Statistical Method For Transport ECV 5602 Individual Assignment
Statistical Method For Transport ECV 5602 Individual Assignment
ECV 5602
INDIVIDUAL ASSIGNMENT
Presented by
Ø Sa’adiah Mohd Yusoh GS46511
INTRODUCTION
OBJECTIVES
The purpose of this task is to study the VKT between Cars and
Motorcycles. The objectives are as follows:
1 To determine annual vehicle kilometer travel (VKT) among
Malaysia motor vehicle users
To compare annual VKT between different groups of
2 motor vehicle users (Age, household income, gender
and etc.)
3
To investigate the impact of demographic factors (income,
age, education level) and family characteristics
(household size, number of kids, and etc.) on annual VKT.
LITERATURE REVIEW
Vehicle Kilometers Travelled (VKT) is the total in kilometers travelled by motor
vehicles on any particular road systems during a given period of time . VKT is an
indicator commonly used; to measure road safety performance in a particular country
and as an indicator for international traffic safety comparison purposes.
The parameter provides additional information about the crash risk in terms of exposure
for certain groups of road users. For instance, drivers who travel long distances are more
at risk to be involved in an accident than those who make shorter trips.
The VKT survey has been conducted in Malaysia since 2003. The results of the
previous VKT survey (2007) showed that the average passenger car in Malaysia
covered a distance of 20,499 km while the VKT for motorcycles were 15,879 km. In
addition to this, VKT is also utilised as a very useful indicator in describing the risk of
travelling on the roads of any particular country .
LITERATURE REVIEW(CONT..)
Figure 1 shows the ranking of countries in terms of deaths per billion VKT in
2009 respectively. In relation to road safety performance, the fatality index
(death/ billion VKT) for Malaysia in the year 2009 was 17.27. It means that
there were 17 deaths per every billion kilometer travelled in Malaysia in that
year. (Source : MIROS)
Figure 1: Deaths per billion Vehicle-kilometers in 2009 (Source: MIROS)
RESEARCH DESIGN
• Descriptive study
Methodology • Inferential study
• Field study
How?
• Personal interview, email, WhatsApp
Approach
• Random survey, families, friends,
Respondent officemates, neighbors
7
METHODOLOGY
1) Set the objectives & hypothesis
2) Preparation of Questionnaires
3) Conduct Survey
4) Analyze the data
5) Conclusion
QUESTIONNAIRE
The questionnaire is divided into TWO (2) parts.
Part A is the Demographic Data, where participants would easily (/) tick
their own description. The data to be collected in this part are Gender, Age,
Education Level, Occupation Field and Monthly income Range.
Part B is the odometer reading of the participants’ vehicle. The reading is
recorded between 9 to 15 days. The date is also recorded to avoid error.
The questionnaire is distributed through e-mail, in-person and by-hand.
Participants are also requested to send the photographs of their odometer
reading to avoid mistakes in recording the data.
This study also involved known respondent where there are only verbal
information collected without questionnaire.
QUESTIONNAIRE
SIMPLE
QUESTIONNAIRE
1) Respondent Name : ______________
2) Age : ______
3) Gender : M/F
4) Monthly Income (RM) : _______
5) Highest education level :_______
6) Vehicle Reg. Number :_________
7) Odometer Reading
i. Reading Day 1 :__________
Date :
ii. Reading Day 14 :__________
Date:
SAMPLE OF THE RAW DATA
The data collected are tabulated and shown in Table 1 and 2:
Odometer
Education Income range Annual
No Name Age Gender Occupation VKT
level RM(K) 1 2 VKT
MOTORCAR
1 Azira 34 F Engineer Degree E 25512 25869 357
4284
2 Mat Yie 38 M Supervisor Degree E 39519 40449 930 11160
TABLE 1 : 3 Amir 35 M Technician Certificate B 56660 56995 335 4020
DATA 4 Fauziah 58 F Entrepreneur Diploma E 243352 243812 460 5520
5 Bayati 29 F Storekeeper Certificate A 163099 163205 106 1272
COLLECTED 6 Zafril 28 M Entrepreneur Certificate C 334281 333545 736 8832
FOR 7 Qila 26 F Student Diploma A 226304 226761 457 5484
MOTORCAR 8 Thilis 28 M Engineer Degree D 99221 100054 833 9996
9 Izman 44 M Manager Degree E 27768 28150 382 4584
10 Hamdan 31 M Engineer Diploma C 192874 193139 265
3180
11 Ijan 35 M Engineer Degree E 184185 184371 2232
186
12 Faizal 26 M Engineer Degree B 61113 61522 409
*Income Range, in RM: 4908
13 Shawal 42 M Executive Diploma C 139012 139396 384 4608
A:<2,000 14 Eyda 24 F Entrepreneur Degree D 32326 32673 347 4164
B:2,001-3000, 15 Yusmira 31 F Admin Certificate A 32904 33169 265
3180
C:3,001-4000, 16 Imi 36 F Engineer Degree E 809 1492 683 8196
Operations
19 Imran 36 M Diploma C 38497 39175 678 8136
Executive
20 Hafiz 41 M Technician Diploma B 166824 167771 947 11364
MOTORCYCLE
1 Dailami 27 M Engineer Diploma C 345270 347115 1845 22140
Secondary
7 Yati 36 F Petrolling A 4276 4600 324 3888
school
Secondary
8 Hidayat 38 M Petrolling A 27001 28007 1006 12072
school
*Income Range, in RM: 9 Saiful 43 M Petrolling Certificate B 27838 28877 1039 12468
A:<2,000 10 Shahrina 39 F Petrolling
Secondary
A 47032 48044 1012 12144
school
B:2,001-3000, 11 Adel 20 M Student Matric A 66896 66999 103 1236
C:3,001-4000, 12 Rezani 29 M Worker SPM A 25362 26262 900 10800
D4,001-5000, 13
14
Qusais
Khaldon
19
31
M
M
Student
Worker
Matric
SPM
A
A
4276
27001
4600
28007
324 3888
1006 12072
E:>5,001 15 Zahran 32 M Worker SPM A 27838 28877 1039 12468
Secondary
16 Azrul 31 M Petrolling A 20821 22128 1307 15684
school
Secondary
17 Badrul 31 M Petrolling A 43558 43989 431 5172
school
18 BLB9968 34 M Business Diploma B 2867 3215 348 4176
19 KCQ629 32 M Business Diploma B 23993 24396 403 4836
20 BNA2683 40 M Worker Diploma B 15890 16385 495 5940
21 BMY1270 44 M Worker Diploma B 8114 8184 70 840
22 JMS5949 36 F Education SPM B 51553 51613 60 720
DISTRIBUTION OF VKT
DESCRIPTIVE ANALYSIS
(SPSS & EXCELL)
Descriptive Analysis for Motorcar & Motorcycle
Histogram & Normal Distribution Curve
This bar represent the
small amount Annual
VKT (Car & Motorcycle)
frequency
The normal distribution is
positive skewed where
the Mean and Median
value is at right side
(positive) of Mode
Descriptive Analysis for Motorcar
Histogram & Normal Distribution Curve
This bar represent the
small amount Annual
VKT (Car) frequency
The normal distribution is
positive skewed where
the Mean and Median
value is at right side
(positive) of Mode
Descriptive Analysis for Motorcycle
Histogram & Normal Distribution Curve
This bar represent the
small amount Annual VKT
(Motorcycle) frequency
The normal distribution is
positive skewed
HYPOTHESIS TEST & MLR
INFERENTIAL ANALYSIS
(SPSS & EXCELL)
HYPOTHESIS TEST 1
TEST HYPOTHESES FOR THE DIFFERENCE BETWEEN
TWO POPULATIONS MEAN, WITH KNOWN VARIANCES
The two populations to be tested are:
1. VKT for Cars, VKT for Motorcycles,
µcars = 6039.82 and σcars = 3458.83
µmotorcycle = 11209.64 and σmotorcycle = 3888.00
H0 : The mean of the VKT for cars and VKT for motorcycles is similar at
significance level, α = 0.05
H0: µa - µb = 0 (claim)
And,
H1: The mean of the VKT for cars and VKT for motorcycles is different at
significance level, α = 0.05
H1: µa - µb ≠ 0
HYPOTHESIS TEST 1 CONT…
Compare z with α = 0.05
Critical Region: Z0.025 > 1.96 and Z0.025 <- 1.96 0.025 0.025
Calculate the sample test statistic 0
-1.96 1.96
Ztest is in the
rejection
region so
Reject H0
Conclusion: Reject H0,
The evidence indicates that the mean of VKT car and VKT
motorcycle is significant different at significance level α = 0.05
(95% confidence level)
HYPOTHESIS TEST 2
Hypothesis Test 2 (Using Excel Data Analysis Tools)
– To Compare Annual VKT Between Household Income
Ho : There is no significant difference in the average of annual VKT between
Income < RM 5000 and Income > RM 5000.
Ho : µ1 - µ2 = 0
H1 : There is a significant difference in the average of annual VKT between
Income < RM 5000 and Income > RM 5000.
H1 : µ1 - µ2 ≠ 0
HYPOTHESIS TEST 2 CONT..
Population normally distributed; large sample size > 30; population
variance unknown; population variance equal
Assume level of significance = 0.05
Ho : µ1 - µ2 = 0, H1 : µ1 - µ2 ≠ 0
Degree of Freedom, df = 44 – 2 = 42
Critical Region :
t < t0.025, 42 = (2.009 + 2.021)/2 = -2.015
t > t0.025, 42 = (2.009 + 2.021)/2 = 2.015
Annual VKT for Income
No Annual VKT for Income > RM5000
HYPOTHESIS
>RM5000
1 4020 4284
2 1272 11160
3 8832 5520
TEST 2 CONT.. 4
5
6
5484
9996
3180
4584
2232
8196
7 4908 1404
8 4608 5088
9 4164
10 3180
11 8136
12 11364
13 14460
Data sorted as Table 14
15
6804
22140
1 before analyze 16 14688
17 37980
18 13812
19 22800
20 16788
21 3888
22 12072
23 12468
24 12144
25 1236
26 10800
27 3888
28 12072
29 12468
30 15684
31 5172
32 4176
33 4836
34 5940
35 840
36 720
From Excel
HYPOTHESIS
TEST 2 CONT..
t0.025,42 t0.025,42
Conclusion :
t statistic = 1.479 < t0.025, 42 = 2.015
Fail to reject hypothesis null. Therefore there is no significant difference in
the average of annual VKT between Income < RM5000 and > RM5000
Gender Income range (Income
Age (FEMALE = 0, <RM4001 =0, Income Annual VKT
MULTIPLE LINEAR 34
MALE = 1)
0
>RM4001 = 1)
1 4284
REGRESSION 38
35
58
1
1
0
1
0
1
11160
4020
5520
29 0 0 1272
28 1 0 8832
26 0 0 5484
28 1 1 9996
44 1 1 4584
Dependent Variable 31 1 0 3180
35 1 1 2232
= Annual VKT
26 1 0 4908
42 1 0 4608
24 0 1 4164
Independent Variable 31 0 0 3180
= Age, Gender and Income 36 0 1 8196
34 0 1 1404
45 0 1 5088
36 1 0 8136
41 1 0 11364
27 0 0 14460
35 1 0 6804
27 1 0 22140
30 1 1 14688
32 1 0 37980
35 1 1 13812
40 1 0 22800
40 1 0 16788
36 0 0 3888
38 1 0 12072
43 1 0 12468
39 0 0 12144
RESULT OF MULTIPLE LINEAR REGRESSION
R2 not close to 1 indicate bad fit
R2 = 0.094 means that only 9.4% of annual
VKT can be explained by gender, age and
income
MLR :
ŷ = 4588.176 + 35.452x1 + 4396.937x2 – 995.518x3
Adjusted R2
P-value
Independent
Variables
Age
Gender
Income
β p value
MULTIPLE LINEAR
REGRESSION CONT..
Regression Line:
Where: Y = Annual VKT; X1 = Age; X2 = Gender; X3 = Income
1. When X1, X2 and X3 are zero, the estimated values of Annual VKT
is 4588.176.
2. When X2 and X3 are held constant, the estimated values of Annual
VKT increase by 35.452 for every 1 unit increase in age.
3. When X1 and X3 are held constant, the estimated values of Annual
VKT increase by 4396.937 for every 1 unit increase in gender.
4. When X1 and X2 are held constant, the estimated values of Annual
VKT decrease by 995.518 for every 1 unit increase in income.
MULTIPLE LINEAR
REGRESSION CONT..
Conclusion:
1) F (1.385) > Significance F (0.261). Therefore the result
is not so reliable (slightly significant)
2) P Value for age, gender and income is more than 0.05
which it is not statistically significant. Therefore this
model can be replace with other independent variable
to get more effective value of Annual VKT