Unformatted Attachment Preview
Follow the 4. Data Collection and Model Formulation in
“traffic accidents in Oman example work.pdf” file. For making
the analysis of below given data.
(Prepare the 5. Results and Analysis and 6. Discussion of
Results.)
PROJECT TITLE
DEVELOPING OF CRASH PREDICTION MODELS FOR SULTANATE OF
OMAN
year
Total Accidents
Fatalities
Injuries
2006
6869
681
7548
2007
8816
798
8531
2008
7982
951
10558
2009
7253
953
9783
2010
7571
820
10066
2011
7719
1056
11437
2012
8209
1139
11618
2013
7829
913
10802
2014
6717
816
3835
2015
6279
675
3624
Total
75244
8802
87802
5. Data Collection
5.1Statistical Data
Traffic accident data was collected from the Ministry of Health and Royal Oman Police for
the last 9 years from 2006 to 2015. SPSS (Statistical Package Computer Program) will be
used to determine the prediction accident models for the last. 9 years according to the
accident report data. The following tables and figures show the statistical data of the
dependent and independent variables.
Table 2: The Eleven Independent Variables
Variable
Independent Variable
x1
Speed
x2
Neglect
x3
Fatigue
x4
Overtaking
x5
Sudden Stopping
x6
Improper Act
x7
drunk Driving
x8
Weather Condition
x9
Safety Distance
x 10
Vehicle DEFECTS
x 11
Road Defects
year
Total Accidents
Fatalities
Injuries
2006
6869
681
7548
2007
8816
798
8531
2008
7982
951
10558
2009
7253
953
9783
2010
7571
820
10066
2011
7719
1056
11437
2012
8209
1139
11618
2013
7829
913
10802
2014
6717
816
3835
2015
6279
675
3624
Total
75244
8802
87802
Table 3: Total Number of Accidents, Fatalities and Injuries (2006-2015)
Table 4: Total Number of Accident Cause (2006-2015)
cause
year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Speed
5236
5344
4734
3744
4032
3861
4328
3952
3510
3411
Neglect
3649
2098
922
615
494
569
721
845
606
655
Fatigue
3
13
13
18
18
27
25
13
10
40
Drunk Driving
76
93
167
172
203
167
134
136
103
53
Overtaking
165
194
419
374
248
397
384
314
224
228
Weather Condition
25
51
36
130
66
22
27
51
19
22
Sudden Stopping
11
36
70
85
47
60
74
25
11
3
Safety Distance
350
269
378
240
252
566
542
553
507
462
Improper Act
263
574
1064
1639
1967
1726
1609
1579
1479
1217
Vehicle Defects
80
128
154
178
188
232
260
243
169
155
Road Defects
11
16
25
58
56
92
105
118
79
33
TOTAL
9869
8816
7982
7253
7571
7719
8209
7829
6717
6279
Table 5: Number of Fatalities with the Accident Causes (2006-2015)
cause
year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Speed
386
458
566
499
487
593
595
471
428
383
Neglect
58
78
74
49
44
70
94
95
87
74
Fatigue
4
9
4
22
6
20
37
12
7
14
Drunk Driving
7
6
7
9
5
15
6
4
10
3
Overtaking
88
122
158
159
97
165
175
125
126
97
Weather Condition
4
6
2
6
5
5
8
2
0
1
Sudden Stopping
0
0
1
2
0
0
0
0
0
1
Safety Distance
3
2
9
9
12
22
16
22
15
8
Improper Act
82
55
80
121
106
92
141
105
86
74
Vehicle Defects
35
58
40
61
47
57
55
61
32
14
Road Defects
14
4
10
16
11
17
12
16
25
6
TOTAL
681
798
951
953
820
1056
1139
913
816
675
Table 6: Number of Injuries With the Accident Causes (2006-2015)
cause
year
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Speed
4454
5038
5395
4074
4340
4932
5424
6923
1752
1728
Neglect
2119
1650
1266
642
457
573
672
845
355
339
Fatigue
4
10
21
30
24
27
33
13
9
49
Drunk Driving
50
77
94
97
110
99
91
136
28
11
Overtaking
357
445
941
913
600
975
908
314
267
196
Weather Condition
39
70
84
198
85
34
51
57
14
17
Sudden Stopping
11
20
82
142
68
106
100
21
9
42
Safety Distance
77
152
576
398
400
1084
941
553
390
285
Improper Act
273
770
1824
2925
3560
3118
2849
1579
835
842
Vehicle Defects
144
289
243
266
333
343
413
243
109
95
Road Defects
20
10
32
89
89
146
136
118
67
20
TOTAL
7548
8531
10558
9783
11437 11437 11618 10802 10802 3624
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
Development of Traffic Accident Models in
Sultanate of Oman
Eman Hamed AlHarrasi, Master Degree in Engineering Project Management, Faculty of Engineering,
Department of Civil Engineering, Al-Isra University
Basim Jrew, Ph.D. Professor
Faculty of Engineering, Department of Civil Engineering, Isra University
Amman, Jordan, Tel 00962-6-4711710, Mobile 00962-7-95974742
E-mail: basim_jrew@yahoo.com
Mohammad Abojaradeh*, Ph.D. Associate Professor, P.E.
Faculty of Engineering, Department of Civil Engineering, Zarqa University P.O. Box 2000, Zarqa
13110, Jordan Tel: 00962-5-3821100, E-mail: abojaradeh@yahoo.com
Abstract
According to Royal Oman Police statistics, more than five hundred people die per year in Oman
because of car accidents. The World Health Organization WHO estimated road traffic death rate
per 100,000 populations to be equal to 30.4 making Oman to be one of the highest countries in
traffic rate accident.
The main objectives of this study are: to analyze traffic accidents in Oman and their main causes
in order to reduce the frequency and severity of traffic accidents. Also, to study the effect of
driver behavior mistakes on traffic safety.
The study will predict traffic accident statistical regression models. These models relate the total
number of accidents, total fatalities, and total injures as dependent variable with possible causes
of accidents as independent variables, focusing on causes that are related to driver behavior as
independent variables. Statistical Package for Social Sciences computer software SPSS was used
for the statistical analysis. Traffic accident data for the last 11 years (2003-2013) was collected
from the Royal Oman Police. The traffic accident causes (independent variables) in Oman are
classified according to Royal Oman Police as: speeding, neglect, fatigue, drunk driving,
overtaking, weather condition, sudden stopping, safety distance, improper act, vehicle defects
and road defects.
The analysis showed that the main causes of accidents are related to the driver behavior. Road
defects, vehicle defects and weather condition were not considered as main causes of accidents
in Oman. Based on the model analysis, countermeasure programs are needed to apply for the
driver behavior that are resulted in the traffic accident statistical regression models.
Key Words: Traffic Accidents, Driver Behavior, frequency, severity, Regression Models, World
Health Organization, Royal Oman Police, Sultanate of Oman.
1|Page
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
1. Introduction
Nearly 1.3 million people die each year as a result of a road traffic collision, more than 3000
deaths each day and more than half of these people are not traveling in a car. Twenty to fifty
million more people sustain non-fatal injuries from a collision, and these injuries are an
important cause of disability worldwide. Ninety percent of road traffic deaths occur in low- and
middle-income countries, which claim less than half the world's registered vehicle fleet. Road
traffic injuries are among the three leading causes of death for people between 5 and 44 years of
age. Unless immediate and effective action is taken, road traffic injuries are predicted to become
the fifth leading cause of death in the world, resulting in an estimated 2.4 million deaths each
year. This is, in part, a result of rapid increases in motorization without sufficient improvement
in road safety strategies and land use planning. The economic consequences of motor vehicle
crashes have been estimated between 1% and 3% of the respective gross national product (GNP)
of the world countries, reaching a total over $500 billion. Reducing road casualties and fatalities
will reduce suffering; unlock growth and free resources for more productive use. (WHO, 2010)
There are many causes for traffic accidents, the human factor, the driver, is the primary cause of
those accidents. The main cause of traffic accidents is disobeying the traffic safety laws, which
include speeding, driver distraction, driving under the influence of drugs or alcohol, close
following between the running cars, yielding for pedestrians and other vehicle etc. The causes of
crashes are usually complex and involve several factors. The main factors can be divided into
four separate categories: the driver, the vehicle, the roadway, and the environment.
1.1 Study Objectives
The main objectives of this study are: to analyze traffic accidents in Oman and their main causes
in order to reduce the frequency and severity of traffic accidents. Also, to study the effect of
driver behavior mistakes on traffic safety.
1.2 Traffic Accident in Sultanate of Oman
Oman has, in contrast to many other countries with high numbers of fatalities in road crashes, a
well-developed road system as well as a relatively new car fleet. According to official reports of
the Ministry of Health (MoH) in Oman, road traffic accident(RTA) problem is the number one
cause of inpatient deaths and the leading cause of serious injury, disability and premature death
among adults (MoH, 2009). The World Health Organization (WHO) has ranked Oman at fourth
place in the Arabian Gulf Co-operation Council (GCC) states and 57th worldwide as far as the
occurrence of traffic accident, injuries and deaths (WHO, 2009). (Mazharul, AlHadhrami, 2012).
WHO present trends in road traffic deaths in Oman from 2001 to 2010 as shown in Figure 1,
(WHO, 2013).
2|Page
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
Figure 1: Trends in Road Traffic Deaths in Oman
2. Review of Related Literature
There are many studies have been conducted in order to study the traffic accident causes, effects
and costs. The following are some of these studies;
Abojaradeh and Jrew (2012) concentrated on the driver behavior mistakes on the traffic safety in
Jordan. They went through a questionnaire to find the main causes of the traffic accidents and
their effect. It was found that there is a direct relationship between driver behavior and the
exposed accidents. (Abojaradeh, Jrew, 2012)
Abojaradeh and Jrew et al., (2009) focuses on major streets and intersections in five main areas
in the Greater Amman Municipality. The study area includes the following areas: Marj AlHamam in Western Amman and Abu Nusair, and Suweileh in North Amman, Jubeeha in Middle
Amman, and Marka in Eastern Amman. The results of the prediction models indicates that the
close following, lane violation, and not taking enough precaution are the most causes of total
accidents. The lane violation and wrong u- turn, and sudden turning are the most causes of
accident injuries. (Abojaradeh, Jrew, 2009)
Mazharul and AlHadhrami (2012) focused on the increase of motorization and road traffic
accidents in Oman. They said that the growth of automobile is faster than the growth of the
Omani population. It has been observed that during the 10 year study period from 2000 to 2009,
the population of Oman increased by about 2.0% per annum, while the automobile fleet in the
country increased by 4.3% per annum. At the same period, the new registration of automobile
increased by 10% per annum. (Mazharul, AlHadhrami, 2012)
Jaddan, et al., (2013) studied the magnitude, cost and potential countermeasures of traffic
accident in Jordan. The study investigates the present and future magnitude of road traffic
accidents in order to provide a better understanding of the road safety trauma and assist strategic
planning and optional allocation of resources. The nimber of accidents has increased five oldfold between 1995 and 2009 reaching 122,793 in 2009 with an estimated cost of JOD 336
million (equivalent to about $504 million). (Jaddan, et al., 2013)
Hong, et al., (2005), they developed models considering the characteristics of roadway
alignments and traffic characteristics. The created models can be used to estimate the accidents
3|Page
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
rates on new or improved roads. In the study, roads were classified into 4 groups based on lanes,
roads level and the existence of median barriers. The regression analysis had been performed for
each group with actual data associated with traffic, roads and accidents. (Hong, et al., 2005)
3. Traffic Accident Causes in Sultanate of Oman
The main objective of the study is to develop traffic accident prediction models that are related to
the total accidents, fatalities and injuries as a dependent variable, with possible causes of
accidents that are related to driver behavior, as independent variables. The independent variables
are shown in Table 1. Traffic accident data was collected from the Ministry of Health and Royal
Oman Police for the last 11 years from 2003 to 2013.
Table 1: Traffic Accident Causes (Independent Variables)
No.
1
2
3
4
5
6
7
8
9
10
11
Independent Variable
Speed
Neglect
Fatigue
Overtaking
Sudden Stopping
Improper Act
Drunk Driving
Weather Condition
Safety Distance
Vehicle Defects
Road Defects
The following are some definitions of accident causes that are related to driver behaviors. Some
causes are clear enough to be defined and others are not related to the driver behaviors.
1. Overtaking
Overtaking is one of the highest risk maneuvers for both drivers and riders because it can put the
overtaking vehicle into the path of oncoming traffic, often at high speeds. If there is a head-on
collision, the speed of both vehicles combines to create a much more severe impact (ROSPA,
2009, www.rospa.com).
2. Neglect
When a person is negligent, it means that he or she has behaved in a thoughtless or careless
manner, which has caused harm or injury to another person. A person can be negligent by doing
something that he or she should not have done (for example, running a red light ), or by failing to
do something that he or she should have done (for example, failing to yield, stop for a pedestrian,
or turn on lights when driving at night).
4|Page
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
Negligence is a legal theory often used in car accident cases. A driver must use care to avoid
injuring other motorists, passengers, or pedestrians -- basically, anyone that he or she encounters
on the road. If a driver is not reasonably careful and injures someone as a result, the driver
is liable for injuring the accident victim (www.nolo.com).
3. Fatigue
Fatigue driving is the driver, after prolonged periods of continuous driving, experiences mental
and physical functional disorder. If the driver does not have a good sleep at the night, even a
short period of driving can still cause fatigue driving. Driving Fatigue will affect the driver's
attention, feeling, perception, thinking, judging, decision making and other aspects. Driving
fatigue is not a morbid state, but a physiological self-protective response. The physical strength
and capabilities could be fully restored after a proper rest. Excessive fatigue driving are the result
of lack of rest compensation and may suddenly show up in some morbid. It will probably lead to
an accident. Fatigue driving cannot be ignored according to the nature of fatigue, Fatigue can be
divided into two Catalogues, one is physical fatigue, and the other is mental fatigue (www.caredrive.com).
4. Safety Distance
It is the safe distance between the vehicle and the vehicle in front, so the vehicle can have
enough time to stop if need to stop suddenly. Drivers must care about their speed when they want
to stop because the faster they are going, the longer it will take to stop.
4. Data Collection and Model Formulation
4.1. Statistical Data
Traffic accident data was collected from the Ministry of Health and Royal Oman Police for the
last 11 years from 2003 to 2013. SPSS (Statistical Package Computer Program) will be used to
determine the prediction accident models for the last 11 years according to the accident report
data. The following tables and figures show the statistical data of the dependent and independent
variables.
Table 2: The Eleven Independent Variables
Variable
x1
x2
x3
x4
x5
x6
x7
x8
x9
x10
x11
5|Page
Independent Variable
Speed
Neglect
Fatigue
Overtaking
Sudden Stopping
Improper Act
Drunk Driving
Weather Condition
Safety Distance
Vehicle Defects
Road Defects
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
Table 3: Total Number of Accidents, Fatalities and Injuries (2003-2013)
Year
Total Accidents
Fatalities
Injuries
2003
10197
578
6735
2004
9460
637
6636
2005
9247
689
6658
2006
9869
681
7548
2007
8816
798
8531
2008
7982
951
10558
2009
7253
953
9783
2010
7571
820
10066
2011
7719
1056
11437
2012
8209
1139
11618
2013
7829
913
10802
TOTAL
94152
9215
100372
Table 4: Total Number of Accidents with the Accident' Causes (2003-2013)
Cause Year
2003
Speed
2831
Neglect
5805
Fatigue
2
Drunk Driving
111
Overtaking
240
Weather Condition
9
Sudden Stopping
23
Safety Distance
446
Improper Act
630
Vehicle Defects
91
Road Defects
9
TOTAL
10197
6|Page
2004
2089
5850
9
118
265
2
18
617
355
128
9
9460
2005
2933
5354
11
79
158
13
4
414
179
91
11
9247
2006
5236
3649
3
76
165
25
11
350
263
80
11
9869
2007
5344
2098
13
93
194
51
36
269
574
128
16
8816
2008
4734
922
13
167
419
36
70
378
1064
154
25
7982
2009
3744
615
18
172
374
130
85
240
1639
178
58
7253
2010
4032
494
18
203
248
66
47
252
1967
188
56
7571
2011
3861
569
27
167
397
22
60
566
1726
232
92
7719
2012
4328
721
25
134
384
27
74
542
1609
260
105
8209
2013
3952
845
13
136
314
51
25
553
1579
243
118
7829
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
Table 5: Number of Fatalities with the Accident' Causes (2003-2013)
Cause Year
Speed
Neglect
Fatigue
Drunk Driving
Overtaking
Weather Condition
Sudden Stopping
Safety Distance
Improper Act
Vehicle Defects
Road Defects
TOTAL
2003
281
139
1
6
83
6
4
1
24
32
1
578
2004
348
143
22
3
91
0
0
1
8
21
0
637
2005
358
193
6
4
75
3
0
1
11
30
8
689
2006
386
58
4
7
88
4
0
3
82
35
14
681
2007
458
78
9
6
122
6
0
2
55
58
4
798
2008 2009 2010 2011 2012 2013
566 499 487 593 595 471
74
49
44
70
94
95
4
22
6
20
37
12
7
9
5
15
6
4
158 159
97
165 175 125
2
6
5
5
8
2
1
2
0
0
0
0
9
9
12
22
16
22
80
121 106
92
141 105
40
61
47
57
55
61
10
16
11
17
12
16
951 953 820 1056 1139 913
Table 6: Number of Injuries with the Accident' Causes (2003-2013)
Cause Year
Speed
Neglect
Fatigue
Drunk Driving
Overtaking
Weather Condition
Sudden Stopping
Safety Distance
Improper Act
Vehicle Defects
Road Defects
TOTAL
7|Page
2003
2533
3092
1
58
361
31
8
92
364
185
10
6735
2004
2012
3363
43
40
548
6
12
134
247
222
9
6636
2005
2644
3183
4
27
404
17
2
108
110
149
10
6658
2006
4454
2119
4
50
357
39
11
77
273
144
20
7548
2007 2008
5038 5395
1650 1266
10
21
77
94
445
941
70
84
20
82
152
576
770 1824
289
243
10
32
8531 10558
2009 2010 2011
4074 4340 4932
642
457
573
30
24
27
97
110
99
913
600
975
198
85
34
142
68
106
398
400
1084
2925 3560 3118
266
333
343
98
89
146
9783 10066 11437
2012
5424
672
33
91
908
51
100
941
2849
413
136
11618
2013
6923
845
13
136
314
57
21
553
1579
243
118
10802
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
Figure 2: Total Number of Accidents, Fatalities and Injuries from 2003 to 2013
4.2 Statistical Model Formulation
SPSS software was used in formulating the regression models by using the method of least
squares. The method of least squares is a procedure to determine the best fit line to date to find
linear relationships between dependent and independent variables. The least square method
(LSM) is probably the most popular technique in statistics. This is due to several factors. First,
most common estimators can be casted within this framework. For example, the mean of a
distribution is the value that minimizes the sum of squared deviations of the scores. Second,
using squares makes LSM mathematically very tractable because the Pythagorean Theorem
indicates that, when the error is independent of an estimated quantity, one can add the squared
error and the squared estimated quantity. Third, the mathematical tools and algorithms involved
in LSM (derivatives, Eigen decomposition and singular value decomposition) have been well
studied for a relatively long time (srmo.sagepub.com).
In this study, the dependent variables are: total number of accidents, fatalities, injuries and
property damage. Otherwise, the independent variables are speed, neglect, fatigue, drunk driving,
overtaking, weather condition, sudden stopping, safety distance, improper act, vehicle defects
and road defects.
The relationship between variables can be written using the following equation:
=
+
+
+ …. +
Where:
= Dependent variable
= Value of the nth independent variable.
= The value of Y when X is equal to zero. This is also called the “Y Intercept”.
= Regression coefficient of the nth independent variable
8|Page
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
This relationship is known as a multiple linear regression model. The multiple linear regression
models are an extension of a simple linear regression model to incorporate two or more
explanatory variable in a prediction equation for a response variable.
The stepwise method was used to form the multiple linear regression models. Stepwise
regression is designed to find the most parsimonious set of predictors that are most effective in
predicting the dependent variable. Variables are added to the regression equation one at a time,
using the statistical criterion of maximizing the R² of the included variables.
Several models were formed and the best models were selected. The selected model has to have
not less than two variables and not more than three or four variables. The multiple linear
regression analysis makes several key assumptions (www.statisticssolutions.com):
Linear relationship
Multivariate normality
No or little multicollinearity
No autocorrelation
Homoscedasticity
Correlation quantifies the extent to which two quantitative variables, X and Y. The correlation
coefficient is scaled so that it is always between -1 and +1. When r is close to 0 this means that
there is little relationship between the variables and the farther away from 0 r is, in either the
positive or negative direction, the greater the relationship between the two variables. The sign of
the correlation coefficient determines whether the correlation is positive or negative. The
magnitude of the correlation coefficient determines the strength of the correlation. The
correlation strength can be described as: (www.sjsu.edu)
0 < |r| < 0.3 weak correlation
0.3 < |r| < 0.7 moderate correlation
|r| > 0.7 strong correlation
4.3 Prediction Regression Models
For the traffic accident prediction models, the independent variables were presented in table 2
and the dependent variables are shown in table 7 below:
Table 7: The Dependent Variables for each Model
Dependent
Variable
Definition
Total Accidents
Total Fatalities
Total Injuries
9|Page
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
5. Results and Analysis
5.1 Development of total traffic accident model (Y1)
Using the SPSS software, the input data are shown in table 4 and the output of the software
program are presented below:
Table 8: Model Summary of the Total Number of Accidents
Model Summaryh
Change Statistics
Std. Error
Mod
R
Adjusted R
of the
R Square
F
Square
Square
Estimate
Change
Change
Sig. F
el
R
1
.903a
.815
.794
456.205
.815
39.654
1
9
.000
2
b
.940
.925
275.068
.125
16.756
1
8
.003
3
c
.986
.972
.961
199.620
.032
8.190
1
7
.024
4
.998d
.996
.993
85.072
.023
32.542
1
6
.001
5
e
.999
.999
33.893
.004
32.800
1
5
.002
f
1.000
1.000
17.313
.000
15.163
1
4
.018
g
1.000
1.000
3.205
.000 113.744
1
3
.002
6
7
.970
1.000
1.000
1.000
df1
df2
Change
a. Predictors: (Constant), x2
b. Predictors: (Constant), x2, x1
c. Predictors: (Constant), x2, x1, x9
d. Predictors: (Constant), x2, x1, x9, x8
e. Predictors: (Constant), x2, x1, x9, x8, x7
f. Predictors: (Constant), x2, x1, x9, x8, x7, x3
g. Predictors: (Constant), x2, x1, x9, x8, x7, x3, x4
h. Dependent Variable: Total_Accidents
The selected model is:
= 1364.124 + 0.868
+ 0.883
+ 1.006
+ 1.306
The model shows a strong correlation (R2 = 0.996) between dependent and independent
variables. The traffic accident model has a strong correlation with neglect, speed, improper act
and safety distance.
10 | P a g e
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
5.2 Development of total fatality model (Y2)
The statistical input data are shown in table 5 and the output of SPSS program are shown as
follows:
Table 9: Model Summary of the Total Number of Fatalities
Model Summaryc
Mod
R
el
R
Adjusted R
Std. Error
Square
Square
of the
Estimate
Change Statistics
R Square
F
df1
Change
Change
df2
Sig. F
Change
1
a
.962
.926
.918
51.686
.926 113.224
1
9
.000
2
.978b
.957
.946
42.088
.030
1
8
.046
5.573
a. Predictors: (Constant), x1
b. Predictors: (Constant), x1, x3
c. Dependent Variable: Total_Fatalities
Table 10: Coefficients of the Total Number of Fatalities
Coefficientsa
Unstandardized
Standardized
Coefficients
Coefficients
Collinearity
Correlations
Statistics
ZeroModel
1
2
B
Std. Error
(Constant)
77.512
73.124
x1
1.659
.156
(Constant)
116.030
61.740
x1
1.480
.148
x3
3.327
1.409
a. Dependent Variable: Total_Fatilities
11 | P a g e
Beta
T
Sig.
1.060
.317
10.641
.000
1.879
.097
.859
10.020
.202
2.361
.962
Toler
order
Partial
Part
ance
VIF
.962
.962
.962
1.000 1.000
.000
.962
.962
.738
.738
1.354
.046
.642
.641
.174
.738
1.354
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
The selected model is:
= 116.030 + 1.480
+ 3.327
The model shows a strong correlation (R2 = 0.957) between dependent and independent
variables. The total traffic accident fatality model has a strong correlation with the speed and
neglect.
5.3 Development of total injury model (Y3)
The statistical input data are shown in table 6 and the output of SPSS program are as follows:
Table 11: Model Summary of the Total Number of Injuries
Model Summaryc
Change Statistics
Mode
l
R
Square
R
a
1
2
.902
.976b
Adjusted R Std. Error of
Square
the Estimate
.813
.953
.792
.942
R Square
Change
894.868
473.647
.813
.140
F
Change
df1
39.175
24.126
Sig. F
Change
df2
1
1
9
8
.000
.001
a. Predictors: (Constant), x8
b. Predictors: (Constant), x8, x4
c. Dependent Variable: Total_Injuries
Table 12: Coefficients of the Total Number of Injuries
Coefficientsa
Model
1
2
Unstandardized
Standardized
Coefficients
Coefficients
B
Std. Error
(Constant)
7052.218
427.134
x8
5.049
.807
(Constant)
5444.379
397.826
x8
3.323
.553
x4
28.989
5.902
a. Dependent Variable: Total_Injuries
12 | P a g e
Beta
Collinearity
Correlations
t
Sig.
16.511
.000
6.259
.000
13.685
.000
.593
6.008
.485
4.912
.902
Statistics
Zero-order Partial Part Tolerance
VIF
.902
.902
.902
1.000
1.000
.000
.902
.905
.458
.596
1.678
.001
.862
.867
.375
.596
1.678
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
The selected model is:
= 5444.379 + 3.323
+ 28.989
The model shows a strong correlation (R2 = 0.953) between dependent and independent
variables. The total traffic accident injury model has a strong correlation with the safety distance
and drunk driving.
6. Discussion of Results
The multiple linear regression analysis was used to find the relationships between the dependent
and independent variables. The eleven independent variables were shown in table 2.
For this data, four models were created and the summary of model results is shown below in
table 15.
Table 15: Summary of the Resulted Regression Models
Model
No.
1
R2
Regression Model
= 1364.124 + 0.868
+ 0.883
+ 1.006
+ 1.306
0.996
2
= 116.030 + 1.480
+ 3.327
0.957
3
= 5444.379 + 3.323
+ 28.989
0.954
6.1 Validation of Model (1) “Total Traffic Accident Number”:
The model is:
= 1364.124 + 0.868
+ 0.883
+ 1.006
+ 1.306
The statistical SPSS output were presented in tables 3.7, 3.8 and 3.9. All independent variables
had negative correlation with the total accidents except x2 and x8 which had a positive
correlation. The variable x2 (neglect) had the highest correlation with accident number. Results
show a strong correlation coefficient between dependent and independent variables (R2 = 0.996).
Which means the proportion of variance in the dependent variable explained by all of the
independent variables is 99.6%. So, variables like speed, neglect, safety distance and improper
act had a 99.6% of predictability about the model. By looking to the P value of the model, it was
found to be less than 0.05 which showed the model significant.
The t statistics for the b coefficients are less than the level of significance of 0.05. That
conducted there is a statistically significant relationship between the number of accident and
speed, neglect, improper act and safety distance.
The value of the constant ( ) is equal to 1364.124, which is the intercept or the predicted value
of X if Y is 0. So, if
scores are 0 the reading score is 1364.124. The values of
( ) are 0.868, 0.883, 1.006 and 1.306 respectively, values that Y will change by if X changes by
13 | P a g e
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
1 unit. All ( ) values are positive, indicating a direct relationships. For example, as the variable
neglect by one increase, the total number of accidents increases by 0.868. By looking closely at
standardized coefficients to compare effect of independent variables, the variable neglect x2
(1.958) had much more influence than the other three variables x1, x9 and x8 which their beta
equal to 0.884, 0.671 and 0.176, respectively.
It is necessary that prediction accident cases from the analysis of data should be verified for
having a confidence with real accident data. The validation of the model can be tested by
comparing between the observed number of accidents and the predicted one. There is a very
slight difference between them. Percentage difference equals the absolute value of the change in
value, divided by the average of the 2 numbers, all multiplied by 100. The percentage difference
can be finding by:
Percentage difference= (| V1 - V2 | / ((V1 + V2)/2)) * 100
The difference was found not more than 1% for each year. For example, the observed number of
total accident in 2003 is 10197 accidents and the predicted number is found to be 10119
accidents. So, the predicted one is 0.8% less than the observed number.
6.2 Validation of Model (2) “Total Fatalities”:
The model is:
= 116.030 + 1.480
+ 3.327
The statistical SPSS output were shown in tables 3.10, 3.11 and 3.12. All independent variables
had positive correlation with the total number of fatalities except x2 and x7 which had a negative
correlation. The variable x1 (speed) had the highest correlation with fatality number. Results
show a strong correlation coefficient between dependent and independent variables (R2 = 0.957).
Which means the proportion of variance in the dependent variable explained by all of the
independent variables is 95.7%. So, variables like speed and fatigue had a 95.7% of
predictability about the model. By looking to the P value of the model, it was found to be less
than 0.05 which showed the model significant.
The constant value ( ) is equal to 116.030; this is the predicted value of total number of
fatalities when all other variables are zero. The coefficient ( ) of the speed is about 1.5. So for
every unit increase in speed, expect an approximately 1.5 point increase in the number of
fatalities, holding all other variables constant. On the other hand, an increase in fatigue will make
a change in the number of fatalities about 3.3 point. By looking closely at standardized
coefficients to compare effect of independent variables, the variable speed x1 (0.859) had much
more influence than the variable fatigue x3 (0.202) on the number of fatalities.
The model validity was tested and the difference was found not more than 10% for each year.
For example, the observed number of fatalities in 2003 was 578 fatalities and the predicted
number was found to be 535 fatalities. So, the predicted one is 8% less than the observed
number.
14 | P a g e
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
6.3 Validation of Model (3) “Total Injuries”:
The model is:
= 5444.379 + 3.323
+ 28.989
The statistical SPSS output were shown in tables 3.13, 3.14 and 3.15. All independent variables
had positive correlation with the total number of injuries except x2 which had a negative
correlation. The variable x2 (neglect) had the highest correlation with the injury number. Results
show a strong correlation coefficient between dependent and independent variables (R2 = 0.953).
Which means the proportion of variance in the dependent variable explained by all of the
independent variables is 95.3%. So, variables like safety distance and drunk driving had a 95.3%
of predictability about the model. By looking to the P value of the model, it was found to be less
than 0.05 which showed the model significant.
The constant value ( ) is equal to 5444.379; this is the predicted value of total number of
injuries when all other variables are zero. The coefficient ( ) of the safety distance is about 3.3.
So for every unit increase in safety distance, expect an approximately 3.3 point increase in the
number of injuries, holding all other variables constant. On the other hand, an increase in drunk
driving will make a change in the number of injuries about 29 point. By looking closely at
standardized coefficients to compare effect of independent variables, the variable safety distance
x8 (0.593) had more influence than the variable drunk driving x4 (0.485) on the number of
injuries.
The model validity was tested and the difference was found not more than 10% for each year.
For example, the observed number of injuries in 2003 was 6735 injuries and the predicted
number was found to be 7431 injuries. So, the predicted one is 10% higher than the observed
number.
7. Conclusions
The following conclusions can be concluded from this research study:
1) The predicted models of the statistical traffic accident data that resulted from SPSS
computer program indicate a strong correlation coefficient between dependent and
independent variables.
2) The statistical results show the models are statically significant and statically highly
significant, where the P-values are less than 0.05 and 0.001.
3) According to the analysis of the traffic accident data and weather condition, vehicle
defects and road defects were not considered to be main causes of traffic accidents in
Oman.
4) Speed appears as the main cause of traffic accidents in the total number of traffic
accidents, fatalities and injuries.
5) Speed, neglect, safety distance between vehicles and improper act are the major causes of
traffic accidents which are related to the driver behaviors. By managing these accident
causes, the number of accidents should be decreased.
15 | P a g e
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
6) The driver behaviors, speed and fatigue are the leading causes of the total number of
fatalities during the eleven years. So, fatality number can be reduced in the future by
managing a countermeasure program especially for the speed and fatigue.
7) The main causes of the accident injuries are the safety distance and drunk driving. Those
causes are related to the driver behaviors. The number of injuries can be reduced by
applying a countermeasure program especially for the safety distance and drunk driving.
8) The validation of the predicted models indicates a small error (less than 10%) for the first
four models. This was found by comparing the predicted results with the actual numbers.
8. Recommendations
According to the previous results the following recommendations can be listed to improve traffic
safety:
1) Traffic accident reduction requires countermeasure programs. The mechanism by which
independent variables (driver behaviors) are selected for application of a countermeasure
should be included as a part of the government highway safety program.
2) A management program (Black spots program) based on before and after study for a
special location in order to determine the accident reduction factor for each
countermeasure.
3) Provide an educational traffic safety programs in schools and universities.
4) Increase low enforcements for speeding, neglecting roles, improper acting overtaking and
drinking of alcohol.
5) More researches are needed with more traffic accident causes due to driver behaviors.
6) A comprehensive traffic safety management program is required for the traffic accident
numbers, fatalities, injuries and property damages.
16 | P a g e
Seventh Traffic Safety Conference, 12-13 May 2015, Amman Jordan
REFERENCES
Abojaradeh M., (2013), “Traffic Accidents Prediction Models to Improve Traffic Safety in
Greater Amman Area". Journal of Civil and Environment Research issued from IISTE USA.
Volume 3, No. 2, 2013. pp 87-101.
Abojaradeh M. A., Jrew B., and D. Abojaradeh, (2009). “Traffic Accidents Prediction Models in
Amman Area in Amman Jordan”. The Forth Conference on Scientific Research in Jordan, Jordan
Society for Scientific Research, 7 November 2009, Amman Jordan. Highway and Traffic in
Jordan, pp 59-80.
Abojaradeh, M., Jrew B. (2012), "The Effect of Driver Behavior Mistakes on Traffic Safety in
Jordan" published and presented in the 6th Traffic Safety in Jordan by the Jordan Traffic
Institute, Jordan 19-20 November 2012.
Hong, D., LEE, Y., YANG, H., KIM, J., W., (2005), “Development of Traffic Accident
Prediction Models by Traffic and Road Characteristics in Urban Areas”, Proceeding of the
Eastern Asia Society for Transportation Studies, Vol. 5, pp. 2046-2061, 2005.
Jaddan, K., Al-Hyari, I., Naghawi, H., Ammourah, R., AlNabulsi, Z., (2013), “ Traffic Safety in
Jordan: Magnitude, Cost and Potential Countermeasures”, Journal of Traffic and Logistics
Engineering, Vol. 1, No. 1, June 2013.
Mazharul, M., Islam and Al Hadhrami A.Y.S (2012), “Increased Motorization and Road Traffic
Accidents in Oman”, Journal of Emerging Trends in Economics and Management Sciences
(JETEMS) 3(6): 907-914
Montgomery D., Runger G. (2010). Applied Statistics and Probability for Engineers Fifth
Edition, USA.
National Highway Traffic Safety Administration (NHTSA) (2010), USA.
Nicholas J. Garber, Lester A. Hoel , (2009)."Traffic and Highway Engineering", Fourth edition,
University of Virginia.
ROSPA, 2009, www.rospa.com
Royal Oman Police (2013), “Facts and Figures 2013”
World Health Organization (2010) , "Global Plan for the Decade of Action for Road Safety
2011-2020", Geneva.
World Health Organization (2013), "Global Status Report on Road Safety: Supporting a Decade
of Action", World Health Organization, Geneva.
srmo.sagepub.com
www.care-driver.com
www.nolo.com
www.sjsu.edu
www.statisticssolutions.com
17 | P a g e