Motorcycle-Riding Experience: Friend or Foe? Understanding Its Effects on Driving Behavior and Accident Risk

Abstract

For those who cannot afford vehicles, motorcycles are a common mode of transportation in Pakistan. Although many motorcycle riders also drive vehicles, they continue to engage in dangerous behaviors such as speeding, weaving through traffic, and disobeying traffic laws, which can result in accidents. The purpose of this study was to investigate the relationship between prior motorcycle-riding experience, dangerous driving behaviors, and traffic accidents. A total of 623 drivers in Pakistan participated in a study in which questions on their demographics, involvement in accidents, and dangerous driving practices were posed. Two statistical models were employed to analyze the data and ascertain how motorcycle-riding experience affected dangerous driving behaviors and accidents. Drivers with past motorcycle-riding experience were found to be more likely to exhibit aggressive and risky driving behaviors, as validated by principal component analysis. Certain demographic characteristics were also linked to dangerous driving behaviors, and prior motorcycle experience was identified as a crucial factor in anticipating traffic collisions. The probability of a traffic accident increased by 67% for each unit rise in motorcycle-riding experience. To lower the incidence of accidents, the study suggests that the government and road safety regulatory authorities impose strict rules and regulations for motorcycle riders.

1. Introduction

There is widespread agreement that free and secure movement is a basic human right. As urbanization and globalization continue to increase, the provision of safe transportation within road traffic has become an urgent global concern. Traffic collisions, which are recognized as the most frequent cause of injury, impairment, and death among individuals, can result from a failure to maintain safe road mobility. Road traffic accidents can also exert psychological and monetary consequences in the lives of those involved. As reported by the World Health Organization (WHO), road traffic accidents (RTAs) are, at present, the eighth leading cause of death for all age categories and the main factor behind the deaths of people aged 5–29 years globally [1]. The RTA burden is disproportionately borne by low- and middle-income countries. Low- and middle-income countries experience major difficulties with congestion and mobility, which leads to more road traffic accidents that ultimately result in injuries and fatalities [2,3,4]. In total, 1.35 million people died in traffic accidents in 2016, as reported by the WHO [1], and despite accounting for only 60% of the world’s registered vehicles, low- and middle-income countries account for 90% of all road traffic deaths [5,6]. The Global Status Report on Road Safety adds that Pakistan has a significantly higher rate of fatalities from RTAs than the United Kingdom (UK) and Germany. In Pakistan, there are 14.2 fatalities for every 100,000 people, compared to 3.1 in the UK and 4.1 in Germany [1]. This suggests that, compared to the UK and Germany, Pakistan has a much higher probability of RTA-related fatalities despite being less motorized. It is estimated by studies from around the world that approximately 70% of accidents are due to human factors [7]. Driver behavior (what a driver chooses to do) is of more concern as compared to other human factors in road safety studies. Driving behaviors other than safe driving behaviors are variable and lack consensus. Driving acts that depart from safe driving procedures and have the potential to hurt or endanger other drivers, other road users, or property are referred to as dangerous driving behaviors.

The phrase “dangerous driving” refers to a wide range of behaviors, including aggressive driving, risky driving, and driving while feeling upset (negative emotions). On a global level, this behavior significantly contributes to traffic accidents [8,9]. Running red lights, yelling at others, gesticulating crudely, weaving in and out of traffic, forcing other cars off the road, and exceeding the speed limit are all dangerous driving practices that can endanger not only the driver but also others on the road. These actions are frequently the outcome of aggression and anger, which might result in harm to oneself or other people involved in the incident. Three dimensions of dangerous driving have been delineated: aggressive driving, risky driving, and negative cognitive/emotional driving [10]. Studies also highlight aggression with intent to harm (behaviors and cognitive or emotional states that make the driving situation more dangerous), negative emotions (frustration, anger, and rumination), as well as risky driving behaviors (lacking actual intent to harm) [10,11,12]. Numerous studies exist in the literature in which the authors have studied dangerous driving behaviors in various forms—for example, aggressive driving behaviors [9,13,14,15,16,17,18], risky driving behaviors [8,19,20,21,22], and negative emotional driving [23,24,25,26]. The Dula Dangerous Driving Index (DDDI), devised by Dula and Ballard in 2003 [10], combines aggressive, dangerous, and negative emotional driving behaviors, as compared to other measures, such as the Driving Anger Scale (DAS) [27], the Propensity for Angry Driving Scale (PADS) [28], the Driving Anger Expression Inventory (DAX) [14], and the Driver’s Angry Thought Questionnaire (DATQ) [29], which measure anger only. Numerous studies have examined the relationship between socioeconomic and demographic factors, dangerous driving habits, and the possibility of being involved in traffic accidents [8,10,11,14]. There is strong evidence from countries all over the world that the DDDI is a well-known and thoroughly researched tool for the evaluation of dangerous driving behaviors. The literature further suggests that the translated versions of the DDDI also show good internal consistency, e.g., the Chinese [12], United States [10], French [30], and Romanian [31] versions support the three-factor structure, while the Belgian and American [11] versions support a four-factor structure. Only a single research study has looked into the viability of the DDDI in Asian countries, i.e., Qu et al. [12]. Apart from this, no such study has been conducted in any Asian nation on dangerous driving behaviors using the DDDI. The bulk of research on dangerous driving behaviors, according to a thorough study of the literature, appears to have been conducted in advanced economies [10,11,30,31]. Moreover, due to the strong safety culture and efficient law enforcement, the frequency of injuries and fatalities has reduced over time. However, the frequency of traffic accidents that result in fatalities is exceedingly high and alarming in developing nations [32]. It is difficult to relate research findings from advanced economies to developing economies due to the significant disparities in their characteristics, such as traffic circumstances, driving patterns, road conditions, legal enforcement methods, socioeconomic levels, and road user attitudes.

In recent years, Pakistan, along with several other emerging economies, has experienced extraordinary motorization and urbanization. The situation has been exacerbated by unplanned road infrastructure development and a lack of government attention. Except for two studies that have been conducted on aberrant driving behaviors [33,34], Pakistan has little information available regarding how drivers behave on roads. The first study [33] focused mostly on cases of noncompliance with traffic rules (violations). According to the study’s findings, drivers in Pakistan frequently engage in aggressive, lawless, risk-taking, and selfish driving behaviors. Meanwhile, the second [34] focused on errors and violations (distracted and risky). According to this study’s findings, the main causes of aberrant driving behaviors in Pakistan include a lack of sufficient driving training and driving without a license. Apart from these two studies, which focused on aberrant driving behaviors, a research study has been conducted in Pakistan to examine the effect of alcohol and marijuana on crash risk perception among commercial drivers [35]. Furthermore, running red lights, yelling at others, gesticulating crudely, weaving in and out of traffic, forcing other cars off the road, tailgating, and exceeding the speed limit are all dangerous driving practices that can endanger not only the driver but also others on the road; these are all common among Pakistani drivers. In addition, non-adherence to seatbelt and helmet usage laws is a very common aspect found among Pakistani drivers.

In Pakistan, motorcycles are a common form of transportation. Pakistan has a 43% usage rate, which places it seventh globally after Thailand (87%), Vietnam (86%), Indonesia (85%), Malaysia (83%), China (60%), and India (47%) [36]. In Pakistan, the majority of road traffic accidents (RTAs) are caused by motorcycle riders. Between January and March 2022, there was a significant increase in the number of motorcycle-related traffic accidents throughout all 36 districts of Punjab province. The total number of such accidents during this time period was 64,964, implying that the province saw an average of 725 such accidents every day [37]. In the cities of Rawalpindi and Islamabad, there were over 27,000 reported accidents during the years of 2017 and 2019, and all of them involved motorcycles [38]. Motorcycle riders in Pakistan commonly engage in dangerous driving behaviors due to a lack of understanding of safety precautions and a system of laws that lacks effective enforcement. On most highways in Pakistan’s urban areas, there are no specific lanes for motorcycles. As a result, motorcycle riders frequently exhibit dangerous behaviors, such as weaving in and out of traffic, switching lanes unnecessarily, exceeding speed limits, following other vehicles too closely, making improper U-turns, passing other vehicles from the inside lane, honking excessively, and stopping or exiting from any point on the road. As previously stated, the dangerous driving behaviors revealed by motorcycle riders have certain parallels with those shown by drivers of other vehicles. Motorcycle riders frequently develop dangerous driving habits, and when they move on to driving motor cars, they frequently still display these tendencies. As this creates a significant risk to both them and other road users, it is critical to address the issue of motorcycle riders transferring dangerous driving behaviors to the driving of motor cars. This is especially important in countries where motorcycle use is widespread, and it is important to investigate how motorcycle-riding experience affects these dangerous behaviors and how they affect accidents involving motor vehicle drivers.

Many researchers have looked into the connection between driving experience, dangerous driving behaviors, and traffic accidents in the literature. The factors that influence driving behaviors include the road conditions, weather conditions, driving time (day vs. night), vehicle condition, driver’s health, and road infrastructure, as reported by [39,40]. Drivers in general tend to drive more cautiously in wet conditions, i.e., heavy rainfall conditions, as compared to dry weather conditions [41]. Driving speeds also affect driving behavior: the higher the speed, the higher the probability of being involved in traffic accidents, because a driver has less time to respond to a safety-critical situation when traveling at higher speeds [42]. In regard to roadway features, it has been observed that on horizontal curves, drivers typically slow down as they approach the curve and progressively speed up as they approach the curve’s exit. Additionally, it was shown that as the horizontal curve radius increases, so does the speed of the vehicle [43]. A research study on the effect of motorcycle-riding experience on traffic accidents and aberrant driving behaviors was performed in Pakistan [5]. The study found no evidence to correlate motorcycle-riding experience with traffic accidents, but it did find that motorcycle-riding experience is a strong predictor of distraction-related driving offenses. This is the only known study on the subject, and it focuses on aberrant driving behaviors. There is presently no research that directly examines how a person’s motorcycle-riding history affects their propensity to participate in dangerous driving practices or their probability of being involved in an accident while operating a vehicle. It is critical to examine the dangerous driving behaviors used by Pakistani drivers in order to understand the underlying causes and create efficient strategies to encourage safer driving habits. In order to achieve the above goal, a self-report questionnaire survey was carried out in Pakistan. The purpose of the survey was to evaluate the participants’ socioeconomic status, experience with motorcycles, and involvement in dangerous driving behaviors. The Dula Dangerous Driving Index (DDDI), which consists of 27 items, was used to gauge the participants’ dangerous driving tendencies. To identify the factor structure of dangerous driving behaviors among Pakistani drivers, principal component analysis with varimax rotation was used. Using a generalized linear model (GLM) and a binary logistic regression model, the impact of motorcycle-riding experience on dangerous driving habits and crashes (RTAs) among Pakistani drivers was also investigated.

2. Materials and Methods

2.1. Participants

On a voluntary and anonymous basis, 796 Pakistani drivers answered a self-report questionnaire either online (N: 481) or through in-person interviews (N: 315). Following the completion of the necessary data screening and filtration steps, 732 replies (432 online responses and 300 responses from in-person interviews) were judged to be legitimate and selected for further investigation. Respondents (N: 109) who exclusively rode motorcycles and were unable to drive other vehicles were eliminated to ensure accuracy, resulting in a final sample size of 623 for this study. Out of 623 valid samples, 373 samples were from online data collection and 250 samples were from physical data collection. The sample was intended to include people of all ages who were licensed to drive both public and private vehicles, regardless of gender. Additionally, effort was made to include drivers of all educational levels. The study included people aged 18 to 65, with an average age of 2.41 and a standard deviation of 1.44. The sample included 506 men (81.2%) and 117 women (18.8%). Moreover, 75% of the participants had a valid driver’s license. Half of the participants were either undergraduate or graduate students or employees at various universities, while the other half were drivers recruited from diverse sites such as bus stops, markets, and residential neighborhoods.

Table 1. Total sample characteristics.

Variable	Category	Combined Sample (N: 623)	Physical Data Collection (N: 250)	Online Data Collection (N: 373)
		Percentage (%)	Percentage (%)	Percentage (%)
Gender	Male	81.2	79.2	82.6
	Female	18.8	20.8	17.4
Age group	18–24 years	37.9	30	43.2
	25–34 years	20.9	19.2	22
	35–44 years	17.2	22	13.9
	45–54 years	12.5	14.4	11.3
	55–64 years	9.6	12.8	7.5
	>65 years	2	1.6	2.1
Education level	PhD	5.6	6	5.4
	Masters	15.4	18	13.7
	Bachelors	29.1	23.2	33
	Higher Secondary School Certification (HSSC)	29.9	25.6	32.7
	Secondary School Certification (SSC)	9.1	9.1	5.9
	Below SSC	8.6	25	5.1
	No education	3.4	2	4.3
Motorcycle-riding experience	No experience	23.9	17.6	11
	1–4 years	30.7	31.2	26.3
	5–8 years	20.7	18.8	24.1
	9–12 years	9.3	10	11.8
	13–16 years	8.8	9.2	10.2
	17–20 years	4.7	9.6	9.9
	>20 years	2.6	3.6	6.7

summarizes the characteristics of the whole sample and the characteristics of the data collected through online and in-person interviews.

2.2. Materials

2.2.1. Dula Dangerous Driving Index (DDDI)

The Dula Dangerous Driving Index (DDDI), a self-report measure developed by Dula and Ballard [10] to determine the personal tendency toward dangerous driving, was used in this study. The original scale consists of 28 items and three components: risky driving (RD;12 items), negative emotions while driving (NE; 9 items), and aggressive driving (AD; 7 items). As alcohol consumption is banned in Pakistan, one item related to drunk driving was removed and a 27-item DDDI was used in this research. The respondents rated the frequency of each item on a five-point Likert scale ranging from 1 (“never”) to 5 (“always”).

Because many drivers have a low educational level and cannot read English, the Pakistani Urdu version of the DDDI was employed in this study. We adopted a unique procedure to translate the original English version. First, four language department professors individually translated the DDDI into Urdu. They then discussed their translations in order to develop a final text. Second, five experienced drivers reviewed the draft to ensure that the items were clear. Finally, we completed the scale after obtaining feedback from a group of five potential participants (drivers) who pretested the translated text.

2.2.2. Sociodemographic Characteristics

The sociodemographic characteristics included in the questionnaire were age, gender, education, and driving experience. The participants were also asked to indicate whether they had a valid driving license, whether they used a seatbelt or helmet while driving or riding a motorcycle, from whom they learned to drive, and whether they had been involved in traffic accidents in the past. The participants were prompted to share details about their motorcycle-riding history as part of the study. They were specifically asked to state how many years of experience they had, with response possibilities ranging from none to more than 20 years, and with each option being given a number between 1 and 7.

2.3. Method

Data collection was performed using an internet-based survey and physical data collection. The online survey questionnaire was developed using Google Forms, and the survey links were published on social media channels. Meanwhile, for physical data collection, in-person interviews were carried out. For in-person interviews, data were collected from drivers available at various locations, such as bus stations, shopping mall parking lots, restaurants, housing complexes, and other commercial areas. The willingness of the participants to participate was confirmed before data collection and they were also briefed about the purpose of the data collection and study. The participants were all Pakistani drivers from various professions. Both strategies used an anonymous and self-report questionnaire. The participants were informed that their comments would be used only for research reasons and would not be made public. SPSS version 26 was used to perform the necessary data screening and statistical analysis.

2.4. Data Analysis

To establish whether the data were appropriate for principal component analysis (PCA), two parameters were evaluated. The first was the fraction of variation among the variables that may represent common variance, as determined by the Kaiser–Meyer–Olkin (KMO) measure of sample adequacy. The second test, known as Bartlett’s Test of Sphericity (BTS), examined the null hypothesis that the variables were uncorrelated in the sample. The sample’s KMO ratio was 0.869, indicating that the sample size was adequate for analysis. Furthermore, the BTS for the data was determined to be statistically significant (p < 0.000), indicating that the data satisfied the requirements for principal component analysis. As a result, principal component analysis with varimax rotation was performed on the data. The Kaiser rule [44], which entails maintaining factors with eigenvalues greater than one, and the scree cutoff points [45] technique are two often used rules for the calculation of the number of latent variables or factors. Based on the eigenvalue criterion, a five-factor solution was initially achieved in the analysis, but the scree cutoff points approach indicated only three factors. The three-factor solution obtained from the scree cutoff points was judged to be appropriate after several iterations. Furthermore, all extraction communalities were constrained to be over 0.4 [46]. The Cronbach’s alpha coefficient was used to evaluate the internal validity and reliability of the identified factors [47]. This statistical metric allows for an assessment of the internal consistency and dependability of the acquired elements.

Furthermore, the validity (degree to which the outcomes accurately reflect the variables that they are designed to capture) of the DDDI scale was determined by calculating its correlation with other variables (or measures of a construct) with which it should be positively, negatively, or not at all associated. Spearman’s rank correlation analysis was performed to check the validity of the DDDI by examining the relationships between motorcycle-riding experience, sociodemographic characteristics, and the dangerous driving behavior dimensions obtained as a result of PCA. In the next step, generalized linear models (GLM) were incorporated to examine the effect of motorcycle-riding experience and other study variables on dangerous driving behaviors. Generalized linear models (GLM) are a versatile and extensively used modeling technique that may be applied to a wide range of modeling scenarios involving response variables that do not have a normal distribution or a link between their mean and variance. GLM can be used in a variety of contexts, including the analysis of continuous variables whose variance is dependent on the mean and even categorical data. Aggressive driving, dangerous driving, and negative emotional driving were incorporated as dependent factors (DVs) in the GLM model, whereas sociodemographic characteristics and motorcycle-riding experience were introduced as independent factors. (IVs). Lastly, the binary logistic regression approach was used to assess the impact of dangerous driving and motorcycle-riding experience on traffic accidents.

3. Results

3.1. DDDI Dimensions

The principal component analysis (PCA) results showed that there were three separate groups or categories of the DDDI that provided a clear and complete overview of a multitude of dangerous driving behaviors displayed by Pakistani drivers.

Table 2. DDDI dimensions and Cronbach’s alpha coefficients.

Factors	Item No.	Factor 1	Factor 2	Factor 3
Negative emotions while driving (NE) (Cronbach’s α: 0.689)	11. When I get stuck in a traffic jam, I get very irritated.	0.754
	12. I get impatient and/or upset when I fall behind schedule when I am driving.	0.723
	16. I get irritated when a car/truck in front of me slows down for no reason.	0.721
	10. I consider the actions of other drivers to be inappropriate or “stupid”.	0.538
	15. I feel that I may lose my temper if I have to confront another driver.	0.510
Aggressive driving (AD) (Cronbach’s α: 0.716)	5. I would tailgate a driver who annoys me.		0.727
	4. I deliberately use my car/truck to block drivers who tailgate me.		0.694
	6. When someone cuts me off, I feel I should punish him/her.		0.641
	2. I make rude gestures (e.g., giving “the finger,” yelling curse words).		0.602
	2. I feel it is my right to strike back in some way, if I feel another driver has been aggressive toward me.		0.547
	3. I verbally insult drivers who annoy me.		0.409
Risky driving (RD) (Cronbach α: 0.702)	26. I consider myself to be a risk-taker.			0.712
	21. I will drive if I am only mildly intoxicated or buzzed.			0.728
	19. I will race a slow-moving train to a railroad crossing.			0.714
	27. I feel that most traffic “laws” could be considered as suggestions.			0.694
Extraction method: principal component analysis. Rotation method: varimax with Kaiser normalization. Rotation converged in 4 iterations.

displays the list of items that were extracted during the PCA analysis. Each component was labeled in accordance with the most prominent contributing factors to this behavior. The first extracted component was identified as “negative emotional driving”, since it had five items that corresponded to negative emotional driving and accounted for 24.40% of the variance. The items “When I get stuck in a traffic jam, I get very irritated”, “I get impatient and/or upset when I fall behind schedule when I am driving”, and “I get irritated when a car/truck in front of me slows down for no reason” for this component had higher factor loadings and represented the negative traits of Pakistani drivers in such traffic situations. The second element was entitled “aggressive driving”, since it was dominated by aggressive driving tendencies. It consisted of six elements (for example, “I would tailgate a driver who annoys me”, “I deliberately use my car/truck to block drivers who tailgate me”, and “when someone cuts me off, I feel I should punish him/her”.) and accounted for 15.21% of the variation. The third and last extracted factor was named “risky driving” because it contained items that were dominated by risky driving behaviors. This factor consisted of four items representing the risk-taking nature of Pakistani drivers and accounted for 8.94% of the variance. The items in this factor with high factor loadings were “I consider myself to be a risk-taker” and “I will race a slow-moving train to a railroad crossing”.

3.2. Internal Consistency

Cronbach’s alpha coefficients were used to analyze the consistency of the DDDI subscales. The findings revealed that all three subscales and the total DDDI score had high internal consistency. The alpha value of the entire DDDI score was 0.771, which was regarded as good [47]. Similarly, the alpha values for each of the three subscales (negative feelings while driving, risky driving, and aggressive driving) ranged from 0.689 to 0.716, respectively. See Table 2 for details.

3.3. Correlation Analysis

The findings of the Spearman’s bivariate correlation test revealed that aggressive driving (Spearman’s rho, r: 0.724, p > 0.01), risky driving (Spearman’s rho, r: 0.837, p > 0.01), and negative emotional driving (Spearman’s rho, r: 0.717, p > 0.01) had a positive association with the overall DDDI score. A decrease in aggressive (Spearman’s rho, r: −0.083, p > 0.05), risky (Spearman’s rho, r: −0.246, p > 0.01), and negative emotional (Spearman’s rho, r: −0.286, p > 0.01) driving behaviors was observed with an increase in age. Females were reported to be more prone to risky driving behaviors (Spearman’s rho, r: −0.084, p > 0.05). Driving experience did not show any association with dangerous driving behaviors, whereas motorcycle-riding experience showed a significant association with all dimensions of the DDDI, indicating that with an increase in motorcycle-riding experience, dangerous driving behaviors also increased. This study discovered no instances of multiple correlations greater than 0.7, which could have impacted the accuracy of the regression results. As a result, it is possible to conclude that there were no concerns with collinearity in the study’s sample. See

Table 3. Correlation analysis of DDDI with variables in study.

Variable	1	2	3	4	5	6	7	8
1. Aggressive driving	-
2. Risky driving	0.540 **	-
3. Negative emotions	0.409 **	0.534 **	-
4. DDDI score	0.724 **	0.837 **	0.717 **	-
5. Age	−0.136 **	−0.284 **	−0.290 **	−0.256 **	-
6. Gender	−0.073	−0.084 *	−0.019	−0.049	−0.046	-
7. Experience	0.045 *	0.013 *	0.018	0.021	0.477 **	−0.043	-
8. Motorcycle-riding experience	0.628 **	0.328 **	0.191 **	0.428 **	0.037 *	−0.076	−0.021	-

* p < 0.05 (1-tailed) and ** p < 0.01 (2-tailed).

for details.

3.4. Factors Affecting Dangerous Driving Behaviors of Pakistani Drivers

A generalized linear model (GLM) was used to assess and quantify the influence of factors such as motorcycle-riding experience and demographic features on the occurrence of dangerous driving behaviors. A GLM applies a well-known technique to a wide range of response modeling situations, such as when the response variables are not normally distributed or when “variance is a function of the mean” for numerous continuous variables. GLM approaches can also be used with categorical data. A GLM has two components: the random component of the probability function, which reflects the variance in the values of the response variable, and the structural component of the probability function, which connects the mean of the response variable to the predictor values [48]. The three extracted dimensions of the DDDI, i.e., aggressive driving, risky driving, and negative emotional driving, were forecasted by demographic variables and the motorcycle-riding experience variable. Three GLM models were employed to identify the factors responsible for predicting the dangerous driving behaviors of Pakistani drivers.

In the first GLM model, the aggressive driving dimension of the DDDI was entered as a DV and the sociodemographic variables and motorcycle-riding experience were entered as the IVs in the model.

Table 4. GLM model 1 results (dependent variable: aggressive driving behavior).

Variable	B	SE	Wald	p (Sig)	OR	95% CI
						Lower	Upper
Aggressive Driving (AD)	Chi-squared: 53.94, p < 0.01
Age (18–24)	0.236	0.1077	4.812	0.028	1.266	1.025	1.564
Age (25–34)	0.221	0.1083	4.174	0.041	1.248	1.009	1.543
Age (over 65)	0 a
Gender	Not Significant
Education	Not Significant
Driving Training (from Traffic Police School)	−0.282	0.0849	11.018	0.001	0.754	0.639	0.891
Driving Training (from Private driving School)	−0.253	0.0742	11.592	0.001	0.777	0.672	0.898
Driving Training (from Friend)	−0.188	0.0738	6.460	0.011	0.829	0.717	0.958
Driving Training (from Relative/Family Member)	−0.229	0.0694	10.897	0.001	0.795	0.694	0.911
Driving Training (No Training)	0 a
Traveling Time	Not Significant
Motorcycle Riding Exp. (No Experience)	−0.473	0.0900	27.585	0.000	0.623	0.523	0.744
Motorcycle Riding Exp. (1–4) Years	−0.449	0.0890	25.441	0.000	0.638	0.536	0.760
Motorcycle Riding Exp. (5–8) Years	−0.255	0.0912	7.784	0.005	0.775	0.648	0.927
Motorcycle Riding Exp. (9–12) Years	0.261	0.0969	7.268	0.007	1.298	1.074	1.570
Motorcycle Riding Exp. (13–16) Years	0.286	0.0970	8.718	0.003	1.332	1.101	1.611
Motorcycle Riding Exp. (17–20) Years	0.288	0.1058	7.393	0.007	1.333	1.084	1.641
Motorcycle Riding Exp. (>20) Years	0 a

Likelihood Chi-squared: 417.245, df: 26, significance: p < 0.01, Akaike’s Information Criterion (AIC): 1506.278, Bayesian Information Criterion (BIC): 1630.446, B: Coefficient, SE: Standard Error, OR: Odds Ratio, CI: Confidence Interval. ^a This parameter was set to zero because it was redundant.

presents the results of the GLM model for aggressive driving behaviors. The results revealed that motorcycle-riding experience significantly affected the aggressive driving behaviors of vehicle drivers. Drivers having no motorcycle-riding experience (β: −0.473, odds ratio: 0.623, p < 0.01) were less involved in aggressive driving behaviors as compared to drivers with more than 20 years of motorcycle-riding experience. Drivers having motorcycle-riding experience of 16 to 20 years (β: 0.288, odds ratio: 1.333, p < 0.01) were more involved in aggressive driving behaviors as compared to drivers with more than 20 years’ experience. On comparing the results of drivers with little or no prior motorcycle-riding experience to those with more experience, it was discovered that drivers with no prior motorcycle-riding experience were less likely to engage in aggressive driving than those with more experience. Drivers aged 18 to 24 years (β: 0.236, odds ratio: 1.266, p < 0.05) were more involved in aggressive driving behaviors as compared to drivers over 65 years of age. Upon analyzing the results, it was revealed that with an increase in the drivers’ age, a reduction in aggressive driving was also observed, but the values did not reach significant levels. It was also observed that drivers who had received training in any form before driving were less prone to aggressive driving behaviors as compared to drivers with no prior training, e.g., driving training from a traffic police driving school (β: −0.282, odds ratio: 0.754, p < 0.01), training from a private driving school (β: −0.253, odds ratio: 0.777, p < 0.01), and training from a friend (β: −0.229, odds ratio: 0.795, p < 0.01). The gender, education, and travel time variables did not show any significant associations with aggressive driving behaviors.

In the second GLM model, the risky driving dimension of the DDDI was entered as a DV and the sociodemographic variables and motorcycle-riding experience were entered as the IVs in the model.

Table 5. GLM model 2 results (dependent variable: risky driving behaviors).

Variable	B	SE	Wald	p (Sig)	OR	95% CI
						Lower	Upper
Risky Driving (RD)	Chi-squared: 28.26, p < 0.01
Age (18–24)	0.467	0.1272	13.479	0.000	1.595	1.243	2.047
Age (25–34)	0.328	0.1280	6.553	0.010	1.388	1.080	1.783
Age (35–44)	0.347	0.1299	7.152	0.007	1.415	1.097	1.826
Age (Over 65)	0 a
Gender (Male)	0.101	0.0449	5.027	0.025	1.106	1.013	1.208
Gender (Female)	0 a
Education	Not Significant
Driving Training (from Relative/Family Member)	−0.198	0.0818	5.853	0.016	0.820	0.699	0.963
Driving Training (No Training)	0 a
Traveling Time (6 a.m.–12 p.m.)	−0.152	0.0717	4.471	0.034	0.859	0.747	0.989
Traveling Time (12 a.m.–6 a.m.)	0 a
Motorcycle Riding Exp. (No Experience)	−0.395	0.1058	13.965	0.000	0.673	0.547	0.829
Motorcycle Riding Exp. (1–4) Years	−0.371	0.1048	12.520	0.000	0.690	0.562	0.848
Motorcycle Riding Exp. (>20) Years	0 a

Likelihood Chi-squared: 189.63, df: 23, significance: p < 0.01; Akaike’s Information Criterion (AIC): 1756.074; Bayesian Information Criterion (BIC): 1866.94; B: Coefficient, SE: Standard Error, OR: Odds Ratio, CI: Confidence Interval. ^a This parameter was set to zero because it was redundant.

presents the results of the GLM model for risky driving behaviors. The results revealed that motorcycle-riding experience significantly affected the risky driving behaviors of vehicle drivers. Drivers having no motorcycle-riding experience (β: −0.395, odds ratio: 0.673, p < 0.01) were less involved in risky driving behaviors as compared to drivers with more than 20 years of motorcycle-riding experience. Drivers having motorcycle riding experience of 1 to 4 years (β: −0.371, odds ratio: 0.690, p < 0.01) were also less involved in risky driving behaviors as compared to drivers with more than 20 years’ experience. On comparing the results of drivers with little or no prior motorcycle-riding experience to those with more experience, it was discovered that drivers with no prior motorcycle-riding experience were less likely to engage in risky driving behaviors than those with more experience, but this did not reach a significant level. It was also observed that drivers who had received training from a family member or relative before driving (β: −0.198, odds ratio: 0.820, p < 0.05) were less prone to risky driving behaviors as compared to drivers with no prior training. Among the demographic variables, drivers aged 18 to 24 years (β: 0.467, odds ratio: 1.595, p < 0.01) were more involved in risky driving behaviors as compared drivers over 65 years of age. Drivers aged 25 to 34 years (β: 0.328, odds ratio: 1.388, p < 0.05) and 35 to 44 years (β: 0.347, odds ratio: 1.415, p < 0.01) also appeared to engage in more risky driving behaviors in comparison with motorists over the age of 65. Upon analyzing the results, it was revealed that with an increase in drivers’ age, a reduction in risky driving was also observed, but the values did not reach significant levels. Male drivers (β: 0.101, odds ratio: 0.1.106, p < 0.05) were more prone to risky driving behaviors as compared to female drivers in the sample. Moreover, it was observed that drivers who traveled from 6 a.m. to 12 p.m. (β: −0.152, odds ratio: 0.859, p < 0.05) were less involved in risky driving behaviors as compared to those who traveled from 12 a.m. to 6 a.m. in the morning.

The last GLM model predicted the factors that were responsible for negative emotional driving among Pakistani drivers. Similar to the first two models, the negative emotional driving dimension of the DDDI was the DV and all other study variables were the IVs in the model. The results revealed that drivers who had motorcycle-riding experience of 16–20 years (β: 0.314, odds ratio: 1.369, p < 0.05), followed by drivers with motorcycle-riding experience of 9 to 12 years (β: 0.270, odds ratio: 1.310, p < 0.05), were more prone to negative emotional driving as compared to drivers with more than 20 years’ experience. Driving training did not show any significant associations with negative emotional driving. Among the demographics, only age had a significant association with negative emotional driving, while gender and traveling time did not reach significant levels. It was observed that drivers aged 18 to 24 years (β: 0.303, odds ratio: 1.354, p < 0.05) were more prone to negative emotional driving, followed by drivers aged 35 to 44 years (β: 0.273, odds ratio: 1.314, p < 0.05), as compared to drivers aged over 65 years. With an increase in drivers’ age, it was revealed that negative emotional driving behaviors were reduced, i.e., drivers aged 55–64 years (β: −0.282, odds ratio: 0.755, p < 0.05) were reported to be less prone to negative emotional driving as compared to younger drivers. See

Table 6. GLM model 3 results (dependent variable: negative emotional driving behaviors).

Variable	B	SE	Wald	p (Sig)	OR	95% CI
						Lower	Upper
Negative Emotions (NE)	Chi-squared: 7.721, p < 0.01
Age (18–24)	0.303	0.1330	5.196	0.023	1.354	1.043	1.757
Age (35–44)	0.273	0.1354	4.064	0.044	1.266	0.974	1.646
Age (55–64)	−0.282	0.1377	4.185	0.041	0.755	0.576	0.988
Age (Over 65)	0 a
Gender	Not Significant
Education	Not Significant
Driving Training	Not Significant
Traveling Time	Not Significant
Motorcycle Riding Exp. (9–12) Years	0.270	0.1188	5.163	0.023	1.310	1.038	1.653
Motorcycle Riding Exp. (17–20) Years	0.314	0.1307	5.777	0.016	1.369	1.060	1.769
Motorcycle Riding Exp. (>20) Years	0 a

Likelihood Chi-squared: 159.66, df: 23, significance: p < 0.01; Akaike’s Information Criterion (AIC): 1762.342; Bayesian Information Criterion (BIC): 1873.206; B: Coefficient, SE: Standard Error, OR: Odds Ratio, CI: Confidence Interval. ^a This parameter was set to zero because it was redundant.

for detailed results.

3.5. Effect of DDDI Dimensions and Motorcycle-Riding Experience on Road Traffic Accidents (RTA)

The aim of this section was to investigate how various factors, such as dangerous driving behaviors, motorcycle-riding experience, and demographic information, influence the occurrence of road traffic accidents. To do this, the sample drivers were divided into two groups: those who had been involved in road traffic accidents (RTAs) and those who had not. The data were then analyzed using a binary logistic regression model, with the presence of RTAs coded as “0” for no and “1” for yes. As independent variables (IVs), the model employed dangerous driving behaviors. In the first stage of the binary logistic model, dangerous driving behaviors, i.e., aggressive, risky, and negative emotional driving, were introduced as IVs to evaluate their impacts on road traffic accidents. Aggressive driving behaviors (β: 0.816, Exp(β): 2.262, p < 0.01) and risky driving behaviors (β: 0.321, Exp(β): 1.379, p < 0.01) were significant predictors of RTAs, while negative emotional driving did not reach a significant level. See

Table 7. Binary logistic model stage 1 results (dependent variable: traffic accidents; independent variable: DDDI dimensions).

Variable	β	SE	Wald	Sign (p)	Exp. (β)
Accident Involvement (criterion)	0.216	0.081	7.178	0.007	1.241
Risky Driving (RD)	0.321	0.108	8.862	0.003	1.397
Negative Emotions (NE)	0.171	0.105	2.667	0.102	1.187
Aggressive Driving (AD)	0.816	0.106	59.053	0.000	2.262

Nagelkerke R²: 0.317, model chi-squared: 168.34, sig: p < 0.01, model percentage correctness: 76%.

for the stage 1 results.

In the second stage, motorcycle-riding experience and other sociodemographic factors were introduced as IVs along with the dangerous driving behaviors to evaluate their effects on RTAs. See

Table 8. Binary logistic model stage 2 results (dependent variable: traffic accidents; independent variable: DDDI dimensions, motorcycle riding experience, and demographics).

Variable	β	SE	Wald	Sign (p)	Exp. (β)
Accident Involvement (criterion)	0.216	0.081	7.178	0.007	1.241
Risky Driving (RD)	0.473	0.127	13.943	0.000	1.604
Negative Emotions (NE)	0.441	0.124	12.707	0.000	1.554
Aggressive Driving (AD)	0.464	0.141	10.797	0.001	1.591
Motorcycle-Riding Experience	0.667	0.105	40.393	0.000	1.948
Age	0.545	0.096	32.040	0.000	1.725
Driving Experience	−0.255	0.090	8.012	0.005	0.775
Gender	Not Significant

Nagelkerke R²: 0.474, model chi-squared: 272.112, sig: p < 0.01, model percentage correctness: 78.0%.

for detailed results. All three dimensions of dangerous driving behaviors had a significant association of p < 0.01, which showed that all three dangerous driving behaviors were responsible for the occurrence of RTAs. Motorcycle-riding experience had a significant association with RTAs, i.e., one unit increase in the motorcycle-riding experience of drivers increased the probability of occurrence of RTAs by 67%. Similarly, drivers’ age and driving experience also showed significant associations with RTAs, i.e., the likelihood of RTAs rises by 54% with every single unit increment in driver age, but it falls by 25% with each unit increment in driving experience. Figure 1 presents a summary of the data analysis.

4. Discussion

The primary goal of this study was to determine whether drivers with motorcycle experience in Pakistan display more dangerous driving behaviors, which could potentially contribute to an increase in road traffic accidents. An increase in road traffic accidents involving drivers who had previously ridden motorcycles was the main reason that this study was performed. The study proposes that dangerous driving behaviors common among motorcycle users, such as speeding, evading traffic signals, using the opposite lanes, risky overtaking, and using medians, may explain the heightened dangerous driving behaviors seen in drivers with prior motorcycle-riding experience. According to the study, drivers who had prior experience in riding motorcycles were more inclined to participate in dangerous driving behaviors and cause traffic accidents. Many studies exist in the literature that have studied the effect of driving experience on dangerous driving behaviors, such as [31,49]. As far as we know, no previous research has looked into how motorcycle-riding experience affects the possibility of vehicle drivers engaging in dangerous driving behaviors and being involved in accidents. This is the first study that probes into this relationship.

The Dula Dangerous Driving Index (DDDI) was used in this study to assess the dangerous driving behaviors of Pakistani drivers. A three-factor solution of dangerous driving behaviors, i.e., aggressive driving, risky driving, and negative emotional driving, was obtained in this research, which is in accordance with previous studies [10,31]. However, this study is not in line with [11,12], which support a four-factor solution of dangerous driving behaviors. Based on the study of the three-factor solution, there is good evidence to justify the usage of the DDDI as a measure of dangerous driving behaviors in Pakistan, because the scale’s three variables can explain a considerable percentage of the variation. This is the first study to utilize the DDDI to explore the dangerous driving behaviors of Pakistani drivers. The three extracted factors and the overall DDDI score have good internal consistency and reliability, with alpha coefficient values greater than 0.7 for each factor. The DDDI dimensions can therefore be used in similar driving situations and adequately depict dangerous driving behaviors in Pakistan.

One of the objectives of this study was to identify the variables responsible for dangerous driving behaviors using motorcycle-riding experience and other sociodemographic variables. Motorcycle-riding experience was identified to be a reliable indicator of dangerous driving behaviors. Motorcycle-riding experience had a strong association with aggressive driving behaviors, i.e., with an increase in the motorcycle-riding experience of drivers, their engagement in aggressive driving behaviors was also increased. Minimal or no experience in riding motorcycles was found to be a negative predictor of risky driving behaviors. With an increase in motorcycle-riding experience, the risky driving behaviors of motor car drivers also increased, but this did not reach a significant level. Drivers with no experience in riding motorcycles were also found to be involved in negative emotional driving in comparison to motorists with more motorcycle experience. The possible reasons for such a trend may be as follows: similar to riding a motorcycle, some motorcycle riders could display over-confidence in their driving skills when switching to driving a car. This may encourage them to take risks or engage in risky activities that they perceive to be manageable. Additionally, car driving may be monotonous to motorcycle riders, who are accustomed to being outside and enjoying the open road. This boredom may cause distractions and a lack of concentration while driving, both of which are risky. Some riders of motorcycles might not be accustomed to a car’s size and weight, and they might not be aware of blind spots or other hazards that are specific to operating a larger vehicle. Moreover, it can be a major change to switch from riding a motorcycle to operating a car, and some motorcycle riders may experience worry or anxiety as a result. This nervousness may result in errors or risky actions while driving.

The previous literature suggests that sociodemographic variables can describe dangerous driving behaviors [11,31]. Among the demographic variables, age was found to be a significant predictor of dangerous driving behaviors (aggressive, risky, and negative emotional driving), and a greater number of teenage motorists engage in aggressive, risky, and negative emotional driving behaviors. However, as the age increases, the involvement in dangerous driving behaviors is reduced. The results of this study are in line with previous work [10], in which age was also found to be a key indicator of dangerous driving behaviors. The reason might be that there were more young drivers (58%) within our sample than older ones and that, as people become older, their propensity for risk taking and aggression declines. Gender was also a key indicator of risky driving behaviors. Male drivers are more likely to take risks as compared to females. This result is consistent with earlier research works [10,31]. The reason for such results could also be that, in our data set, female driver representation was much lower (19%) as compared to males (81%). This can also be explained by the fact that the number of female drivers in the country is very low due to cultural obligations [50]. Drivers who receive training before driving engaged significantly less in dangerous driving behaviors, suggesting that trained drivers are less likely to display aggressive and risky driving as compared to untrained drivers, while no association with negative emotional driving was observed. This could be explained as follows: drivers who received training from any source (driving schools or friends or family members) were less involved in aggressive driving behaviors, while drivers who received training from a relative or family member were less involved in risky driving behaviors as compared to untrained drivers. The results of this study are at odds with those of a prior investigation carried out in Pakistan [34], in which driving training was not a significant predictor of aberrant driving behaviors. The other reason for such a trend could be that trained drivers are more aware of the traffic rules and regulations and how to behave in different traffic situations, so they are less prone to dangerous driving behaviors as compared to untrained drivers.

This study sought to evaluate the association between motorcycle-riding experience, dangerous driving behaviors, and the occurrence of traffic accidents among respondents from Pakistan. RTAs are predicted by dangerous driving behaviors, according to a substantial number of investigations that have been published [12,16,17,18]. All three dangerous driving behaviors were found to be significant predictors of traffic accidents. Aggressive driving and risky driving behaviors were the strongest predictors of road traffic accidents. Numerous studies have repeatedly indicated that aggressive driving and risky driving behaviors are a strong indicator of being involved in a road traffic accident. As a result, the findings of this study are congruent with the current body of research on the topic [12,16,17,18,51,52]. Driving influenced by negative emotions, i.e., negative emotional driving, also significantly predicted road traffic accidents. This finding is also in line with previous studies [12,53]. Another reason for such results is the unequal representation of drivers of all age groups in the sample. Younger drivers tend to drive more aggressively, take more risks, and are influenced by their emotions, so the probability of involvement in accidents also increases as compared to older drivers. The motorcycle-riding experience variable was identified as a critical indicator of road traffic accidents. Individuals with past motorcycle-riding experience have a much higher risk of being involved in road traffic accidents (RTAs), according to the present study, which is a novel conclusion because earlier studies on dangerous driving behaviors did not take this element into consideration. Car driving is a very different experience from riding a motorcycle. Some motorcycle riders may find it difficult to adjust to driving a car and may lack the experience necessary to properly negotiate traffic and, as a result, may be involved in accidents.

The current study has a few limitations. The first is that, although it is a cost-effective strategy, relying on drivers’ self-reports to identify accidents and unsafe driving behaviors may be inappropriate. The outcomes of self-report research entirely depend on how accurately and truthfully drivers respond to the questionnaire. In order to enhance the self-reporting and move beyond the limits of our study, we advise the use of more cutting-edge methodologies such as driving simulators or naturalistic driving tactics. These methods may be able to address the biases present in self-reports and offer more unbiased and trustworthy information on driving behavior. In addition, 81% of respondents in our sample of 623 participants were men, which means that more women must participate in future studies in order to generalize the findings to the entire population. The DDDI measure was studied here for the first time in Pakistan, and further research is therefore encouraged to verify it in our country, with a particular focus on drivers of commercial vehicles.

5. Conclusions

The likelihood of RTA occurrence increases by 67% for every unit increase in a driver’s motorcycle experience, demonstrating the strong correlation between motorcycle-riding experience and RTAs. The likelihood of RTAs increases by 54% with every single unit escalation in driver age, while it decreases by 25% with each unit increment in driving experience.

This study suggests looking further into the association between motorcycle experience and a person’s propensity for dangerous driving behavior. In order to be consistent with the norms of developed nations, the fines and punishments for breaking traffic regulations while operating a motorcycle or other motor vehicle should also be reviewed. The establishment of strict rules and regulations for motorcycle riders in accordance with international standards, such as the system being followed in the UK [54], by the government and road safety regulatory authorities is important to reduce the number of road traffic accidents (RTAs) in Pakistan. An internet database should be built that is available to all traffic police departments across the country, not only in specific cities, so that violators’ traffic offence records can be updated wherever they commit a violation. The implementation of penalties and an electronic ticketing system throughout the country, as well as stringent monitoring of traffic offenders using a cutting-edge traffic monitoring technology, is necessary. Moreover, educating road users about traffic regulations and road safety through driver training programs and road safety campaigns would be beneficial. This will guarantee that road safety is maintained when drivers start operating motor vehicles. This study has highlighted the link between dangerous motorcycle-riding behaviors and dangerous driving behaviors among vehicle drivers. We strongly recommend that more research be undertaken in the future to further explore this topic.