalities, and the values were >0.5.Factor 1 comprises 10 indicators, making a substantial contribution to stress-related riding behavior factors.It accounts for 50.982% of the total explained variance, with an eigenvalue of 13.255.This highlights its significant association with stress-inducing riding situations, as evidenced by 80% of the indicators displaying factor loadings above 0.7.Specifically, these significant factor loadings were worrying when riding on a road with many junctions (0.785), worrying when riding at night (0.780), worrying when riding on a road with recklessly parked vehicle (0.778), worrying when riding in traffic jam (0.758), worrying when riding in bad weather (0.755), worrying when riding on hilly roads (0.753), worrying when riding in a road that I've never been to (0.749), and worrying when riding on roads with poor geometry and asphalt surface (0.728).Other lower significant indicators are worrying when riding in the wrong lane (0.697), and worrying when riding with fragile goods (0.588).These collectively showed an elevated level of apprehension among motorcyclists when dealing with specific environmental and traffic conditions, signifying a prevailing sense of anxious riding within this group.
Factor 2 comprised 11 items, explaining 9.196% of the variance, with an eigenvalue of 2.391.All items in this factor are strongly associated with aggressive riding, showing factor loadings ranging from 0.474 to 0.793.The most significant indicators, with factor loadings exceeding 0.7, include riding fast until a sudden lane change is required (0.793), riding aggressively when in a bad mood (0.757), and becoming agitated with other motorists during a traffic jam and subsequently riding aggressively (0.749).Other indicators exhibit lower significance for factor two, such as reacting to other imprudent actions of motorcyclists (0.680), engaging in disputes with fellow road users (0.667), and failing to overtake other vehicles (0.639), forcing to ride without enough rest (0.634), when riding a worn-out motorcycle (0.629), when traffic light turns red (0.580), when there is a demanding task (0.550), and when riding with high status people (0.474).
Factor 3, which we named "anger," comprised 5 items closely associated with the angry condition.This factor had the smallest impact on stress-related riding behavior factor, accounting for only 4.199% of the variance with an eigenvalue of 1.092.In comparison to the indicators listed in the previous two factors, the factor loadings of each, showed a lower contribution which is less than 0.70.These indicators included feeling angry when riding behind a slow-moving vehicle (0.683), expressing anger through flashing lights or horns (0.666), experiencing frustration in traffic jams without a U-turn option (0.663), impatience during rush hour (0.554), and proceeding through a yellow-to-red traffic signal (0.547).In a parallel study, Liew et al. (2022) [33] conducted a comprehensive examination of anger among Malaysian motorcyclists.Similar to the method used in this study, EFA was adopted to generate 8 factors consisting of 59 indicators.The most significant factor related to anger was labeled 'unsafe or inappropriate actions,' comprising 14 associated items such as challenges in changing lanes and an inclination to accelerate while overtaking.Overall, all 26 indicators used in our study exhibited factor loadings exceeding 0.4, affirming their significance and inclusion, with none necessitating removal.

Conclusions
In conclusion, this study collected data from 300 motorcyclists in Banda Aceh City and Aceh Besar Regency, Aceh Province, Indonesia, in order to identify the risk factor linked to stress-related riding behavior.The questionnaire included 26 indicators to describe the conditions of stress-related riding behavior of motorcyclists.By using the Exploratory Factor Analysis (EFA), the data was classified into three factor: anxious riding, aggressive riding, and anger, with eigenvalues surpassing 1.0 and a total variance of 64.378%.The communalities value for the 26 indicators exceeded the defined threshold of 0.30, ranging from 0.454 to 0.774.Factor loadings for all indicators ranged from 0.474 to 0.793.The Cronbach's alpha values for all three factors exceeded 0.60, affirming robust reliability and internal consistency.Significantly, anxious riding stood out as the primary determinant of stress-related riding behavior among motorcyclists in Aceh Province.This study presents initial findings regarding factors and indicators that potentially contribute to stress-related riding behavior within the Indonesian context.Future study may explore the intricate relationships between these indicators and accident experiences.This will be a central focus of upcoming investigations.

2
According to[6], it accounts for approximately 80 and 95% of accidents[7]  and accident-related factor, respectively.Vivo et al.(2006)  emphasized that human factors, including the psychological condition of motorcyclists, contribute to 60-80 percent of traffic accidents[8].This point out the importance of a comprehensive understanding of human aspects for enhancing riding safety.Studies have shown that young motorcyclists often overestimate their riding abilities, showing higher confidence levels while perceiving lower accident risks [9] [10].According to the report by Ulleberg and Rundmo (2003), a motorcyclist's attitude toward traffic safety can mediate the relationship between personality characteristics and riding behavior [11].Furthermore, studies from [12] and [13] motorcyclists negligence and lack of discipline are the leading causes of traffic accidents.The Traffic Regulations Discipline, as stipulated in Law Number 22 of 2009 regarding Road Traffic and Transportation [14], concerning the conduct of motorcyclists on the road.It emphasizes adherence to established laws and traffic regulations for both motorized and non-motorized users.This discipline is directly related to aggressive riding tendencies that often result in violation of traffic laws [15].

Table 1 .
Descriptive statistics of respondents.

Table 2 .
Mean and standard deviation of each indicator.