safety

Article
Nonlinear Analysis of the Effects of Socioeconomic, Demographic,
and Technological Factors on the Number of Fatal Traffic Accidents
Nassim Sohaee 1, *,† and Shahram Bohluli 2,†

1 Information Technology and Decision Science, University of North Texas, Denton, TX 76203, USA
2 Gradient Systematics LLC, Dallas, TX 75202, USA; [email protected]
* Correspondence: [email protected]
† These authors contributed equally to this work.

Abstract: This study explores the complex connections among various socioeconomic, demographic,
and technological aspects and their impact on fatal traffic accidents. Utilizing the Lasso polynomial
regression model, this study explores the impact of demographic variables, including income, ed-
ucation, unemployment rates, and family size. Additionally, socioeconomic factors such as Gross
Domestic Product (GDP) per capita, inflation rate, minimum wage, and government spending on
transportation and infrastructure are examined for their impact on the occurrence of fatal accidents.
This study also investigates the influence of technological advances in vehicles on the outcomes
of traffic safety. The findings of this research reveal that certain factors, such as average, alcohol
consumption, unemployment rate, minimum wage, and vehicle miles traveled (VMT), among others,
have a substantial impact on the multifactorial model and play a considerable role in the frequency of
fatal accident rates. The research results have significant implications for policymakers, highlighting
the need for a comprehensive approach that accounts for the interdependence of economic indicators,
behavioral patterns, and traffic safety outcomes. This study underscores the importance of consid-
ering a wide range of socioeconomic, demographic, and technological factors to develop effective
policies and strategies to reduce fatal traffic accidents.

Keywords: fatal accidents; Lasso; polynomial regression; socioeconomic; automation

Citation: Sohaee, N.; Bohluli, S.

Nonlinear Analysis of the Effects of
Socioeconomic, Demographic, and 1. Introduction
Technological Factors on the Number
Fatal traffic accidents are a major and widespread problem in the United States and
of Fatal Traffic Accidents. Safety 2024,
10, 11. https://doi.org/10.3390/
worldwide. In the U.S. alone, thousands of individuals lose their lives each year due
safety10010011
to these incidents, while road accidents are considered the ninth leading cause of death
worldwide. According to the World Health Organization, approximately 1.3 million deaths
Academic Editor: Tom Brijs occur annually from road accidents, causing around 50 million injuries across different
Received: 13 October 2023 countries [1]. However, initiatives such as the United Nations Global Plan aim to reduce
Revised: 6 December 2023 road fatalities and injuries by half by 2030, giving optimism for improvement. [2].
Accepted: 5 January 2024 To effectively prevent fatal accidents, it is crucial to develop strategies grounded in a
Published: 10 January 2024 thorough understanding of the various contributing factors. These encompass a range of
elements from human behavior and vehicle malfunctions to environmental conditions and
broader macrosocioeconomic influences [3]. Recent extensive research has demonstrated
a mutual coordination between road transportation and economics [4], highlighting how
Copyright: © 2024 by the authors. economic status can influence regional disparities in crash occurrences [5]. This underscores
Licensee MDPI, Basel, Switzerland. the importance of considering a diverse array of factors in devising comprehensive accident
This article is an open access article
prevention strategies.
distributed under the terms and
Understanding the impact of socioeconomic factors on fatal collisions requires a clear
conditions of the Creative Commons
definition and comprehension of these factors themselves. These factors include broader
Attribution (CC BY) license (https://
social, economic, and demographic conditions such as Gross Domestic Product (GDP),
creativecommons.org/licenses/by/
inflation rate, income levels, educational attainment, poverty rates, employment rates,
4.0/).

Safety 2024, 10, 11. https://doi.org/10.3390/safety10010011 https://www.mdpi.com/journal/safety

Safety 2024, 10, 11 2 of 18

average cost per mile traveled, daily vehicle miles traveled by drivers, and the extent
of urban development in a particular area. Researchers have extensively studied these
factors [5–7]. Previous studies have typically focused on individual contributing factors or
a limited combination. However, it is crucial to analyze the combined influence of various
macrosocioeconomic factors on the occurrence of deadly accidents.
This research aims to unravel the combined impact of socioeconomic, demographic,
and technological factors that, through a complex interplay, contribute to the overall
number of traffic-related fatalities. In addition to examining individual factors, this study
seeks to shed light on the intricate relationships between these variables and how they can
affect the total number of deaths resulting from traffic incidents. It is crucial to understand
that this research is exploratory in nature, rather than confirmatory. This means that the
primary objective is not to establish causal relationships, but rather to observe patterns and
correlations that have already been identified and substantiated in the field. It is important
to note that this approach is useful for generating hypotheses and identifying potential
areas for further research. In addition, the investigation of such a complex structure could
be based on a myriad of other influencing factors. In this study, our focus is on a country like
the United States, which presents a unique context with its specific socioeconomic dynamics,
demographic diversity, and technological advances. By concentrating on this particular
setting, we aim to provide insights that are relevant and adaptable to similar contexts,
while acknowledging that findings may vary in different national or regional scenarios.
After conducting a thorough review of the existing literature, this study identifies
three primary categories of factors that contribute to fatal accident rates: demographic,
socioeconomic, and technological. In the following sections, each of these categories will
be discussed in detail, drawing from relevant research and integrating the findings into a
cohesive analytical framework.
This article is structured in a methodical manner with subsequent sections to ensure
clarity and continuity. Section 2 will explain the methodologies used to acquire and prepare
the datasets for this study. Following that, Section 3 will describe the analytical approach
adopted, specifically the Lasso polynomial regression model, and how it will be applied to
the data. Finally, Section 4 will present the findings of the model, highlighting new insights
and contributing to the ongoing discussion within the research community on traffic safety
and its underlying factors. Through this structured discourse, this study aims to improve
the understanding of traffic fatalities and provide a foundation for future research in this
critical domain.

1.1. Impact of Demographic Factors

In the collection of factors that contribute to traffic safety, demographic characteristics
provide the foundational context to understand patterns and trends in road use and its
associated risks. Average family size, higher educational attainment, and population
collectively paint a picture of the social makeup that is intrinsic to the way individuals
interact with transportation systems. Demographic elements such as average family size,
higher educational attainment, and population serve as the underlying strata upon which
socioeconomic influences and technological advancements exert their effects, making them
indispensable for a holistic analysis of traffic fatality risks.
High population density and urbanization levels generally lead to more traffic fatalities
due to increased traffic density and complex road networks and age distribution within
the population [8,9]. The age distribution of drivers is important for road safety, as both
younger and older drivers pose specific risks. [10].
Research shows that young drivers possess an exaggerated sense of their driving
ability, which apparently leads to a lower appreciation of the level of danger associated with
different driving actions and therefore to greater risk-taking behaviors while driving [11,12].
Larger family sizes may influence vehicle occupancy and travel frequency, while higher
levels of educational attainment often correlate with greater awareness of safety protocols
and potentially influence commute patterns [13,14].
Safety 2024, 10, 11 3 of 18

1.2. Impact of Socioeconomic Factors

Socioeconomic factors refer to the social and economic aspects of a community or
society, which are often related to economic activity and policy. It is important to note
that while some factors are clearly socioeconomic or demographic in nature, others can
potentially intersect both categories. For example, higher educational attainment can have
socioeconomic implications, influencing income levels and employment opportunities.
Therefore, the classification provided here is based on the primary association of each factor.
Socioeconomic factors play a crucial role in shaping transportation behaviors and
influencing the risk of traffic-related fatalities. These factors include a wide range of
economic activities, policies, and societal behaviors. For example, the number of registered
cars in a society is indicative of its mobility needs and economic capabilities. The number
of registered vehicles has increased over time, but the impact on the number of accidents is
not clear. Some studies suggest that accidents cannot influence vehicle registration, while
others indicate that accidents can cause changes in vehicle registration [15].
The GDP is a measure of the economic health of a community. GDP per capita is a
crucial indicator of a country’s economic output and living standards. Research suggests
that an increase in GDP per capita is associated with a decrease in traffic accidents due
to better infrastructure, improved road safety measures, and increased awareness of and
education on traffic regulations [16–18]. Furthermore, the inflation rate has a significant
impact on the occurrence of fatal crashes; higher inflation rates may result in higher vehicle
maintenance costs, leading to delayed or inadequate vehicle upkeep [19,20].
The economic stability of individuals is directly affected by the annual average un-
employment rate, the average household income, and the minimum wage. This, in turn,
affects their travel habits and vehicle maintenance practices. The role of the minimum
wage in traffic safety is complex, but critical. Lower minimum wages are often associated
with higher poverty rates and income inequality, which, in turn, correlate with inadequate
vehicle maintenance and limited access to safer transportation options. However, this
relationship is not linear. Studies show that an increase of 10% in the minimum wage could
correspond to an increase of 7% to 11% in traffic accidents [21]. Various studies suggest
that higher income and education levels are associated with a lower probability of deadly
accidents [22,23] and that poverty rates are positively correlated with the occurrence of
fatal accidents [24].
Investing in transportation and infrastructure development is crucial to ensure the
safety and efficiency of road networks. Countries that allocate a larger portion of their
budget to these sectors generally report fewer traffic fatalities [25]. However, it is important
to note that the immediate impact of such investments may not always result in lower
fatal traffic. As infrastructure and transportation systems improve, increased construction
and maintenance activities may temporarily contribute to increased traffic volumes and
potential hazards [26,27].
VMT, average total cost per mile, and average annual fuel price are used as indicators
of the economic burden of transportation and the broader economic patterns that influence
individual behavior. Although VMT has a clear impact on motor vehicle fatalities, it is
difficult to isolate and quantify its individual effects due to the interplay between driving
demand and other factors, such as fuel price or average cost per mile. Studies have shown
that fuel prices are significantly correlated with the rate of fatal accidents [28]. When fuel
prices rise, there is often a corresponding decrease in fatal accidents, likely due to reduced
travel demand, shorter travel distances, and more efficient route planning [29,30].
Alcohol consumption is often used as a measure of social behavior and economic
spending patterns. It is influenced by various factors, such as disposable income, cultural
practices, social norms, and stress levels, all of which are related to socioeconomic condi-
tions. Driving under the influence of alcohol and/or psychoactive substances increases
the risk of serious and even fatal motor vehicle accidents [31,32]. In addition, the inter-
action of average alcohol consumption with other socioeconomic factors adds a layer of
complexity to its impact on traffic safety. Economic downturns, characterized by factors
Safety 2024, 10, 11 4 of 18

such as reduced household income and austerity measures, could alter drinking behav-
iors, limiting opportunities for consumption or, conversely, increasing it due to stress and
social disruption.
The laws and regulations imposed by the party in power can have a direct influence
on many of the socioeconomic factors discussed in this document. We would like to further
explore whether a shift in political party has a substantial effect on the total number of fatal
accidents, and thus, we include this as one of the variables in our research.
The interplay of various socioeconomic factors creates a complicated network that
influences how people use vehicles, affects road safety, and ultimately determines the
number of fatal traffic accidents each year. It is crucial to study these factors to develop
well-informed policies that address the economic causes of transportation safety issues.

1.3. Impact of Technological and Safety Features

The evolution of technology within the automotive industry has led to the implemen-
tation of various safety features and technologies aimed at reducing the number of fatal
traffic accidents. These enhancements, encompassing a wide array of features, technologies,
and design principles, are primarily focused on augmenting the safety of drivers and
passengers. The primary objective of these innovations is to decrease the likelihood of
accidents, mitigate the severity of crash injuries, and enhance road safety.
Advanced Driver Assist Systems are central to these technological advancements,
including adaptive cruise control, lane-keeping assistance, blind spot monitoring, and auto-
matic emergency braking. These systems are classified into different levels of automation,
from partial to conditional automation. Research indicates that the utilization of ADAS
can reduce traffic fatalities by diminishing human error [33]. It is suggested that fully
autonomous driving systems are necessary to completely eliminate fatal road accidents
in the future [34]; another critical safety feature is electronic stability control, which helps
drivers maintain vehicle control in difficult driving conditions [35].
Most road accidents arise from human error. Advanced driver assistance systems are
developed to enhance vehicle safety by automating and adapting vehicle systems. These
automated features have been shown to reduce traffic fatalities by minimizing human
errors. ADAS technology has gained significant attention as a research topic over the
past decade, designed to provide critical information for drivers in order to decrease the
number of traffic accidents. A variety of ADASs currently under investigation for intelligent
vehicles are based on Artificial Intelligence, Laser, and Computer Vision technologies with
the potential to decrease both the frequency and severity of traffic accidents. This leads
us towards fully autonomous driving systems as essential to achieving zero fatal road
traffic accidents.
To support this investigation, a report by the National Highway Traffic Safety Admin-
istration highlighted a significant decrease of 81% in the fatality rate per vehicle mile of
travel from 1960 to 2012 [36]. This decline has been attributed, in part, to the introduction
of vehicle safety technologies such as seatbelts, airbags, and electronic stability control
(ESC) in conjunction with programs aimed at promoting the usage of these safety features.
This study will also consider the influence of these technological factors in the context of
road safety.

2. Model Development
The impact of socioeconomic, demographic, and technological factors on the preva-
lence of fatal traffic accidents is a complex issue. Studies have shown that the correlation
between economics and traffic fatalities follows a curved pattern similar to an inverted
U [37,38]. At the beginning of an economic downturn, the initial decline in traffic volume
due to factors such as job loss and reduced consumer spending can result in a decrease in
accidents. But as the recession deepens, other factors can increase the risk of accidents, such
as reduced investment in road infrastructure, decreased spending on vehicle maintenance,
and potential increases in risky behaviors such as driving under the influence, possibly
Safety 2024, 10, 11 5 of 18

due to stress or other social problems related to the recession. These factors can lead to an
eventual increase in fatal accidents, even with lower traffic volumes [39]. At the onset of
economic growth, there is often an increase in vehicle ownership and usage, as more people
can afford cars and there is greater mobility for work and leisure. This increase in traffic
volume can lead to an initial increase in accidents. However, as the economy continues to
grow and stabilize, factors such as improved road safety measures, better infrastructure,
greater public awareness, and the ability to afford safer vehicles come into play, leading to
a reduction in the number of fatal accidents [26,40].
The study of the correlation between traffic accidents and socioeconomic factors is
a multifaceted issue that requires careful examination. Various nonlinear modeling tech-
niques, such as decision trees and random forests, have been used to investigate the severity
of accidents and their relationship with socioeconomic factors [41–45]. Furthermore, ad-
vanced machine learning models like neural networks have been widely used to explore
the impact of socioeconomic factors on fatal accidents [46–48]. These approaches have
provided valuable information on the complex relationship between traffic accidents and so-
cioeconomic factors and have the potential to inform the development of effective accident
prevention strategies.
It is crucial to conduct comprehensive research on the intricate interplay of socioeco-
nomic factors that lead to fatal accidents. Although advanced machine learning techniques
such as neural networks have been adopted, comprehending and interpreting the results of
these models still pose a significant challenge. Addressing these research gaps is paramount
to develop effective policies, gain insight into the underlying determinants of fatal acci-
dents, and create preventive measures. Consequently, it is critical that we prioritize efforts
to bridge these gaps and continue to explore this vital area of study.
In this study, we conducted a comprehensive analysis of various factors, as outlined in
Table 1, with the aim of establishing a causal relationship between these variables and traffic
fatalities. To examine the complex nonlinear dynamics at play, we employed a polynomial
regression model of degree 2, utilizing a Lasso model to optimize the machine learning
approach. This approach is well known for its efficacy in variable selection, regularization,
and reduction in overfitting. By minimizing residual error while imposing constraints
on the model parameters, we encourage sparsity and ensure optimal results [49]. It is
important to note that the utilization of this technique has provided us with an effective
means of examining the multifaceted nature of the data, allowing us to draw meaningful
conclusions that have significant implications for this field of study.
In order to model the relationship between the independent variables listed in Table 1
and the number of fatal crashes, we utilize the equation

Y = W0 + W1 X + W2 X 2 .

This second-degree polynomial transformation is employed to capture any nonlinear

associations that may exist between the variables and the occurrence of deadly accidents.
Lasso regression is then used to identify the pivotal variables and their coefficients, which in
turn minimizes prediction errors. To ensure that these coefficients remain within reasonable
bounds, a regularization parameter is incorporated to constrain the sum of the absolute
value of the regression coefficients. This approach yields a more accurate and meaningful
understanding of the factors that contribute to fatal crashes, thereby informing efforts to
prevent such tragic events in the future.
Safety 2024, 10, 11 6 of 18

Table 1. List of all variables considered in the model discussed in this paper.

Demographics - Average family size Average number of individuals living in a household.

- Higher educational attainment Proportion of individuals who have completed advanced levels of education.
- Number of registered vehicles Total count of registered motor vehicles that are officially registered on public roadways.
- New vehicle sales Total number (in millions) of brand-new motor vehicles sold to consumers in a year.
- Population
Socioeconomic - GDP Total monetary value of all finished goods and services produced in one year.
- Unemployment rate Percentage of the labor force that is unemployed and actively seeking employment,
averaged over a year.
- Average income Average income of all households.
- Minimum wage The lowest legally permissible hourly pay that employers can offer to workers,
established by laws or government regulations.
- Total government spending on The aggregate amount of public funds allocated by a government for the construction,
transportation and infrastructure maintenance, and improvement of transportation systems and physical infrastructure.
- VMT Total number of miles traveled by all vehicles in a year.
- Average cost per mile Average vehicle operating expense for each mile driven.
- Average fuel price Average fuel cost per gallon calculated over a year.
- Average alcohol consumption Average amount of alcoholic beverages consumed per person within a population
over a year.
Other factors - Recession The length of time, measured in months in a year, during which an economy
experiences a significant decline in activity.
- Governing political party The political group that holds the primary authority in a government through election
to executive positions.
Technological - Level 1 The proportion of vehicles that have basic driver assistance systems, such as cruise control
or lane-keeping assistance, where the driver is still primarily in control.
Features - Level 2 The proportion of vehicles that have advanced driver assistance systems capable of controlling
both steering and acceleration/deceleration under certain conditions, while still requiring
the driver’s active engagement.
- Level 3 The proportion of vehicles capable of performing all driving tasks under specific conditions,
but requiring the driver to be ready to take control when the system requests.
- Seatbelt Seatbelt usage under primary seatbelt law enforcement.
- ESC The proportion of vehicles equipped with electronic stability control.
- Safety app Percentage of vehicles equipped with a software application designed to enhance road safety.
- Five-star enhancement Implementation and regulation of five-star safety enhancement.
Safety 2024, 10, 11 7 of 18

3. Data
This study is based on an extensive dataset that covers the number of fatal accidents
in the United States from 1975 to 2020. The data were sourced from reputable government
agencies like the Federal Highway Administration, Department of Labor, and Department
of Commerce, and they include a range of socioeconomic indicators. To ensure accurate
analysis, all monetary values have been adjusted to their 2020 equivalent to account for the
effects of inflation. Overall, the dataset provides a solid foundation for this academic re-
search.
The visual representation shown in Figure 1 presents a comprehensive overview of the
demographic, socioeconomic and various other variables that have been taken into account
in this study. It should be noted that the data do not reveal any direct linear correlation
between any single variable and the number of fatal accidents.

Figure 1. Trends in various societal, economic, and technological indicators such as average family
size, college enrollment, minimum wage, government spending on transportation and infrastructure,
VMT, unemployment rate, GDP, new vehicle sales, average total cost per mile, average gas price,
instances of recession, average alcohol consumption, political party, seatbelt usage, the percentage of
vehicles on the road equipped with levels 1, 2, and 3 of vehicle automation, and, finally, culminating
in the percentage number of fatal accidents between 1975 and 2020.
Safety 2024, 10, 11 8 of 18

The column labeled recession provides a detailed account of the duration of economic
recessions in months. Economic recessions are defined as significant and prolonged declines
in a country’s economic activity, often marked by decreased GDP and negative trends in
employment, consumer spending, business investment, and industrial production.
The variable governing political party serves as an indicator of the governing political
party during the specified study period. As political parties typically have diverse policies
and priorities, particularly in domains such as infrastructure and transportation, they can
have diverse impacts on road safety in both the short and long run. For this particular study,
the Republican and Democratic parties were assigned binary codes of 0 and 1, respectively.
To further analyze the data presented in the seatbelt and autonomous feature columns,
it is important to dig deeper into the academic research surrounding these topics. Numerous
studies have shown that seatbelt use significantly reduces the risk of injury or death in a
car accident. Additionally, research on autonomous features has shown promising results
in terms of improving road safety. As these technologies continue to advance, it is essential
to remain informed and educated about their potential benefits and limitations.
The dataset used in this empirical investigation was divided into two distinct segments,
in which a considerable proportion of 90% of the total data was apportioned for the purpose
of model training and adjustment, while the remaining 10% of the data was used for the
evaluation of the generalizability of the model to novel data. It is worth noting that the
Lasso polynomial regression model is used primarily for data mining activities and not for
predictive analytics. Nonetheless, its competence and dependability remain of paramount
importance, which is ensured through rigorous training and validation on both datasets.
To optimize the parameters, a grid search was performed, resulting in the attainment
of model training and test r2-scores of 0.98 and 0.94, respectively, with a regularization
parameter λ = 5. The r2-scoring function in machine learning, also known as the coefficient
of determination, measures the proportion of the variance in the dependent variable that is
predictable from the independent variables, providing an indication of the goodness of fit
of a model.
Tibshirani [50] suggested that the Lasso model has a significant advantage over other
models due to its L1 regularization. This regularization technique promotes sparsity in
feature coefficients, simplifying the model and reducing overfitting by preventing noise
capture. Additionally, the Lasso model effectively addresses multicollinearity, allowing for
the identification of significant individual features in highly correlated datasets. As shown
in Figure 2, the coefficients derived from the Lasso model exclude features with a coefficient
of zero.
In a Lasso regression model, coefficients play a crucial role in understanding the
influence of predictor variables on the response variable. Positive coefficients in a Lasso
model indicate a direct relationship, meaning that as the predictor variable increases,
the response variable tends to increase as well. This positive correlation suggests that
the feature positively contributes to or is associated with an increase in the outcome
being predicted. On the other hand, negative coefficients signify an inverse relationship.
In this case, an increase in the predictor variable is associated with a decrease in the
response variable. This negative correlation implies that the feature is inversely related
to or diminishes the outcome. The Lasso model is particularly adept at feature selection,
as it can shrink less important variable coefficients to zero, effectively removing them from
the model. This helps in isolating the variables that are most predictive and simplifying
the model by eliminating irrelevant features, leading to a more interpretable and less
complex model.
Safety 2024, 10, 11 9 of 18

Figure 2. Computed coefficients of Lasso polynomial regression model.

4. Results
The coefficients derived from the Lasso regression model are a critical analytical tool
to determine the importance of various predictors for the frequency of fatal traffic accidents.
The model shows that certain features, such as government spending on infrastructure and
transportation and average total cost per mile, have a positive linear correlation with the
number of fatal accidents. This means that an increase in these variables is associated with
a corresponding increase in the predicted number of fatal accidents. This conclusion is
consistent with previous scholarly investigations of these factors [25,26,40,51,52].
The results of this model indicate that there is an inverse relationship between the
frequency of fatal accidents and several other variables, such as governing political party,
level 3 autonomous features, ESC, five-star enhancement, and the use of a safety app. This
means that improvements or increases in these features are likely to reduce the number of
fatal crashes. These findings support earlier discussions and research on the impact of these
factors on road safety, highlighting the complex nature of influences on traffic accident
rates. This underscores the importance of a comprehensive approach in the management
of traffic safety and the formulation of policies. References supporting these findings
include [33–35,53–56].
Our model discovered a nonlinear relationship between the frequency of fatal accidents
and four independent variables. These variables are the average family size, the unemploy-
ment rate, the average alcohol consumption, and their squared values. The correlation is
particularly strong when the nonlinearities of these variables are taken into account. One
noteworthy finding is that the average alcohol consumption has a positive correlation when
squared. This indicates that the number of fatal accidents increases nonlinearly in propor-
tion to the square of the average alcohol consumption. This positive quadratic coefficient
implies a U-shaped relationship, wherein the dependent variable initially increases and
subsequently decreases as the independent variable ascends.
Safety 2024, 10, 11 10 of 18

There is also an inverse relationship between the unemployment rate and average
family size with the number of fatal accidents. This relationship is indicated by a negative
degree-two polynomial correlation. This suggests that an increase in these variables is
associated with a more significant decrease in fatal accidents. Additionally, the presence of
a negative quadratic coefficient indicates an inverted U-shaped curve.
Furthermore, Figure 2 elucidates the interactions among higher-order features, es-
sentially delineating the interplay of the primary features. The linear coefficient for an
interaction term, denoted as Xi × X j , indicates the variation in the dependent variable as a
result of a unit change in one of the interacting variables (Xi or X j ), while maintaining the
other constant. A positive linear coefficient for Xi × X j implies that an increase in the value
of Xi exerts a beneficial influence on the dependent variable when X j is constant, whereas
a negative coefficient suggests an adverse effect.
In terms of quadratic coefficients for interaction terms, these coefficients elucidate
the nonlinear or curvilinear dynamics between the interaction term and the dependent
variable. A positive quadratic coefficient for Xi × X j indicates a U-shaped curve in the
relationship, suggesting that as the product of Xi and X j increases, the dependent variable
initially rises before eventually declining. On the contrary, a negative quadratic coefficient
represents an inverted U-shaped curve.
The interpretation of interaction terms with quadratic coefficients can be complicated
considering the complex interaction and its nonlinear impact on the dependent variable.
Thus, the use of graphical representations or contour plots is often beneficial for clarifying
these relationships. Figure 3 presents a contour plot generated from the Lasso polynomial
regression analysis of the dataset. Contour plots graphically depict the relationships
between two predictor variables and the response variable, utilizing the x and y axes for
the predictor variables and contour lines or color gradations to represent the response
variable’s values. This visualization facilitates an understanding of how the response
variable changes across the ranges of the predictor variables.
In contour plots, the shading density is integral to data interpretation. Regions with
denser shading, depicted as darker areas, typically indicate a higher concentration of
data points, suggesting a more robust correlation or significant relationship between the
variables at these points. Conversely, lighter shaded areas denote a smaller density of data
points, which may point to a weaker or less defined relationship.
Contour plots are invaluable for visualizing both the direction and magnitude of
relationships between variables. The patterns of the contour lines provide key insights;
converging contour lines imply a synergistic interaction, indicating that the combined
influence of the variables is greater than their individual effects. On the other hand, diver-
gent contour lines indicate an antagonistic interaction, where the combined effect of the
variables is weaker than expected. In Figure 3, this concept is visually demonstrated: darker
regions signal areas of convergence, suggesting more powerful synergistic interactions,
while lighter regions point to divergence, indicating weaker antagonistic interactions. Thus,
the darker regions in the plot are pivotal in identifying where the response variable reaches
its optimal values, shedding light on the intricate relationships among the variables. Based
on previous research, there is a negative correlation between the unemployment rate and
the average family size and the number of fatal accidents [57,58]. This means that an
increase in these variables is associated with a decrease in fatal accidents. Additionally,
the presence of a negative quadratic coefficient indicates an inverted U-shaped curve.
Furthermore, Figure 2 illustrates the interactions among higher-order features, es-
sentially delineating the interplay of the primary features. The linear coefficient for an
interaction term, denoted as Xi × X j , indicates the variation in the dependent variable due
to a unit change in one of the interacting variables (Xi or X j ), while maintaining the other
constant. A positive linear coefficient for Xi × X j implies that an increase in the value of
Xi exerts a beneficial influence on the dependent variable when X j is constant, whereas a
negative coefficient suggests an adverse effect.
Safety 2024, 10, 11 11 of 18

These quadratic coefficients for interaction terms elucidate the nonlinear or curvilinear
dynamics between the interaction term and the dependent variable. A positive quadratic
coefficient for Xi × X j indicates a U-shaped curve in the relationship, suggesting that as the
product of Xi and X j increases, the dependent variable initially increases before declining.
On the contrary, a negative quadratic coefficient represents an inverted U-shaped curve.
Interaction between minimum wage and recession: Extensive scholarly inquiries have
delved into the complex dynamics between minimum wage policies and economic reces-
sions. According to established research [59,60], increases in minimum wage levels can
significantly impede employment and income growth for low-skilled workers during peri-
ods of economic downturn, such as the Great Recession. Furthermore, a recent study [61]
highlighted that an increase in the minimum wage resulted in a notable reduction in em-
ployment of 5.6 percentage points among people aged 16 to 30 years who have less than a
high school education.

Figure 3. Contour plot of interaction terms in Lasso polynomial regression.

The interaction between minimum wage regulations and economic contractions also
extends its influence on road safety and fatal accidents. According to established re-
search [59,60], economic policy changes, such as adjustments in the minimum wage, can
indirectly influence various aspects of societal behavior, including road usage and safety.
One possible justification could be that increasing the minimum wage in such a period
may lead to higher disposable income among specific population segments. This increase
in spending power could encourage more commutes and travel, increasing traffic density
and the potential for accidents. Furthermore, suppose that the increase in minimum wage
adversely impacts employment among certain groups. In that case, it may lead to changes
in traffic patterns, for instance, an increase in job-seeking travel or shifts in the times and
Safety 2024, 10, 11 12 of 18

frequencies of travel, which could contribute to a higher risk of road accidents, hence
fatal accidents.
Interaction between minimum wage and average alcohol consumption: The impact
of minimum wage increases on alcohol consumption has been extensively studied in
academia, with varying results. Research in this field has documented both positive and
negative outcomes. For instance, Ref. [62] found that eligibility for increasing minimum
wage significantly raises the frequency of drunkenness among adolescents without a
corresponding increase in overall alcohol consumption. On the other hand, Ref. [63]
reported that the additional income generated from the increases in minimum wage was
primarily allocated to essential expenditures, indicating that low-income households tend
to prioritize fundamental necessities over discretionary or addictive products such as
alcohol. However, Ref. [64] identified a positive correlation between minimum wage hikes
and sales, suggesting that such policy changes could favor the consumption of nondurable
goods, including alcohol. The interaction between minimum wage adjustments and alcohol
consumption is multifaceted and heavily dependent on specific contexts.
Regarding fatal accidents, an increase in minimum wage could potentially have
adverse effects. With more disposable income, individuals might engage more in social
and recreational activities, which may involve alcohol consumption. This increase in
alcohol consumption, coupled with possible risky driving behaviors, could lead to a higher
number of fatal accidents. The combination of higher minimum wages and increased
alcohol consumption can significantly affect decision making processes, increasing the risk
of traffic accidents.
Interaction between VMT and new vehicle sales (in millions): Empirical research
has established a notable correlation between VMT and new vehicle sales, suggesting a
nuanced relationship between car ownership trends and transportation behaviors [65].
An econometric study highlighted that changes in vehicle ownership patterns and prefer-
ences for varying vehicle sizes significantly influence VMT and gasoline consumption [65].
Furthermore, mandated advances in fuel efficiency and the growing adoption of hybrid or
alternative-fuel vehicles are expected to affect fuel tax revenues, a critical funding source
for road maintenance and improvement [66]. Intriguingly, an analysis of new vehicle
sales data has shown that owners of fuel-efficient vehicles tend to increase their driving
frequency each month [67]. These insights are vital for a comprehensive understanding
of transportation safety studies, emphasizing the need to concurrently evaluate VMT and
new vehicle sales.
As shown in the contour plot of VMT versus new vehicle sales in Figure 3, the duration
of recessions about the minimum wage within this dataset produces plausible results.
Data points are more densely concentrated where the minimum wage is relatively high,
coinciding with recession durations of approximately 2 to 6 months, suggesting a scenario
in which higher minimum wages, within the specified duration of economic downturns,
are associated with a higher number of fatal accidents [66]. Interpreting these visual data
underlines the need for a nuanced understanding of the economic factors that can contribute
to road safety outcomes. Possible scenarios could be the following. Higher minimum wages
could increase mobility as people have more disposable income [65]. According to previous
studies, higher minimum wages can attract or retain workers in the labor market who
otherwise would not participate, such as younger or less experienced drivers, who are
statistically more likely to be involved in accidents [66]. Also, a higher minimum wage
might act as a safety net during economic downturns, encouraging continued commuting
and maintenance of prerecession lifestyle patterns that include frequent travel, thus not
reducing the accident rates as might happen with lower wages. This complex behavior
needs further in-depth causal study [66].
Interaction between VMT and average cost per mile: The relationship between VMT
and the average cost per mile of vehicles has been the subject of several studies [68,69].
A study found that households with higher driving intensity are less responsive to the
fuel cost per mile [68]. Meanwhile, another study estimated VMT elasticities in relation to
Safety 2024, 10, 11 13 of 18

the cost per mile and reported that these magnitudes are significant [69]. These findings
suggest that the cost per mile of vehicles can influence VMT and that a VMT tax may be an
effective policy tool to control emissions and raise highway revenue.
According to the contour plot of the VMT-average cost per mile in Figure 3, the high-
est density of data points is found at higher levels of VMT and in the midrange of the
average total cost per mile. This could suggest that as vehicles are driven more miles,
the average cost per mile tends to centralize within a certain range, which might reflect
economies of scale in vehicle operation costs or a common pricing structure for fuel and
maintenance services.
The variability in this graph indicates the influence of various factors on the cost per
mile, such as vehicle type, fuel efficiency, maintenance costs, driving habits, and fuel price,
which may not scale directly with the number of miles driven.
Interaction between unemployment rate and new vehicle sales (in millions): Based
on various studies, it has been established that there exists a strong correlation between
the unemployment rate and the sales of new vehicles [70,71]. When the unemployment
rate increases, the retail sales of new cars tend to decrease. In fact, during the 2007–2009
recession, there was a significant drop of about 20% in vehicle purchases, which can be
attributed to stringent auto loan requirements [72]. The unemployment rate substantially
affects the sales of new vehicles, particularly during economic downturns. This crucial fact
also plays an important role in ensuring transportation safety [70].
The contour plot shown in Figure 3 of the model discussed in this article indicates that
economic conditions, as noted in the unemployment rate, strongly influence the purchasing
behavior of new vehicles. During low unemployment, consumers may feel more financially
secure and are more likely to invest in new cars. On the contrary, when the unemployment
rate increases, economic uncertainty may reduce such significant expenditures.
Interaction between unemployment rate and recession: The relationship between the
length of the recession and the unemployment rate is complex and nonlinear. Studies
have shown that the unemployment rate tends to have an upward trend during frequent
recessions before 1983, whereas long expansions beginning in 1983 are associated with a
downward trend in the unemployment rate [73]. Furthermore, shocks to the unemployment
rate appear to persist in recessions, supporting the hysteresis that workers may lose valuable
job skills in prolonged slumps [74].
The nonlinear relationship between the unemployment rate and recession duration can
be observed in the contour plot of these two variables resulting from the model in Figure 3.
The pattern of the contour lines suggests a nonlinear relationship between the two variables.
The lines are closely packed at the lower end of both axes, showing a steep relationship
between increasing unemployment rates and the lengthening of the recession period.
The plot then fans out, indicating a plateau or a more gradual relationship as unemployment
rates increase. According to the model presented in this article, the complex interaction of
these two variables has a minor positive impact on the number of fatal accidents.
Interaction between unemployment rate and average alcohol consumption: Unem-
ployment is associated with increased alcohol consumption [75]. Studies have shown
that unemployed individuals report higher levels of psychological distress and consume
more alcohol compared to employed individuals [75,76]. Furthermore, higher unemploy-
ment insurance benefits are associated with increased alcohol consumption among the
unemployed [76].
However, some studies have found a negative relationship between unemployment
and alcohol consumption, indicating that economic factors dominate stress-induced changes
in alcohol use [77]. Nevertheless, Popovici and French found that unemployment had a pos-
itive and significant effect on overall alcohol consumption and binge drinking episodes [78].
The nonlinear relationship between unemployment and alcohol consumption is com-
plex and is highlighted by the contour plot of the unemployment rate and the average
alcohol consumption. The denser central region could suggest a peak or a particular range
Safety 2024, 10, 11 14 of 18

of the unemployment rate where alcohol consumption is notably higher, while the spread-
ing out of the lines indicates a less pronounced relationship outside of this central range.
At moderate unemployment levels, where the economic and psychological impacts
of job loss could encourage greater alcohol consumption, there may be a higher risk of
alcohol-related accidents. However, this risk may not increase linearly with unemployment
rates, potentially due to varying social safety nets, community support, or changes in law
enforcement vigilance during different economic conditions.
Interaction between recession and average alcohol consumption: The relationship
between the recession and average alcohol consumption varies across countries and de-
mographics. Some studies suggest that alcohol consumption decreases during a reces-
sion [79,80]. On the contrary, others indicate that certain types of economic loss, such as loss
of employment or housing, can lead to increased alcohol consumption and problems [81].
Factors such as cultural attitudes towards alcohol, government policies, and social pro-
grams also play a role in shaping this relationship [58]. In England, the recession was
associated with a decrease in frequent drinking among the general population, but an
increase in binge drinking among unemployed people [82]. Additionally, the reduction in
fatal automobile accidents during the recession can be attributed to both decreased driving
and changes in driving behavior, including alcohol consumption.
According to the contour plot of recession and average alcohol consumption in
Figure 3, the lighter shaded areas, which suggest fewer observations, are more preva-
lent as the duration of recession either decreases (towards 0 months) or increases (towards
12 months). This may imply that very short or very long recessions do not coincide with
the same levels of alcohol consumption as medium-length recessions. This plot also reveals
that the relationship between the duration of the recession and alcohol consumption is not
linear. Instead, there appears to be a peak in alcohol consumption at a certain point within
the duration of the recession before it either increases less dramatically or decreases.
In the context of fatal accidents, we can observe that higher alcohol consumption
increases the likelihood of alcohol-related accidents. The presented plot highlights a
specific period during the recession where the risk of such accidents is higher. This
information is highly significant for public health officials and policymakers, as it suggests
that interventions aimed at reducing alcohol consumption and impaired driving should be
focused primarily on the middle stages of recessions.

5. Conclusions
In this study, the Lasso polynomial regression model was utilized to explore the
nonlinear relationships between various socioeconomic, demographic, and technological
factors and their impact on fatal traffic accidents. The research findings indicate that
factors such as minimum wage, family size, alcohol consumption, unemployment rate, new
vehicle sales and government spending on transportation and infrastructure are significant
factors impacting the likelihood of fatal accidents. Furthermore, the study revealed that
technological factors such as a higher rate of vehicles with level 3 automation or safety
measures such as safety updates, enhancements, and applications have a direct effect on
reducing the number of fatal accidents.
Furthermore, including a degree-two polynomial regression represents the correlation
between the number of fatal accidents and specific independent variables of higher order.
This discovery underscores the fact that the association between the number of deadly
accidents and socioeconomic factors can be linear and it can also be influenced by higher-
level interactions.
Empirical analysis conducted within this research highlights the significant impact
of higher-order interactions among critical factors such as average alcohol consumption,
unemployment rate, minimum wage, and VMT on the incidence of fatal accidents. These
interactions suggest a complex, multifaceted influence that requires a nuanced examination
of their collective effects on traffic fatalities. Notably, the unemployment rate and average
Safety 2024, 10, 11 15 of 18

alcohol consumption have emerged as significant determinants in the multifactorial model

explored in this study, exerting a considerable influence on fatal accidents.
These findings have important policy implications, particularly during periods of
economic downturn where the employment rate is a crucial factor. Policymakers should
develop comprehensive and multifaceted regulatory frameworks that address these in-
tricate relationships. Such frameworks should aim to mitigate the risks associated with
elevated unemployment rates and alcohol consumption, both of which are implicated in
the increased probability of fatal traffic incidents. This could involve stricter enforcement
of traffic safety laws and regulations and broader social and economic policies that support
employment and manage alcohol consumption during periods of economic instability.
In conclusion, this study highlights the need for policymakers to adopt a holistic
perspective that considers the interconnected nature of economic indicators, behavioral
patterns, and traffic safety outcomes. Integrating these dimensions into a cohesive policy
strategy makes it possible to enhance road safety effectively and sustainably, particularly
when economic indicators predict a heightened risk of fatal accidents.
We suggest several promising directions for future research. Conducting a comparative
analysis in different regions or countries would be informative, as it could reveal how
various socioeconomic, demographic, and technological factors impact traffic deaths in
various geographical contexts. In addition, longitudinal studies over several years would
provide a deeper understanding of trends and evolutions in traffic safety, offering a dynamic
perspective on the interplay of these factors. Another vital area for future exploration is the
implementation and evaluation of policy interventions. Based on the insights gained from
this research, targeted policies could be tested in different settings and their effectiveness in
reducing traffic deaths could be critically evaluated. Lastly, as technological advancements
continue to shape the automotive industry, further research on how emerging technologies,
such as autonomous vehicles and advanced traffic management systems, influence road
safety would be particularly relevant. These potential research paths not only promise to
extend the findings of our current study but also aim to contribute to the broader goal of
improving road safety and reducing traffic-related deaths worldwide.

Author Contributions: N.S.: conceptualization, methodology, software, formal analysis, data cura-
tion, visualization, and writing, original draft. S.B.: conceptualization, validation, and writing. All
authors have read and agreed to the published version of the manuscript.
Funding: A summer research grant from University of North Texas, G. Brint Ryan College of Business,
partially supported this work.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: All data used in this research are within public domain and accessible
via FHWA (https://highways.dot.gov/research/safety/hsis/data) and US Census Bureau (https:
//www.census.gov/).
Conflicts of Interest: Shahram Bohluli was employed by Gradient Systematics LLC, and other
authors declare no conflicts of interest.

References
1. World Health Organization. World Health Organization Road Traffic Injuries. 2023. Available online: https://www.who.int/
news-room/fact-sheets/detail/road-traffic-injuries (accessed on 1 October 2023).
2. Ahmed, S.K.; Mohammed, M.G.; Abdulqadir, S.O.; El-Kader, R.G.A.; El-Shall, N.A.; Chandran, D.; Rehman, M.E.U.; Dhama, K.
Road traffic accidental injuries and deaths: A neglected global health issue. Health Sci. Rep. 2023, 6, e1240. [CrossRef] [PubMed]
3. Ning, H.; Yu, Y.; Bai, L. Analyzing the Impact of Traffic Violation Behaviors on Traffic Fatal Crashes Using Multilevel Models on
Expressways. Math. Probl. Eng. 2022, 2022, 9917877. [CrossRef]
4. Zhu, F.; Wu, X.; Peng, W. Road transportation and economic growth in China: Granger causality analysis based on provincial
panel data. Transp. Lett. 2022, 14, 710–720. [CrossRef]
5. Sagar, S.; Stamatiadis, N. GIS techniques to analyze factors associated with crash occurrence. Traffic Saf. Res. 2022, 2, 000005.
[CrossRef]
Safety 2024, 10, 11 16 of 18

6. Bates, L.J.; Davey, J.; Watson, B.; King, M.J.; Armstrong, K. Factors contributing to crashes among young drivers. Sultan Qaboos
Univ. Med J. 2014, 14, e297.
7. Chand, S.; Li, Z.; Alsultan, A.; Dixit, V.V. Comparing and Contrasting the Impacts of Macro-Level Factors on Crash Duration and
Frequency. Int. J. Environ. Res. Public Health 2022, 19, 5726. [CrossRef] [PubMed]
8. Eksler, V. Road mortality in Europe: How sensitive is it to demographic structure and population dynamics? IATSS Res. 2007,
31, 80–88. [CrossRef]
9. Onishi, H.; Fujiu, M.; Morisaki, Y.; Takayama, J. Analysis of the Relationship between Age and Violation of Traffic Laws and
Ordinances in Traffic Accidents on Children. Sustainability 2022, 14, 12778. [CrossRef]
10. Ulak, M.B.; Ozguven, E.E.; Spainhour, L. Age-based stratification of drivers to evaluate the effects of age on crash involvement.
Transp. Res. Procedia 2017, 22, 551–560. [CrossRef]
11. DeJoy, D.M. An examination of gender differences in traffic accident risk perception. Accid. Anal. Prev. 1992, 24, 237–246.
[CrossRef]
12. Gabany, S.G.; Plummer, P.; Grigg, P. Why drivers speed: The speeding perception inventory. J. Saf. Res. 1997, 28, 29–35. [CrossRef]
13. Factor, R.; Mahalel, D.; Yair, G. Inter-group differences in road-traffic crash involvement. Accid. Anal. Prev. 2008, 40, 2000–2007.
[CrossRef] [PubMed]
14. Mon, E.E.; Jomnonkwao, S.; Khampirat, B.; Satiennam, W.; Ratanavaraha, V. Willingness to pay for mortality risk reduction for
traffic accidents in Myanmar. Accid. Anal. Prev. 2018, 118, 18–28. [CrossRef]
15. Botha, G.J. Road traffic report 2008–2009. In Proceeding of 28th Annual Southern African Transport Conference, Pretoria, South
Africa, 6–9 July 2009.
16. Antoniou, C.; Yannis, G.; Papadimitriou, E.; Lassarre, S. Relating traffic fatalities to GDP in Europe on the long term. Accid. Anal.
Prev. 2016, 92, 89–96. [CrossRef] [PubMed]
17. Yannis, G.; Papadimitriou, E.; Folla, K. Effect of GDP changes on road traffic fatalities. Saf. Sci. 2014, 63, 42–49. [CrossRef]
18. Kuo, P.F.; Lord, D.; Sulistyah, U.D.; Jhamb, A. The impact of the great recession on the spatial patterns of traffic fatalities in Texas:
A spatial point pattern test. In Transportmetrica A: Transport Science; Taylor & Francis: London, UK, 2022; pp. 1–19.
19. Ayodeji, O.A.; Dauda, O.; Oyelami, L.O. Economic growth, motorisation and road traffic safety in Nigeria. Ilorin J. Econ. Policy
2020, 7, 16–34.
20. Akinyemi, Y. Relationship between economic development and road traffic crashes and casualties: Empirical evidence from
Nigeria. Transp. Res. Procedia 2020, 48, 218–232. [CrossRef]
21. Sabia, J.J.; Pitts, M.M.; Argys, L.M. Are minimum wages a silent killer? New evidence on drunk driving fatalities. Rev. Econ. Stat.
2019, 101, 192–199. [CrossRef]
22. Sagar, S.; Stamatiadis, N.; Wright, S.; Cambron, A. Identifying high-risk commercial vehicle drivers using sociodemographic
characteristics. Accid. Anal. Prev. 2020, 143, 105582. [CrossRef]
23. Cox, R.G.; Grant, D.P. Traffic Safety and Human Capital; Department of Economics and International Business, Sam Houston State
University: Huntsville, TX, USA, 2017; Volume 1701.
24. Barajas, J.M. Not all crashes are created equal. J. Transp. Land Use 2018, 11, 865–882. [CrossRef]
25. Albalate, D.; Fernández, L.; Yarygina, A. The road against fatalities: Infrastructure spending vs. regulation?? Accid. Anal. Prev.
2013, 59, 227–239. [CrossRef]
26. Sun, L.L.; Liu, D.; Chen, T.; He, M.T. Road traffic safety: An analysis of the cross-effects of economic, road and population factors.
Chin. J. Traumatol. 2019, 22, 290–295. [CrossRef]
27. Traynor, T.L. Regional economic conditions and crash fatality rates–a cross-county analysis. J. Saf. Res. 2008, 39, 33–39. [CrossRef]
28. Naqvi, N.K.; Quddus, M.; Enoch, M. Modelling the effects of fuel price changes on road traffic collisions in the European Union
using panel data. Accid. Anal. Prev. 2023, 191, 107196. [CrossRef] [PubMed]
29. Chi, G.; Cosby, A.G.; Quddus, M.A.; Gilbert, P.A.; Levinson, D. Gasoline prices and traffic safety in Mississippi. J. Saf. Res. 2010,
41, 493–500. [CrossRef] [PubMed]
30. Naqvi, N.K.; Quddus, M.A.; Enoch, M.P. Do higher fuel prices help reduce road traffic accidents? Accid. Anal. Prev. 2020,
135, 105353. [CrossRef] [PubMed]
31. Baruah, A.M.; Chaliha, R. A study of incidence of alcohol use in fatal road traffic accidents. J. Indian Acad. Forensic Med. 2015,
37, 12–15. [CrossRef]
32. Papalimperi, A.H.; Athanaselis, S.A.; Mina, A.D.; Papoutsis, I.I.; Spiliopoulou, C.A.; Papadodima, S.A. Incidence of fatalities of
road traffic accidents associated with alcohol consumption and the use of psychoactive drugs: A 7-year survey (2011–2017). Exp.
Ther. Med. 2019, 18, 2299–2306. [CrossRef]
33. Batkovic, I.; Ali, M.; Falcone, P.; Zanon, M. Safe trajectory tracking in uncertain environments. IEEE Trans. Autom. Control. 2022,
68, 4204–4217. [CrossRef]
34. Rudin-Brown, C.M.; Burns, P.C. The secret of electronic stability control (ESC). In Proceedings of the Canadian Multidisciplinary
Road Safety Conference XVII, Fredericton, NB, Canada, 3–6 June 2007; pp. 3–6.
35. Elvik, R. Can electronic stability control replace studded tyres? Accid. Anal. Prev. 2015, 85, 170–176. [CrossRef]
36. Kahane, C.J. Lives Saved by Vehicle Safety Technologies and Associated Federal Motor Vehicle Safety Standards, 1960 to 2012—Passenger
Cars and LTVs—With Reviews of 26 FMVSS and the Effectiveness of Their Associated Safety Technologies in Reducing Fatalities, Injuries,
and Crashes; Report No. DOT HS 812 069; National Highway Traffic Safety Administration: Washington, DC, USA, 2015.
Safety 2024, 10, 11 17 of 18

37. Van Beeck, E.F.; Borsboom, G.J.; Mackenbach, J.P. Economic development and traffic accident mortality in the industrialized
world, 1962–1990. Int. J. Epidemiol. 2000, 29, 503–509. [PubMed]
38. Anbarci, N.; Escaleras, M.; Register, C. Income, Income Inequality and the “Hidden Epidemic” of Traffic Fatalities; Technical Report;
Working Paper 05002; Department of Economics, College of Business, Florida Atlantic University: Boca Raton, FL, USA, 2006.
39. Greenwalt, K.R. The Effect of Macroeconomic Conditions on Traffic Fatality Rates across the United States; Wright State University:
Dayton, OH, USA, 2006.
40. Wu, W.; Jiang, S.; Liu, R.; Jin, W.; Ma, C. Economic development, demographic characteristics, road network and traffic accidents
in Zhongshan, China: Gradient boosting decision tree model. Transp. A Transp. Sci. 2020, 16, 359–387. [CrossRef]
41. Ariyathilake, S.N.; Rathnayaka, R.K.T. DNN Classifier and Decision Tree-Based Novel Methodology for Analyzing Road
Accidents. In Proceedings of the 2019 19th International Conference on Advances in ICT for Emerging Regions (ICTer), Colombo,
Sri Lanka, 3–4 September 2019; Volume 250, pp. 1–8.
42. Campos, C.I.d.; Santos, M.C.d.; Pitombo, C.S. Characterization of municipalities with high road traffic fatality rates using macro
level data and the CART algorithm. J. Appl. Res. Technol. 2018, 16, 79–94. [CrossRef]
43. Abellán, J.; López, G.; De OñA, J. Analysis of traffic accident severity using decision rules via decision trees. Expert Syst. Appl.
2013, 40, 6047–6054. [CrossRef]
44. Zermane, A.; Tohir, M.Z.M.; Zermane, H.; Baharudin, M.R.; Yusoff, H.M. Predicting fatal fall from heights accidents using
random forest classification machine learning model. Saf. Sci. 2023, 159, 106023. [CrossRef]
45. El Mallahi, I.; Dlia, A.; Riffi, J.; Mahraz, M.A.; Tairi, H. Prediction of Traffic Accidents using Random Forest Model. In Proceedings
of the 2022 International Conference on Intelligent Systems and Computer Vision (ISCV), Fez, Morocco, 18–20 May 2022; pp. 1–7.
46. Prato, C.G.; Gitelman, V.; Bekhor, S. Pattern recognition and classification of fatal traffic accidents in Israel: A neural network
approach. J. Transp. Saf. Secur. 2011, 3, 304–323. [CrossRef]
47. Gorzelanczyk, P. The Use of Neural Networks to Forecast the Number of Road Accidents in Poland. Sci. J. Silesian Univ. Technol.
Ser. Transp. 2023, 118 45–54. [CrossRef]
48. Cansiz, O.F.; Calisici, M.; Miroglu, M.M. Use of artificial neural network to estimate number of persons fatally injured in motor
vehicle accidents. In Proceedings of the 3rd International Conference on Applied Mathematics, Simulation, Modelling, Circuits,
Systems and Signals, Athens, Greece, 29–31 December 2009; pp. 136–142.
49. Li, H.; Yamamoto, S. A model-free predictive control method based on polynomial regression. In Proceedings of the 2016 SICE
International Symposium on Control Systems (ISCS), Nagoya, Japan, 7–10 March 2016; pp. 1–6.
50. Tibshirani, R. Regression shrinkage and selection via the lasso. J.R. Stat. Soc. Ser. B Stat. Methodol. 1996, 58, 267–288. [CrossRef]
51. Fosser, S. An experimental evaluation of the effects of periodic motor vehicle inspection on accident rates. Accid. Anal. Prev. 1992,
24, 599–612. [CrossRef]
52. Barrios, J.M.; Hochberg, Y.V.; Yi, H. The cost of convenience: Ridehailing and traffic fatalities. J. Oper. Manag. 2023, 69, 823–855.
[CrossRef]
53. Furusawa, S.; Nakamura, H. Human assist control considering shape of target system. Asian J. Control 2021, 23, 1185–1194.
[CrossRef]
54. Gajera, H.; Pulugurtha, S.S.; Mathew, S. Influence of Level 1 and Level 2 Automated Vehicles on Fatal Crashes and Fatal Crash Occurrence;
MTI Publications: New York, NY, USA, 2022.
55. Vorko-Jović, A.; Kern, J.; Biloglav, Z. Risk factors in urban road traffic accidents. J. Saf. Res. 2006, 37, 93–98. [CrossRef] [PubMed]
56. Bener, A.; Al Humoud, S.M.; Price, P.; Azhar, A.; Khalid, M.K.; Rysavy, M.; Crundall, D. The effect of seatbelt legislation on
hospital admissions with road traffic injuries in an oil-rich, fast-developing country. Int. J. Inj. Control Saf. Promot. 2007,
14, 103–107. [CrossRef] [PubMed]
57. Gaygisiz, E. Economic and cultural correlates of road-traffic accident fatality rates in OECD countries. Percept. Mot. Skills 2009,
109, 531–545. [CrossRef] [PubMed]
58. Cotti, C.; Tefft, N. Decomposing the relationship between macroeconomic conditions and fatal car crashes during the great
recession: Alcohol-and non-alcohol-related accidents. BE J. Econ. Anal. Policy 2011, 11. [CrossRef]
59. Clemens, J.; Wither, M. The minimum wage and the Great Recession: Evidence of effects on the employment and income
trajectories of low-skilled workers. J. Public Econ. 2019, 170, 53–67. [CrossRef]
60. Clemens, J.; Wither, M. Additional evidence and replication code for analyzing the effects of minimum wage increases enacted
during the great recession. Munich Personal RePEc Archive. 2017. Available online: https://mpra.ub.uni-muenchen.de/80155/
(accessed on 30 November 2023)
61. Clemens, J. The Minimum Wage and the Great Recession: Evidence from the Current Population Survey; Technical Report; National
Bureau of Economic Research: Cambridge, MA, USA, 2016.
62. Strobel, S.; Peden, A.; Forget, E. Minimum Wages and Adolescent Alcohol Use: Evidence from a Natural Experiment; Working Paper
140011; Canadian Centre for Health Economics: Toronto, ON, Canada, 2014.
63. Bittner, J. How Do Changes in the Minimum Wage Affect Household Consumption? Stat. Stat. Econ. J. 2023, 103, 5–29. [CrossRef]
64. Sabokkhiz, L.; Guven Lisaniler, F.; Nwaka, I.D. Minimum wage and household consumption in Canada: Evidence from high and
low wage provinces. Sustainability 2021, 13, 6540. [CrossRef]
65. Springer, R.K.; Resek, R.W. Econometric model of gasoline consumption, vehicle miles traveled, and new car purchases. Energy
Syst. Policy 1981, 5, 73–87.
Safety 2024, 10, 11 18 of 18

66. Morey, G. Tax Analysis of Vehicle-Miles Traveled. Pol’y Persp. 2022, 29, 1. [CrossRef]
67. Vasudevan, V.; Nambisan, S. A model to estimate passenger vehicle fleet composition, vehicle miles traveled, and fuel
consumption. Public Work. Manag. Policy 2013, 18, 56–81. [CrossRef]
68. Alberini, A.; Burra, L.T.; Cirillo, C.; Shen, C. Counting vehicle miles traveled: What can we learn from the NHTS? Transp. Res.
Part D Transp. Environ. 2021, 98, 102984. [CrossRef]
69. Carroll-Larson, J.; Caplan, A.J. Estimating the effectiveness of a vehicle miles travelled tax in reducing particulate matter
emissions. J. Environ. Plan. Manag. 2009, 52, 315–344. [CrossRef]
70. Sivak, M.; Schoettle, B. Economic Factors and Retail Sales of New Vehicles in the United States; Technical Report, Report No.
SWT-2017-13; The University of Michigan Sustainable Worldwide: Annarbour, MI, USA, 2017.
71. Barreira, N.; Godinho, P.; Melo, P. Nowcasting unemployment rate and new car sales in south-western Europe with Google
Trends. NETNOMICS: Econ. Res. Electron. Netw. 2013, 14, 129–165. [CrossRef]
72. Johnson, K.; Pence, K.M.; Vine, D.J. Auto Sales and Credit Supply; Finance and Economics Discussion Series Divisions of Research
& Statistics and Monetary Affairs Federal Reserve Board: Washington, DC, USA, 2014.
73. Lunsford, K.G. Recessions and the Trend in the US Unemployment Rate; Economic Commentary 2021-01; Federal Reserve Bank of
Cleveland: Cleveland, OH, USA, 2021. [CrossRef]
74. Cevik, E.I.; Dibooglu, S. Persistence and non-linearity in US unemployment: A regime-switching approach. Econ. Syst. 2013,
37, 61–68. [CrossRef]
75. Walia, N.; Bhetuwal, R.; Acosta, L.; Muddasani, S.; Kamwal, C.; Mishra, V.D.; Leszkowitz, D.; Mishra, V. Employment status and
its association with psychological distress and alcohol consumption. Cureus 2021, 13 e16054. [CrossRef]
76. Lantis, R.; Teahan, B. The effect of unemployment insurance on alcohol use and abuse following job loss. Econ. Hum. Biol. 2018,
30, 92–103. [CrossRef]
77. Nelson, J.P. Alcohol, unemployment rates and advertising bans: International panel evidence, 1975–2000. J. Public Aff. Int. J. 2010,
10, 74–87. [CrossRef]
78. Popovici, I.; French, M.T. Does unemployment lead to greater alcohol consumption? Ind. Relat. A J. Econ. Soc. 2013, 52, 444–466.
[CrossRef]
79. Dziadkiewicz, J.; Pisani, A.; Wong, R. Does a Recession have an Impact on Alcohol Consumption? Deakin Pap. Int. Bus. Econ.
2015, 8. [CrossRef]
80. Mulia, N.; Zemore, S.E.; Murphy, R.; Liu, H.; Catalano, R. Economic loss and alcohol consumption and problems during the 2008
to 2009 US recession. Alcohol. Clin. Exp. Res. 2014, 38, 1026–1034. [CrossRef] [PubMed]
81. Harhay, M.O.; Bor, J.; Basu, S.; McKee, M.; Mindell, J.S.; Shelton, N.J.; Stuckler, D. Differential impact of the economic recession
on alcohol use among white British adults, 2004–2010. Eur. J. Public Health 2014, 24, 410–415. [CrossRef] [PubMed]
82. Frone, M.R. The great recession and employee alcohol use: A US population study. Psychol. Addict. Behav. 2016, 30, 158.
[CrossRef] [PubMed]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.