CHAPTER 4

MULTIVARIATE ANALYSIS
 CONTENT

1. Reliability testing of scales

2. Exploratory Factor Analysis (EFA)
3. Exercises

2
1. RELIABILITY TESTING OF SCALES

1.1. Scales and reliability of scales

1.2. Reliability testing of scales using Cronbach’s alpha
coefficient
1.3. Applying SPSS software to calculate the reliability
of scales

3
1.1. Scales and reliability of scales
❖ A scale is a set of observed variables with specified
attributes that together measure a certain concept (Nguyen
Dinh Tho, 2011).

Concept
Concept

Component Component

Item Item Item Item

Item Item Item Item Item

4
For example: Ponzi & et al (2011) developed a scale to
measure Corporate reputation with four observed
variables.
Level of agreement
Name Label
1 2 3 4 5
Is a company I have a good feeling
CREP1
about
CREP2 Is a company that I trust
Is a company that I admire and
CREP3
respect
CREP4 Has a good overall reputation

5
❖ Reliability reflects the consistency or trustworthiness of
a measurement. In other words, when we apply the same
method to the same sample under the same conditions,
you will receive the same results.
❖ Different statistical methods determine different types of
reliability, such as:
▪ Test-retest reliability
▪ Inter-rater reliability
▪ Internal consistency reliability

6
1.2. Reliability testing of scales using Cronbach’s
alpha coefficient

❖ Testing the reliability of a scale using Cronbach’s alpha

is a method for assessing internal consistency reliability.
❖ According to Nunnally (1978) and Peterson (1994), a
scale is considered acceptable and good if it meets both
of the following conditions:
▪ Cronbach’s Alpha coefficient > 0.6
▪ Corrected Item-Total Correlation > 0.3

7
7.Note:
MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ
▪ Apply the calculation of Cronbach’s alpha coefficient for
quantitative scales;
▪ Do not apply the calculation of Cronbach’s alpha coefficient for
scales with only one observed variable/item;
▪ When analyzing Cronbach’s alpha, only include the observed
variables/items within the same scale to assess reliability;
▪ A high Cronbach’s alpha coefficient (> 0.6) indicates a strong
correlation between the observed variables/items within the
same scale. However, this does not mean that the higher the
coefficient, the better, because if the Cronbach’s alpha
coefficient approaches 1, it is likely that the content of the
observed variables/items is overlapping.

8
1.3 Applying SPSS software to calculate the reliability
of scales

Application Model: Factors affecting the attraction of foreign

direct investment (FDI) in industrial zones.

Theory: According to Dunning (1977); Paul Krugman (1991);

World bank (2004); Romer & Lucas (2007).

9
Research hypothesis:

H1: Infrastructure has a positive impact on the satisfaction

of FDI investors
H2: Investment policy have a positive impact on the
satisfaction of FDI investors.
H3: Living and working environment has a positive impact
on the satisfaction of FDI investors
H4: Investment advantage has a positive impact on the
satisfaction of FDI investors.
H5: The quality of public services has a positive impact on
the satisfaction of FDI investors.
H6: Local brand has a positive impact on the satisfaction of
FDI investors

10
Research hypothesis:

H7: Human resources have a positive impact on the

satisfaction of FDI investors.
H8: Competitive input costs have a positive impact on the
satisfaction of FDI investors

11
 The quality
Infrastructure of public
services

H1 (+) H5(+)
Investment Local brand
policy H2(+) H6(+)
The
satisfaction
H3(+) of FDI H7(+)
Living and investors
Human
working
resources
environment
H4(+) H8(+)

Investment Competitive
advantage input costs
Research model
12
Measurement scales in the research model:

Scale Name
1. Infrastructure (CSHT)

The electricity supply system meets the requirements. CSHT1

The water supply and drainage system is adequate. CSHT2

Convenient communication (telephone, internet, etc.) CSHT3

Convenient transportation (time, cost) CSHT4

The premises meet the requirements CSHT5

The banking system meets the requirements CSHT6

13
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ
2. Investment policy (CSDT)
Local leaders are dynamic in supporting businesses CSDT1
Legal documents are quickly disseminated to CSDT2
businesses
There are attractive investment incentives CSDT3
Businesses will still invest even if the local area CSDT4
does not have attractive policies
There is a clear tax system (tax officials do not CSDT5
exploit it for personal gain)

14
3. Living and working environment (MTS)
School system meets the needs MTS1
Healthcare system meets the needs MTS2
Environment is not polluted MTS3
Attractive recreational spots MTS4
Friendly residents MTS5
Reasonable living costs MTS6
Disputes between workers and businesses are MTS7
resolved

15
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ
4. Investment advantage (LTDT)
Convenient access to main production materials LTDT1
Convenient access to main consumption markets LTDT2
Close to partner businesses (key distributors or LTDT3
suppliers)
Market competition with main competitors LTDT4

5. The quality of public services (DVC)

Simple and quick administrative procedures DVC1
Local government provides attentive support when DVC2
businesses need it
Streamlined customs procedures DVC3
Investment promotion and trade centers provide DVC4
good support for businesses

16
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ
6. Local Brand (THDP)
The local area is an impressive brand. THDP1
I think this place is becoming a destination for THDP2
investors.
I believe many people have successfully invested THDP3
here, and I want to be like them.
I invest here simply because I want to invest in this THDP4
locality.

17
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ
7. Human Resources (NNL)
Vocational schools meet the requirements of NNL1
enterprises.
There is an abundant supply of unskilled labor. NNL2
The workforce has high discipline. NNL3
The ability to absorb and apply technology is good NNL4
among the labor force.
The company does not face language barriers. NNL5
It is easy to recruit skilled managers locally NNL6

18
7. MÔ HÌNHInput
8. Competitive PHÂN TÍCH
Costs NHÂN TỐ KHÁM PHÁ
(CPCT)
Low land rental prices CPCT1
Cheap labor costs CPCT2
Reasonable electricity, water, and transportation CPCT3
costs
Competitive communication service prices CPCT4
The satisfaction of FDI investors (SAT)
The company's revenue will grow as expected. SAT1
The company's profits have/will meet expectations. SAT2
Our company will continue to invest in long-term SAT3
business in the area.
I will recommend this area to other companies. SAT4
Our company is very satisfied with the investment in SAT5
this location.

19
Open SPSS Data file: P1-DAUTUFDI
Analyze > Scale > Reliability analysis

Select the
observed
variables and
place them in the
Items box

Select Statistics

20
 Select the
functions of
statistics

Continue > OK

21
 The result of the first analysis

Cronbach’s Alpha = 0,651 > 0,6

The Corrected Item - Total correlation coefficient of

CSHT5 = 0,136 < 0,3 → Remove variable CSHT5.
22
 The result of the second analysis

Hệ số Alpha = 0,707 > 0,6

The Corrected Item - Total correlation coefficient of

CSHT4 = -0,077<0,3 → Remove variable CSHT4.
23
The result of the third analysis

Cronbach’s Alpha = 0,861 > 0,6

Corrected
Item - Total
correlation >
0,3

The scale achieved reliability

24
Similarly, perform the analysis for the remaining variables

SUMMARY TABLE OF VARIABLES AND EXCLUDED

SCALES
Scale Excluded Items
CSHT CSHT4 CSHT5
CSDT CSDT4
MTS MTS1 MTS7
LTDT LTDT4
DVC Excluded scale
THDP
NNL NNL5 NNL6
CPCT
SAT
25
2. EXPLORATORY FACTOR ANALYSIS (EFA)

2.1. Exploratory factor analysis (EFA)

2.2. Apply SPSS software to perform Exploratory
Factor Analysis (EFA)

26
2.1. Exploratory factor analysis (EFA)

Exploratory Factor Analysis is used to reduce a set of

k observed variables into a set of F more meaningful
factors (F < k)

27
 X1 Convergent

X2
F1 Factor
X3

X4

Items X5 Discriminant

X6

X7 Factor
F2
X8

X9

28
2.2. Apply SPSS software to perform
Exploratory Factor Analysis (EFA)

Application Model: Factors affecting the attraction

of foreign direct investment (FDI) in Industrial
Zones.
Data: P1-DAUTUFDI

29
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ
Analyze > Dimension Reduction > Factor
Place the
variables that
have been
tested with
Cronbach’s
alpha into the
Variables box
(excluding the
dependent
variable SAT).

30
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ

Select Descriptives... Select KMO…

Continue

31
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ

Select Rotation... Chọn Varimax

Continue

32
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ

Select Options... Enter 0.5

Continue
33
The results of the EFA analysis (The first time)

(1) Test the suitability of EFA

KMO ≥ 0.5 → Factor analysis is appropriate for the
actual data

(2) Test the correlation between observed

variables
Bartlett’ Test
If sig.<= 0,05 → The observed variables are
correlated within the same factor.

34
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ

(1) KMO = 0,743 > 0,5 → Factor analysis is

appropriate for the actual data
(2) Sig. = 0,00 < 0,05 → The observed variables
are correlated within the same factor

35
7.
(3)MÔ HÌNH variance
Extraction PHÂN TÍCH
testingNHÂN TỐ KHÁM
(% cumulative PHÁ
variance)

Acceptance
criterion:
Extraction
variance
>50%

❖ The number of factors extracted from 28 observed variables

is 8 factors.
❖ The 8 factors explain 73.2% of the variance of the observed
variables
36
(4) Factor loading

Remove
variable
CSDT3

Remove
variable
MTS4
The results of the EFA analysis (The second time)

Remove
variable
CSDT5

Remove
variable
NNL4
The results of the EFA analysis (The third time)

F1: CPCT1, CPCT2, CPCT3,

CPCT4
F2: CSHT1, CSHT2, CSHT3,
CSHT6
F3: NNL1, NNL2, NNL3

F4: MTS2, MTS3, MTS5,

MTS6
F5: LTDT1, LTDT2, LTDT3

F6: THDP1, THDP2, THDP3,

THDP4
F7: CSDT1, CSDT2
7.
(4)MÔ HÌNHthe
Rename PHÂN TÍCH NHÂN TỐ KHÁM PHÁ
factors
F1: CPCT1, CPCT2, CPCT3, Competitive Input Costs
CPCT4 (CPCT)
F2: CSHT1, CSHT2, CSHT3,
Infrastructure (CSHT)
CSHT6
F3: NNL1, NNL2, NNL3 Human Resources (NNL)

F4: MTS2, MTS3, MTS5, Living and working

MTS6 environment (MTS)
F5: LTDT1, LTDT2, LTDT3 Investment advantage (LTDT)
F6: THDP1, THDP2, THDP3, Local Brand (THDP)
THDP4
F7: CSDT1, CSDT2 Investment policy (CSDT)

40
7. MÔ
For theHÌNH PHÂN variable
dependent TÍCH NHÂN TỐ
(SAT), doKHÁM PHÁ
the same

Analyze > Dimension Reduction > Factor

Add the 5
observed variables
into the Variables
frame... and
perform the 4-step
test as with the
independent
variables

41
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ

KMO=0,767 > 0,5 => Factor analysis is appropriate

for the actual data
Sig.=0,00<0,05 => The observed variables are
correlated within the same factor

42
7. MÔ HÌNH PHÂN TÍCH NHÂN TỐ KHÁM PHÁ

❖The number of factors extracted from 5 observed

variables is 1 factors.
❖The 1 factors explain 61,2% of the variance of the
observed variables

43
SAT: SAT1, SAT2, SAT3,
SAT4, SAT5
7.
(5)MÔ HÌNH PHÂN
Determine TÍCH for
the scores NHÂN TỐ KHÁM PHÁ
the factors
Transform > Compute variable

▪ CPCT = Mean(CPCT1,CPCT2,CPCT3,CPCT4)
▪ CSHT = Mean(CSHT1, CSHT2, CSHT3, CSHT6)
▪ NNL = Mean(NNL1, NNL2, NNL3)
▪ MTS = Mean(MTS2,MTS3,MTS5,MTS6)
▪ LTDT = Mean(LTDT1,LTDT2,LTDT3,)
▪ THDP = Mean(THDP1,THDP2,THDP3,THDP4)
▪ CSDT = Mean(CSDT1,CSDT2)
▪ SAT = Mean(SAT1,SAT2,SAT3,SAT4,SAT5)

45
3. EXERCISES

Study the satisfaction of foreign direct

investment (FDI) enterprises with the quality of
tax support services at the tax department of
Province X.
Data: BAI TAP 2-HAILONGTHUE, Data from a
survey of 222 FDI enterprises.

46
 The research model

Reliability
H1(+)

Responsiveness
H2(+)

Satisfaction
H3(+)
Empathy with service
quality
H4(+)
Service
capability
H5(+)

Tangibles
47
Requirements:
1. Perform the Cronbach's Alpha test to determine
the reliability of the scales.
2. Conduct the KMO and Bartlett's tests and
check the extracted variance to identify suitable
scales for Exploratory Factor Analysis (EFA).
3. Identify the appropriate measurement scale
system.