Jurnal Varian

Vol. 8, No. 1, October 2024, pp 1–10

e-ISSN: 2581-2017

Finding the Factors Influencing the Severity of Traffic Accident Victims

in Sleman Regency Using Ordinal Logistic Regression Analysis
Amalia Rizqi Cahyani, Mujiati Dwi Kartikasari
Universitas Islam Indonesia, Yogyakarta, Indonesia

Article Info ABSTRACT

Article history:
Special Region of Yogyakarta (Daerah Istimewa Yogyakarta, DIY) is well-known for its tourist,
Received : 01-10-2024 cultural, and educational attractions, but it also has a high accident rate. Sleman Regency is among
Revised : 08-08-2024 the DIY regions with the greatest number of traffic accidents. To stop the rising number of people
Accepted : 09-06-2024 injured in road accidents, this issue needs to be taken into account. The objective of this study was
to examine
the profile of traffic accidents that happened in Sleman Regency between January and April of 2023
Keywords: and use the ordinal logistic regression method to find characteristics that influence the severity of
traffic accidents. Sleman Regency traffic accident data was used in this study. The opponent’s vehicle
Traffic Accident;
factor, with the category of four or more wheeled vehicles and non-motorized vehicles, is one of the
Ordinal Logistic
elements that influences the severity of traffic accident victims in Sleman Regency, according to the
Regression; Severity of
study’s results. It is intended that local governments will be able to use the study’s results as a
Victims.
guide when
deciding what measures to take to lower the number of traffic accidents.

Accredited by Kemenristekdikti, Decree No:

200/M/KPT/2020 DOI:
[Link]

Corresponding Author:
Copyright ©2024 The
Mujiati Dwi Kartikasari, Authors. This is an open access article under the CC BY-
Department of Statistics, Universitas Islam Indonesia, SA license.
Email: mujiatikartikasari@[Link]

How to Cite:
Cahyani, A. R., & Kartikasari, M. D. (2024). Finding the Factors Influencing the Severity of Traffic Accident Victims in Sleman
Regency Using Ordinal Logistic Regression Analysis. Jurnal Varian, 8(1), 1-10.
This is an open access article under the CC BY-SA license ([Link]

A. INTRODUCTION
The rising number of traffic accidents in Indonesia each year has elevated them to the third-largest cause of death in the
country, trailing only tuberculosis and coronary heart disease (Hidayati & Hendrati, 2017). The number of accidents increases
along with the growing number of motorized vehicles in Indonesia. Traffic accidents affect traffic flow and traffic operations, and
causing serious injuries and even fatalities. The World Health Organization (WHO) has stated that 1.25 million people die every
year worldwide because of traffic accidents (Aqib et al., 2020). Between 20 and 50 million individuals experience injuries each
year because of road traffic collisions, and a significant number of them end up with disabilities resulting from these injuries
(Bahiru et al., 2018). It is evident that road traffic accidents have emerged as one of the leading causes of death and injury
worldwide. The prevention and prediction of traffic accidents have become major topics in the fields of traffic science and
intelligent vehicle research (Yang et al., 2023).
According to Law of the Republic of Indonesia Number 22 of 2009 concerning Road Traffic and Transportation, a traffic accident

1
2 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

is an unexpected and unintentional event on the road involving vehicles with or without other road users resulting in human
casualties or property damage (Law of the Republic of Indonesia No. 22 of, 2009). Traffic accidents are events involving other
road users, either unexpected or intentional. An accident is not a random occurrence; it is accompanied by an identifiable cause,
allowing for preventive action to be taken.
Special Region of Yogyakarta (Daerah Istimewa Yogyakarta, DIY) is a popular tourist destination with a lot of cultural and
educational offerings, but it also has a high accident rate. Under the jurisdiction of the DIY Regional Police, there were 5,420
traffic accident cases in 2022, totaling IDR 23,867,402,885 in losses. Sleman Regency had more traffic accidents than Bantul,
Kulon Progo, Gunung Kidul, and Yogyakarta, according to data from the DIY Central Bureau of Statistics. Yogyakarta City had
490 incidents, Bantul Regency 1,557 accidents, Kulon Progo Regency 786 accidents, Gunung Kidul Regency 762 accidents,
Sleman Regency 1,825 accidents. Yogyakarta’s Sleman Regency has one of the highest rates of road accidents. Because of this,
the research uses Sleman Regency traffic accident data (Statistics Indonesia, 2023). Table 1 presents an analysis of the number of
accidents compared.

Table 1. Number of Regency/City Traffic Accidents in DIY in 2022

Regency/City Number of Minor Injuries Number of Major Injuries Number of Death Total Material Losses (IDR)
Bantul 1,000 1 31 1,557 558,511,800
Kulon Progo 708 12 18 786 21,468,345,763
Gunung Kidul 513 0 18 762 666,255,004
Sleman 605 7 30 1,825 931,120,318
Yogyakarta 448 8 0 490 243,170,000
Total 3,274 28 97 5,420 23,867,402,885

According to reports of traffic accidents, nearly 90% of accidents are caused by human factors, and humans become one of the
most unstable factors in causing traffic safety problems. Traffic accidents are often related to the driver’s driving behaviour, which
is mainly decided by his or her character, such as age, driving experience, and driving style (Hu et al., 2020). Many factors can
influence traffic accidents, including traffic conditions, violations of existing traffic laws by drivers, driving at high speeds, a lack
of a driving licence, failure to obey traffic signs and road markings, and various other forms of violations Wicaksono et al. (2014).
Traffic accidents need to be a shared concern among various stakeholders. Conducting a more in-depth analysis of the factors
influencing traffic accidents is crucial for obtaining reliable results. These results can then serve as a basis for formulating effective
policies to reduce the incidence of traffic accidents.
To find factors that influence traffic accidents, one way to do this is through regression analysis. Regression analysis is a
statistical process that estimates the relationship between variables, including techniques for modelling and analysing several
variables based on the form of the relationship between one or more dependent variables and one independent variable (Armstrong,
2012). The issue is that categorical data is present in traffic accident data. Common regression analysis cannot handle the presence
of categorical data. Using logistic regression is one solution. Logistic regression has various types, including binary logistic
regression, multinomial logistic regression, and ordinal logistic regression. This research discusses ordinal logistic regression
applied to traffic accident cases. Ordinal logistic regression is used to model the relationship between an ordered multilevel
dependent and independent variable. In the modelling, the values of the dependent variable have a natural order or ranking (Lu et
al., 2019). Ordinal logistic regression is a regression analysis used to analyze the relationship between the dependent variable,
which is polycotomous with an ordinal scale, and the independent variable (Zain & Zakariyah, 2015).
Several researchers have conducted research on traffic accidents using ordinal logistic regression. Maylinda & Permana
(2019) used the ordinal logistic regression analysis method to determine the factors that influence the severity of traffic accident
victims in Padang City. Silvia et al. (2015) uses ordinal logistic regression to classify the severity of traffic accident victims in
Semarang. Addini et al. (2022) uses ordinal logistic regression to classify the risk level of accident losses based on driver
characteristics in Bali. Referring to previous research and based on the earlier description that the Special Region of Yogyakarta
has a high accident rate, this study will discuss the factors influencing the severity of traffic accident victims using logistic
regression. This discussion on traffic accidents in the Special Region of Yogyakarta is a novel contribution, as it has not been
addressed before.
In this study, we chose the ordinal logistic regression method to understand the relationship between the dependent variable,
which has more than two categories, and the independent variables. The model used for ordinal logistic regression is a logit model
known as a cumulative logit model. The case involves a dependent variable with more than two ordinal categories, specifically the
severity of traffic accident victims, classified into three levels: accidents with minor injuries, accidents with severe injuries, and
Vol. 8, No. 1, October 2024, pp 1–10
DOI: [Link]
3 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017
accidents resulting in death. Given the high number of traffic accidents in Sleman Regency, researchers conducted an analysis
using

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
4 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

the ordinal logistic regression method to determine the factors significantly affecting the severity of traffic accident victims in
Sleman Regency from January to April 2023.

B. RESEARCH METHOD
This research utilizes secondary data of traffic accident victims from January to April 2023, obtained through the traffic
accident reports of the Sleman District Police, accessed on June 6, 2023. The dependent variable (Y) used is the severity level of
accident victims (minor injuries, serious injuries, and fatalities), and there are 9 independent variables used, namely age (0-25 years
old, 26-45 years old, and more than 46 years old), driver’s license (yes and no), time (crowded and deserted), gender (male and
female), victim’s role (rider, passenger, pedestrians, and cycling), type of accident (front-front, side collision, front-end, single
vehicle accidents, hit-and-run, and other), victim’s profession (self-employed, private employee, student, housewife, civil servant,
teacher, and other), opponent’s vehicle (motorized vehicle, four wheel vehicle/more, and non-motorized vehicle), and the victim’s
vehicle (motorized vehicle, four wheel vehicle/more, and non-motorized vehicle). The analysis methods employed in this research
include descriptive analysis and Ordinal Logistic Regression Analysis. Descriptive analysis was conducted to provide an overview
of the profile of traffic accidents in Sleman Regency from January to April 2023, while Ordinal Logistic Regression Analysis
aimed to identify factors influencing the severity of traffic accident victims. The data is then processed using the ordinal logistic
regression method with the following steps, as the flowchart is presented in Figure 1.

1. Descriptive Statistics
Descriptive analysis, as defined by Misbahuddin & Hasan (2022), is a form of research data analysis used to test the
generalization of research results based on one sample. The variables employed for descriptive analysis include one or more
independent variables, and it does not involve forming comparisons or relationships. Descriptive statistics are part of statistics
that play a role in data collection, presentation, and interpretation of data both in graphic and descriptive form, so that the data
presented can be easily understood or read.

2. Ordinal Logistic Regression

Ordinal logistic regression is an analytical method used to describe the relationship between the dependent variable (Y )
and several independent variables (X). The dependent variable is ordinal, having more than two categories, with each category
being rankable (Hosmer & Lemeshow, 2000). The model used in ordinal logistic regression is a logit model characterized by
cumulative probability. The cumulative logit model is one that can be compared with the cumulative probability, specifically
the probability that is smaller or equal to the j response category on i independent variables, denoted as P (Y ≤ j). The
cumulative probability is defined as shown in Equation 1, —
exp(aj + Σp βkxk)
P (Y ≤ j) = k=1 (1)
p
1 + exp(aj + Σk= β k xk
1
— where P (Y ≤ j) is the value of the cumulative probability of the j response category, αj is a constant value where j =
1, 2, . . . , K − 1, and β1, β2, . . . , βk is the value of the regression coefficient. According to Benoit (2012), if there are
several response categories, then K-1 logit functions will be formed. Based on the cumulative probability, the probability per
response category will be obtained, as expressed in Equation 2.

P (Y = 1) = Σp
exp(α1+
Σk =p1 βkxk )
1+exp(α1+ k=
βkxk )
1

P (Y = j) = Σp Σp
exp(αj + exp(αj−1+
Σk = 1 βkxk ) Σk = 1βkxk (2)
)
1+exp(αj + p −βkxk 1+exp(αj−1 p
βkxk )
k= k=
1 ) + 1

P (Y = K) = 1 Σp
− exp(αK +
Σk =p1 βkxk )
1+exp(αK + k=
βkxk )
1

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
5 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

3. Independence Test
The independence test is a test used to determine the relationship between the independent variable and the dependent
variable. One widely known method for analyzing categorical data is the chi-square test (χ2). This test is valuable for assessing

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
6 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

differences in observations, estimating the frequency of an observation within a specific category, and detecting relationships
between a characteristic and other characteristic in a contingency table. The test is conducted using the chi-square test, and the
test statistics utilized in the independence test are outlined in Equation 3.

I J 2
Σ (nij − eˆi j )
χ2 = Σ eˆij (3)
i=1 j=1

The decision criterion for this test is to reject H0 if χ2(count) > χ2(table) or if the significance value is less than α.
This study adopts a significance level of α (0.05) with the following hypotheses:
H0: There is no relationship between the dependent variable and the observed independent variable.
H1: There is a relationship between the dependent variable and the observed independent variable.

4. G Test (Overall)
According to Hosmer & Lemeshow (2000), the overall test is conducted to examine the role of each independent variable
in the model together or in other words to determine whether the independent variable has an overall effect on the dependent
variable. To obtain a decision, the test value of G is compared with the value X2(α,p) with the rejection criterion of rejecting H0
if the value of G > (α,p) . The test statistics used are shown in Equation 4.
X2

likelihood (model B)
G = −2 ln likelihood (model A) (4)

5. Wald Test
The Wald test was conducted to test whether the independent variables were able to influence the dependent variable in
the research. The significance level used is 5% or 0.05. The Wald test statistic follows a normal distribution, so to obtain the
test decision results by comparing the W value with the Z α value. The hypothesis used is to reject H0 if |W| > Z α or if the
2 2

p − value < α. The Wald statistical formula is contained in Equation 5.

βˆ j
W= (5)
SE(βˆj )

6. Backward Elimination
Backward elimination is a step backward by regressing all independent variables with the dependent variable. The
elimina- tion process targets independent variables with a significance value exceeding the α (0.05) threshold. The backward
step involves systematically removing variables with high significance values one by one, and this process is repeated until all
variables have significance values less than α (0.05) (Maulidina et al., 2021).

7. Model Interpretation
The interpretation of coefficients in the ordinal logistic regression model can be accomplished by utilizing the odds ratio
(OR). The odds ratio represents the ratio between the likelihood of an event occurring in a case group compared to a control
group. In interpreting ordinal logistic regression coefficients, the odds ratio used expresses the comparison of risk levels among
two or more categories on an explanatory variable, with one of the categories serving as a reference. The interpretation of the
model is conducted by initially calculating the odds ratio obtained from exp(β) (Putri & Budyanra, 2020).

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
7 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

Figure 1. Conceptual Framework of the Research

C. RESULT AND DISCUSSION

1. Overview of January-April 2023 Traffic Accidents in Sleman Regency
The severity of the victims’ injuries is categorized into three levels: minor injuries, serious injuries, and fatalities.
Research data indicates that traffic accidents predominantly result in minor injuries, constituting the highest percentage at 75.98%,
followed by victims with severe injuries at 20.72%, and fatalities at 3.30%. Additionally, there is a fluctuation in the number
of traffic accidents in Sleman Regency from January to April 2023. In January, there were 314 accidents, followed by 173 in
February, 30 in March, and 120 in April. The variation can be attributed to increased traffic, especially during the New Year’s
celebration holiday at the beginning of the year, leading to elevated vehicle mobility and consequently, a higher number of
accidents in January. Subsequently, from February to March, there was a decline in accidents due to the approach and
implementation of the month of Ramadan, prompting a reduction in outdoor activities. However, in April, there was a surge in
accidents due to the commencement of the Eid homecoming flow, leading to increased vehicle traffic, particularly on roads
such as Jalan Magelang, Jalan Wates, and Jalan Solo (Janati, 2023).

2. Independence Test
The results of the independence test of the victim’s severity with the variables age, driver’s license, time, gender, victim’s
role, accident type, victim’s profession, opponent’s vehicle, and victim’s vehicle are shown in Table 2.

Table 2. Independence test of victim severity with independent variables

Variables df χ2 count χ2 table P − value Decision
Age 4 8.471 9.488 0.076 Fail to Reject H0
Driver’s License 2 4.516 5.991 0.105 Fail to Reject H0
Time 2 6.720 5.991 0.035 Reject H0
Gender 2 5.582 5.991 0.061 Fail to Reject H0
Victim’s Role 6 29.085 12.592 0.000 Reject H0
Type of Accident 10 57.943 18.307 0.000 Reject H0
Victim’s Profession 14 26.026 23.685 0.026 Reject H0
Opponent’s Vehicle 4 42.074 9.488 0.000 Reject H0
Victim’s Vehicle 4 20.075 9.488 0.000 Reject H0

The results show that the variables of time, victim’s role, accident type, victim’s profession, opponent’s vehicle, and victim’s
vehicle have a χ2count value that is greater than the χ2table value and a p-value that is each smaller than α (0.05) which means reject
H0 so that it can be concluded that using a confidence level of 95% there is a relationship between the severity of traffic
accident victims in January-April 2023 in Sleman Regency with time, victim’s role, accident type, victim’s profession, opponent’s
vehicle, and victim’s vehicle.

3. G Test (Overall)
G-Test is a likelihood ratio test used to test the role of the independent variables in the model jointly carried out with the
G-Test. The results obtained from the simultaneous test are shown in Table 3.

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
8 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

Table 3. Independence test of victim severity with independent variables

df G χ2 P-Value
table
20 107.91 31.410 0.000

Table 3 shows that the G value is 107.91, the χ2table value is 31.410, the df obtained is 20, and the p-value is 0.000. This
means that the value of G (107.91) > χ2(0.05,2 (31.410) or p-value (0.000) < α (0.05), it can be concluded that with a
0) H0, meaning that there is at least one variable that has a significant effect on the
confidence level of 95% the existing data reject
severity of traffic accident victims.

4. Wald Test
The Wald test is used to test the parameters partially, using the results of the Wald statistic to test the effect of the
independent variables of time, victim’s role, type of accident, victim’s profession, opponent’s vehicle, and victim’s vehicle
partially on the dependent variable of the severity of traffic accident victims. The results obtained from the partial test are
contained in Table 4.

Table 4. Independence test of victim severity with independent variables

Variable Value P-Value Decision
Victim’s Role: Passenger (X5(2)) 0.667 0.005 Reject H0
Type of Accident: Hit-and-Run (X6(5)) -1.920 0.002 Reject H0
Opponent’s Vehicle: Four-Wheeled Vehicles/More (X8(2)) 0.812 0.001 Reject H0
Victim’s Vehicle: Four-Wheeled Vehicles/More (X9(2)) 0.909 0.002 Reject H0

Based on Table 4, which has a significant status means that the variables are factors that affect the severity of victims of
traffic accidents, namely the victim’s role (passenger), type of accident (hit-and-run), opponent’s vehicle (four-wheeled
vehicles/more), and victim’s vehicle (four-wheeled vehicles/more) variable. While the other variables are results that do not
have a significant effect on the dependent variable, this is indicated by the p-value which is greater than the α value (0.05).

5. Backward Elimination
The results of the backward elimination process until a significant variable is obtained are listed in Table 5.
Table 5. The process of Backward Elimination
P-value
Variable Category
Start Model 1 Model 2 Model 3 Model 4 Model 5
Time Deserted 0.704 0.904 0.866 - - -
Victim’s Role Passenger 0.005 0.001 0.006 0.006 - -
Pedestrians 0.894 0.727 0.738 0.736 - -
Cycling 0.894 0.596 0.772 0.774 - -
Type of Accident Side Collision 0.929 - - - - -
Front-End 0.699 - - - - -
Single-Vehicle Accidents 0.327 - - - - -
Hit-and-Run 0.002 - - - - -
Other 0.565 - - - - -
Victim’s Profession Self-Employed 0.101 0.059 - - - -
Private Employee 0.151 0.100 - - - -
Student 0.557 0.560 - - - -
Housewife 0.299 0.223 - - - -
Civil Servant 0.627 0.392 - - - -
Teacher 0.914 0.984 - - - -
Other 0.915 0.964 - - - -
Opponent’s Vehicle Four-Wheeled Vehicles/More 0.001 0.000 0.000 0.000 0.000 0.000
Non-Motorized Vehicle 0.515 0.000 0.000 0.000 0.000 0.000
Victim’s Vehicle Four-Wheeled Vehicles/More 0.002 0.009 0.002 0.002 0.003 -
Non-Motorized Vehicle 0.532 0.376 0.272 0.271 0.065 -

Backward model selection involves systematically removing variables one by one, starting with the ones with the highest
p-value. The red mark in Table 5 shows the largest p-value in the last model, which is more than the significance level α

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
9 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017
(0.05),

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
10 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

so the backward process is carried out at that stage. The backward process concludes when all remaining variables have a p-
value less than the significance level α (0.05). The results obtained from backward elimination are shown in Table 6.

Table 6. The result of Backward Elimination

Variabel Value P-Value Decision Description
Opponent’s Vehicle: Four-Wheeled Vehicles/More (X8(2)) 0.828 0.000 Reject H0 Significant
Opponent’s Vehicle: Non-Motorized Vehicle (X8(3)) -2.590 0.000 Reject H0 Significant
Minor Injuries — Serious Injuries 1.138 0.000 Reject H0 Significant
Serious Injuries — Fatalities 3.434 0.000 Reject H0 Significant

After knowing the influential variables, the next is to form the final ordinal logistic regression in traffic accident cases in
January-April 2023 in Sleman Regency after backward and simultaneous and partial parameter testing results in variables that
have a significant effect, namely the opponent’s vehicle, the logit function model obtained is:

Logit 1 = 1.138 + 0.828X8(2) − 2.590X8(3)

Logit 2 = 3.434 + 0.828X8(2) − 2.590X8(3)
The factors that influence the severity of traffic accident victims in Sleman Regency are the opponent’s vehicle factor for
four-wheeled vehicle/more, and the opponent’s vehicle factor for non-motorized vehicle. Based on the logit value, the
probability of the severity of traffic accident victims can be calculated.

Table 7. Probability of victim severity

Minor Serious
Independent Variable Fatalities
Injuries Injuries
Opponent’s Vehicle (four-wheeled vehicle/more) 0.877 0.109 0.014
Opponent’s Vehicle (non-motorized vehicle) 0.190 0.510 0.301

Based on Table 7, it can be observed that the probability of opponent’s vehicle with four-wheeled vehicle/more
experiencing minor injuries is 0.877, the probability of serious injuries is 0.109, and the probability of fatality is 0.014.
Meanwhile, the probability of opponent’s vehicle with non-motorized vehicle experiencing minor injuries is 0.190, the
probability of serious injuries is 0.510, and the probability of fatality is 0.301.

6. Model Interpretation
The odds ratio value based on the severity of the victim’s injury and the category of the opponent’s vehicle, the
interpretation of the model containing the odds ratio value are shown in Table 8.

Table 8. Odds ratio

Variable Odds Ratio
Opponent’s Vehicle (four-wheeled vehicle/more) 2.290
Opponent’s Vehicle (non-motorized vehicle) 0.075

The odds ratio value contained in Table 8 is obtained for victims of traffic accidents with opponent’s vehicle (four-
wheeled vehicle/more) have a risk of experiencing accidents from minor injuries to serious injuries or from serious injuries to
fatalities by 2,290 times than victims of traffic accidents using motor vehicle. This means that the risk of victims of accidents
with opponent’s vehicle types (four-wheeled vehicle/more) is greater than that of victims of accidents with opponent’s vehicle
types (motorcycles). Furthermore, victims of traffic accidents with opponent’s vehicle (non-motorized vehicle) have a risk of
experiencing accidents from minor injuries to serious injuries or from serious injuries to fatalities of 0.075 times that of victims
of traffic accidents using motor vehicle. This means that the risk of accident victims with the opponent’s vehicle type
(motorized vehicle) is greater than accident victims with the opponent’s vehicle (non-motorized vehicle).
Research by Maylinda & Permana (2019) shows that the age of the victim, type of accident, and opponent’s vehicle are
factors that influence the severity of traffic accident victims in Padang City. Research by Silvia et al. (2015) indicates that the
age of the victim and opponent’s vehicle are factors influencing the severity of traffic accident victims in Semarang City.
Research by Addini et al. (2022) demonstrates that opponent’s vehicle, gender, education level, religion, and age of the victim
are factors
Vol. 8, No. 1, October 2024, pp 1–10
DOI: [Link]
11 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

affecting the severity of traffic accident victims in Bali. Each area has its characteristics, including traffic characteristics. In
the Special Region of Yogyakarta, the factor that influences the severity of traffic accident victims is opponent’s vehicle factor,
with the category of four or more wheeled vehicles and non-motorized vehicles. As one of the areas in DIY with the highest
traffic accident rate, it is necessary to study further related factors that can affect the severity of traffic accident victims in
Sleman Regency. In addition, it is required to conduct an analysis using other methods as a comparison of the ordinal logistic
regression used in this study.

D. CONCLUSION AND SUGGESTION

Based on the results obtained, several conclusions were obtained from the analysis carried out. In January-April 2023 there
were 637 traffic accidents in Sleman Regency, Yogyakarta. The highest number of accidents occurred at the beginning of the year,
namely in January, while the lowest number of accidents occurred in March. Generally, accident victims suffer minor injuries.
Victims aged 0-25 years are more often involved in accidents. Not having a driver’s license is the highest number compared to
people who have a driver’s license. The number of accidents during busy times is higher than during quiet times. The male gender
is more likely to experience traffic accidents than the female gender. The role of the victim who experiences accidents more often
is the driver. The most frequently involved types of accidents were other accidents (side-by-side and pedestrian/cyclist). The
profession of victims more often involved is students. The type of vehicle for both opponents and victims that is often involved is
generally a motorbike. Based on the output from backward elimination, the factors that influence the severity of traffic accident
victims in Sleman Regency in January-April 2023 are the opponent’s vehicle factor for four/more wheeled vehicles, and the
opponent’s vehicle factor for non- motorized vehicles. It is hoped that local governments would be able to use the study’s findings
as a reference for determining what steps should be made to lower the frequency of traffic accidents.

REFERENCES
Addini, F. F., Haryanto, D., & Maolani, R. A. (2022). Klasifikasi Tingkat Risiko Kerugian Kecelakaan berdasarkan Karakteristik
Pengemudi dengan Analisis Regresi Logistik Ordinal. Jurnal Matematika Integratif, 18(2), 167. https : / / doi . org /
10 . 24198/jmi.v18.n2.41317.167-177
Aqib, M., Mehmood, R., Alzahrani, A., & Katib, I. (2020). In-Memory Deep Learning Computations on GPUs for Prediction of
Road Traffic Incidents Using Big Data Fusion [Series Title: EAI/Springer Innovations in Communication and
Computing]. In R. Mehmood, S. See, I. Katib, & I. Chlamtac (Eds.), Smart Infrastructure and Applications (pp. 79–
114). Springer International Publishing. [Link] 4
Armstrong, J. S. (2012). Illusions in regression analysis. International Journal of Forecasting, 28(3), 689–694. [Link]
10.1016/[Link].2012.02.001
Bahiru, T. K., Kumar Singh, D., & Tessfaw, E. A. (2018). Comparative Study on Data Mining Classification Algorithms for Pre-
dicting Road Traffic Accident Severity. 2018 Second International Conference on Inventive Communication and
Computational Technologies (ICICCT), 1655–1660. [Link]
Benoit, K. (2012). Multinomial and Ordinal Logistic Regression.
Hidayati, A., & Hendrati, L. Y. (2017). Traffic Accident Risk Analysis by Knowledge, the Use of Traffic Lane, and Speed. Jurnal
Berkala Epidemiologi, 4(2), 275. [Link]
Hosmer, D. W., & Lemeshow, S. (2000, September). Applied Logistic Regression (1st ed.). Wiley. https : / / doi . org / 10 .
1002 / 0471722146
Hu, L., Bao, X., Wu, H., & Wu, W. (2020). A Study on Correlation of Traffic Accident Tendency with Driver Characters Using
In-Depth Traffic Accident Data. Journal of Advanced Transportation, 2020, 1–7. [Link]
Janati, C. D. (2023, April). Dear Pemudik, Ini Titik-Titik Rawan Kecelakaan Lalu Lintas di Sleman. Retrieved October 28, 2024,
from [Link]
Law of the Republic of Indonesia No. 22 of. (2009). Traffict and Road Transport. [Link]
no-22-tahun-2009

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
12 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

Lu, P., Wang, H., & Tolliver, D. (2019). Prediction of Bridge Component Ratings Using Ordinal Logistic Regression Model (J. Li,
Ed.). Mathematical Problems in Engineering, 2019(1), 9797584. [Link]
Maulidina, F., Rustam, Z., Hartini, S., Wibowo, V. V. P., Wirasati, I., & Sadewo, W. (2021). Feature optimization using Backward
Elimination and Support Vector Machines (SVM) algorithm for diabetes classification. Journal of Physics: Conference
Series, 1821(1), 012006. [Link]
Maylinda, R., & Permana, D. (2019). Faktor-Faktor Yang Mempengaruhi Tingkat Keparahan Korban Kecelakaan Lalu Lintas Di
Kota Padang Menggunakan Analisis Regresi Logistik Ordinal [Number: 3]. Journal of Mathematics UNP, 4(3), 55–59.
[Link]
Misbahuddin, M., & Hasan, I. (2022, August). Analisis Data Penelitian dengan Statistik (Edisi Kedua). Bumi Aksara.
Putri, N. I., & Budyanra, B. (2020). Penerapan Regresi Logistik Ordinal dengan Proportional Odds Model pada Determinan
Tingkat Stres Akademik Mahasiswa. Seminar Nasional Official Statistics, 2019(1), 368–378. https : / / doi . org /
10 . 34123 / semnasoffstat.v2019i1.104
Silvia, C., Wilandari, Y., & Hoyyi, A. (2015). Ketepatan Klasifikasi Tingkat Keparahan Korban Kecelakaan Lalu Lintas
Menggunakan Metode Regresi Logistik Ordinal dan Fuzzy K-Nearest Neighbor in Every Class [Number: 3 Publisher:
Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro]. Jurnal Gaussian, 4(3), 441–
451. [Link] 10.14710/[Link].4.3.441-451
Statistics Indonesia. (2023). Daerah istimewa yogyakarta province in figures 2023. BPS Statistics of D.I Yogyakarta
Province. [Link]
Wicaksono, D., Fathurochman, R. A., Riyanto, B., & Wicaksono, Y. (2014). Analisis Kecelakaan Lalu Lintas (Studi Kasus -Jalan
Raya Ungaran - Bawen) [Number: 2 Publisher: Departemen Teknik Sipil, Fakultas Teknik Universitas Diponegoro].
Jurnal Karya Teknik Sipil, 3(2), 345–355. [Link]
Yang, J., Han, S., & Chen, Y. (2023). Prediction of Traffic Accident Severity Based on Random Forest (I. Ghosh, Ed.). Journal of
Advanced Transportation, 2023, 1–8. [Link]
Zain, I., & Zakariyah, Z. (2015). Analisis Regresi Logistik Ordinal pada Prestasi Belajar Lulusan Mahasiswa di ITS Berbasis
SKEM [Publisher: Sepuluh Nopember Institute of Technology]. Jurnal Sains dan Seni ITS, 4(1), 15585.
[Link] 12962/j23373520.v4i1.9403

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]
13 | Amalia Rizqi Cahyani JURNAL VARIAN | e-ISSN: 2581-2017

[This page intentionally left blank.]

Vol. 8, No. 1, October 2024, pp 1–10

DOI: [Link]