Các nhân tố ảnh hưởng đến ý định phàn nàn của khách hàng sử dụng dịch vụ điện thoại di động khu vực đồng bằng sông cửu long tt tiếng anh

MINISTRY OF EDUCATION AND TRANING
CANTHO UNIVERSITY

SUMMARY OF DOCTORAL THESIS
In Business Administration
CODE: 62340102

NGUYEN THANH LIEM
FACTORS INFLUENCE ON THE COMPLAINT INTENTION OF
THE CUSTOMER USING TELECOM SERVICE
IN MEKONG DELTA

Cần Thơ, năm 2020
1


Reviewer 1: .....................................................................................................
Reviewer 2: .....................................................................................................

2


LIST OF PUBLICED PAPERS
1. Nguyễn Thanh Liêm, Bùi Văn Trịnh (2017). Antecedents of complaint
intention: case study of mobile phone service in Can Tho. Economics
& Development Journal, 236: 61-71.
2. Nguyễn Thanh Liêm, Bùi Văn Trịnh (2017). Antecedents of complaint
intention: case study of mobile phone service in Mekong delta. Science
Journal of Can Tho University, 236: 61-71.

3

CONTENT.
Page
CHAPTER 1: INTRODUCTION.............................................................. 6
1.1 BACKGROUND ................................................................................. 6
1.2 RESEARCH GOAL ............................................................................ 6
1.2.1 Research goal................................................................................. 6
1.2.2 Research objectives ....................................................................... 6
1.3 RESEARCH QUESTIONS ................................................................. 7
1.4 RESEARCH OBJECT AND SCOPE ................................................. 7
CHAPTER 2: LITERATURE REVIEW AND RESEARCH MODEL 7
2.1 BACKGROUND ................................................................................. 7
2.2 COMPLAINT AND COMPLAINT INTENTION ............................. 8
2.3 THEORIES AND RESEARCH HYPOTHESICS .............................. 8
2.3.1 Justice Theory ................................................................................ 8
2.3.2 Trust ............................................................................................... 9
2.3.3 Customer satisfation and complaint intention ............................... 9
2.3.4 Moderator - Perceived responsiveness ........................................ 10
2.4 RESEARCH MODEL ....................................................................... 10
CHAPTER 3: RESEARCH METHODOLOGY ................................... 10
3.1 RESEARCH STRATEGY ................................................................ 10
3.2 RESEARCH DESIGN ....................................................................... 11
3.3 MEASUREMENT SCALE AND QUESTIONNAIRE DESIGN .... 11
3.4 SAMPLING ....................................................................................... 12
3.5 DATA ANALYSIS METHODS ....................................................... 13
CHAPTER 4: RESULT AND DISCUSSION ........................................ 13
4

4.1 DESCRIPTIVE STATISTICS .......................................................... 13
4.2 RELIABILITY ANALYSIS.............................................................. 14
4.3 EXPLORATORY FACTOR ANALYSIS (EFA) ............................. 14
4.4 FINAL CONSTRUCTS BASED ON EFA RESULTS .................... 16
4.5 CORRELATION ANALYSIS .......................................................... 18
4.6 CONFIRMATORY FACTOR ANALYSIS - CFA .......................... 18
4.7 STRUCTURAL EQUATION MODEL - SEM ................................ 19
4.8 BOOSTRAP ANALYSIS.................................................................. 21
4.9 MODERATOR ANALYSIS – PERCEIVED RESPONSIVENESS 21
4.10 CONTROL VARIABLE - GENDER.............................................. 22
CHAPTER 5: CONCLUSION AND IMPLICATION .......................... 23

5


CHAPTER 1
INTRODUCTION
1.1 BACKGROUND
Complaints have an important impact on customer retention and profitability
(Holloway, Wang & Parish, 2005). Hart et al. (1990) stated: there is a clear
difference between companies that well manage complaints (of customer) and
ones that not. Successful companies often encourage customers to complain
through “positive cooperation” policies and proactive actions, while others
often apply passive approaches in complaints management (Firnstahl, 1989).
The result of literature review leads to the perception that there is a few research
in Vietnam focused on customer complaint, and there is still lack of positive
view on complaining behavior. Breazeale (2009) reveals that the cost of
persuading a new customer is five times higher than maintaining an existing
customer. Especially in the field of mobile phone services, when competition
between companies is becoming increasingly fierce, maintaining existing loyal

customers is critical. Therefore, good complaints management can help the
company attain an opportunity to continue serving customers and improve the
quality of business operations in the future. Based on such arguments, the
proposed topic “Factors influence of the complaint intention of the customer
using telephone service in Mekong delta” is essential and needs to be
conducted.
1.2 RESEARCH GOAL
1.2.1 Research goal
The goal of this study is to find out the factors affecting complaint intention
of customers who using telecom service in Mekong Delta. The research result
is expected to provide academic evidences that contribute to a better
understanding of customer complaint behaviour and suggest practical
implications for telecom service providers.
1.2.2 Research objectives
In order to reach the reseach goal, this study needs to achieve three objectives
as follows: (1) Systematizing literature related to the topic of customer
complaints; (2) Identify and measure the influence of atecedents of compaint
intention of customers who using telecom service in Mekong Delta region; (3)
Proposing management implications to motivate consumers to complain
directly to telecom service providers in the Mekong Delta region.

6


1.3 RESEARCH QUESTIONS
The research questions for the study is proposed as follow: (1) What are the
antecedents to complaint intention of customer who using telecom services in
Mekong Delta?; (2) How much does each antecedent impact the level of
complaint intention of customer who using telecom services in Mekong Delta?;
(3) Which policies should be applied to motivavte direct complain?

1.4 RESEARCH OBJECT AND SCOPE
Research object: The study focuses on complaint intention of customer
who using telecom service in Mekong Delta. Therefore, the research object is
the customer who using mobile phone services in the Mekong Delta region
and had experienced incidents related to mobile phone services.
Geographic scope: Based on the data of telecom service providers and
current population, research space is proposed as provinces and city including:
Can Tho, Long An, Tien Giang and Kien Giang. These are the 4 largest markets,
accounting for 51% of the telecom service market share in Mekong Delta.
Time scope: Data was collected through a cross-sectional survey from May
2016 to February 2017. Data analysis was conducted after that time.

CHAPTER 2
LITERATURE REVIEW & RESEARCH MODEL
2.1 BACKGROUND
Garding & Bruns (2015) conducted a review and made the judgment that the
research and theory related to complaints is very diverse, and therefore the
researcher should focus on a specific direction to work. Based on a review of
more than twenty studies published between 1977 and 2014, Garding & Bruns
(2015) proposed that five dominant factors are considered to affect customer
satisfaction and thereby impacting complaint intention, including: distributive
justice, procedural justice, interactional justice, attitude, and social interaction.
The result of literature review also reveals that Justice Theory is the most
used theory and can be considered good enough to be used as a foundation
theory. In Vietnam, the research related to the topic of complaint behaviour is
not much and no research has evaluated the role of Justice Theory in the field
of services. Therefore, this study is proposed a research model based on Justice
Theory, combined with other variables to increase the explanatory level of the
model.

7


2.2 COMPLAINT AND COMPLAINT INTENTION
Complaint behavior is defined as an action taken by an individual regarding
the communication of negative things about a product/ service/ business with
the third parties (Jacoby & Jaccard, 1981). By 1990s, Singh & Widing (1991)
suggest that customer complaint behaviour is a “set of activities” that is
triggered by customer dissatisfation. McQuilken & Robertson (2011) defines
complaining behavior as a multi-feedback set, that originating from
dissatisfaction.
Intentions are assumed to capture the motivational factors that influence a
behavior; they are indications of how hard people are willing to try, of how
much of an effort the are planning to exert, in order to perform the behavior
(Ajzen, 1991). Many studies have also confirmed intention as a motivational
factor, which motivates an individual to be willing to take action. Combining
the concept of complaint behaviour and intention, in this study, the author
focuses on the concept: the level of effort that customer is planning to exert, in
order to perform the complaint behavior with telecom service providers in
Mekong Delta.
2.3 THEORIES AND RESEARCH HYPOTHESICS
2.3.1 Justice Theory
Colquitt et al. (2001) comprehensively reviewed 183 justice-related studies
from the literature and integrated them into three major justice constructs:
Distributive justice (DISJUS): refers to the perceived fairness when
individuals assess the fairness of an exchange by comparing their inputs to
outcomes to form an equity score (Martinez-Tur et al., 2006).
Procedural justice (PROJUS): Leventhal (1980) first discussed procedural
justice in non-legal contexts such as organisational settings and pointed out six
evaluation criteria that perceived procedural fairness should meet. Procedures

should (1) be consistent: applying procedure consistently across individuals and
time, (2) suppress bias: omitting the self-interest of the decision maker, (3) be
accurate: ensuring that accurate information is collected and used in making
decisions, (4) be corrective: having some mechanisms to correct wrong
decisions, (5) be representative: ensuring that the opinions of all parties affected
by the decision have been taken into account, and (6) be ethical: meeting the
ethical and moral values of the social system.
Interactional justice (INJUS): Bies & Moag (1986) separated out the
interpersonal aspect of procedural justice that is termed interactional justice.
There are four criteria to assess interactional justice: (1) justification for
decisions, (2) truthfulness, (3) respect, and (4) propriety.
8


Much research on service recovery has generally identified consistent
relationships between justice perception and customer satisfaction (Ruyter &
Wetzels, 2000; Holloway et al., 2005). However, there are differences in the
level of impact in different fields as well geographic areas. Therefore, it is
necessary to identify the relationship between 3 justice constructs, and
determine the level of impact of each component on customer’s satisfaction in
Mekong Delta. Research hypotheses are proposed:
H1. Distributive justice has a positive effect on consumer satisfaction, in
telecom service in Mekong Delta.
H2. Procedural justice has a positive effect on consumer satisfaction, in
telecom service in Mekong Delta.
H3. Interactional justice has a positive effect on consumer satisfaction, in
telecom service in Mekong Delta.
2.3.2 Trust
Pavlou & Fygenson (2006) argue that trust is the belief of the buyer that the
seller will behave ethically. Some scholars have carried out a review in social

exchange context, and conclude that trust has an effect on satisfaction (Singh &
Sirdeshmukh, 2000; Chiou, 2004). Studies in e-commerce also suggest trust
should be considered as an important aspect in technology-related transactions
(Gefen, 2000; Reichheld & Schefter, 2000). Reichheld & Schefter (2000)
provided evidence that customers often stay away from service providers whom
they don't trust. Further studies also show that trust is a critical determinant of
customer satisfaction, in a mobile commerce context (Lin & Wang, 2006). The
same argument can also be applied in telecom service context. Hence, it is
hypothesized that:
H4. Trust has a positive effect on consumer satisfaction, in telecom service
in Mekong Delta.
2.3.3 Customer satisfation and complaint intention
Voorhees and Brady (2005) proposed a model for exploring initial drivers of
complaint intentions from the persepctive of social justice in service encounter
when the effects of jusitce drivers on complaint intentions may be mediated by
customer satisfaction. This indicates a fact that customer satisfaction derived
from a failed service encounter may have a negative influence on complaint
intentions. Cronin et al. (2000) claimed the effects of service quality on
consumer’s behavioural intentions may be mediated by customer satisfaction.
The rationale for the argument lies in the conceptualisation of service quality
and customer satisfaction in terms of cognitive and affective evaluations of
service experience that further prompt conative response for compaint
9


intentions. That is, when consumers feel more dissatisfied with their services,
complaint intentions therefore increase. Hence, it is hypothesised
H5. Customer satisfaction negatively affects complaint intentions, in telecom
service in Mekong Delta.
2.3.4 Moderator - Perceived responsiveness

Perceived responsiveness is defined as the consumer’s perception of an evendor’s willingness to address a service failure and to provide a complaint
handling mechanism (Tax, Brown, & Chandrashekaran, 1998). A positive
relationship between customer satisfaction and further complaint intentions, in
traditional shopping, has been proposed (Voorhees & Brady, 2005). That is, as
consumers’ perception of responsiveness increases, they become more likely to
complain to vendors (Richins, 1987). Recent research by Voorhees & Brady
(2005) and Wu (2016) reveal perceived responsiveness as a moderator in the
relationship between satisfaction and complaint intention. Hence, it is
hypothesized that:
H6. The higher the perceived responsiveness, the stronger is the link between
customer satisfaction and complaint intentions, in online shopping.
2.4 RESEARCH MODEL
A research model is proposed to determine the antecedents of the intention
to complain among consumers who using telecom service in Mekong Delta.

Figure 2.1: Proposed research model

CHAPTER 3
RESEARCH METHODOLOGY
3.1 RESEARCH STRATEGY
In a study from Ritchie and O’Connor (2004), they mentions about two kinds
of methodologies which can be applied in academic study to gain knowledge in
social contexts. “These methods are qualitative and quantitative research. “In
this study, in order to solve the research questions that have been raised in
Chapter 1, the author has built a research model which was improved based on
the Justice theory.” Because of the backgrounds amd hypotheses were proposed
10


in Chapter 2, according to criteria by Neuman (2003), this study could be

considered as a quantitative research. Howeverm qualitative method was also
applied in the process of designing measurement scales and questionnaire.
3.2 RESEARCH DESIGN
A research design provides the basic plan to guide the research process in
order to achieve research objectives most efficiently (Hair et al. 2003). In this
study, research design can be divided into three processes: exploratory, pilot,
and main study.

Figure 3.1: Research design
3.3 MEASUREMENT SCALE AND QUESTIONNAIRE DESIGN
Measurement scales are adapted based on previous published studies. The
author also consulted experts to find and adjust points that could lead to
misunderstanding or bias.
The process of determining measurement scale and designing questionnaitre
is described as follows:
- Step 1: English original scales were summarized based on articles
published on journals of ISI or Scopus system. The measurement items
for justice constructs, including distributive justice, procedural justice
and interactional justice, were adapted from the measurements
developed by Martinez – Tur et al. (2006), Turel et al. (2008), Chiu et al.
(2009), and Wu (2013). Trust is measured with items adapted from Wu
(2013), and Turel et al.(2008) ; Satisfaction, complaint intention, and
perceived responsiveness are measured using items based on Liu & Wu
(2013) and Voorhees & Brady (2005).

11


- Step 2: Consult ideas from scholars of Can Tho University to short-list
the potential scales which could be suitable inthe context of telecom

service.
- Step 3: Translate the original scales into Vietnamese with the
consultation of a linguistic expert to control the misunderstanding of
vocabulary and culture.
- Step 4: A group interview with managers working in telecom service was
organized to discuss and finalize the questionnaire.
As the final step, a pilot study with a sample of 150 observations in Can Tho
city was also conducted to check the reliability and validity of the questionnaire
and help the author recorded some possible problems may happen during field
survey process.
3.4 SAMPLING
Sampling proccess usually involves three steps: determining the targeted
population, identify sampling method, and calculate sample size. Since the
study focused on considering complaint intention of customers who using
telecom services in Mekong Delta, the target population is Mekong Delta
consumers who using telecom services and already registered. However, due to
limited resources such as finance, time, and human force, the survey was
conducted in four provinces and cities: Can Tho, Long An and Tien Giang. ,
and Kien Giang. These are 04 localities with large population size and account
for the largest number of telecom services’ customers in the Mekong Delta area.
Non-probability sampling was chosen to collect the primary data for this
study, and it is convenient method where the researchers use any readily
available individuals as participants (Cooper & Chindler, 2006). The most
adventage of this method is it allows the author to achieve a large number of
samples with an optimum resources. However, the author clearly awares that
the convenient sampling method has limitations which will be discussed in t the
final chapter of this study.
“Kline (2011) said that a usual sample size where Structural Equation Model
used is about 200 cases. Barret (2007) also recommended that reviewers of
journal submissions will reject for publication if Structural equation modeling

has a number of sample size lower than 200.” In this situation, Structural
Equation Model is the major technique to analyze data and calculate the results.
Therefore, the sample size should be at least 200. Based on the ability to collect
data, the selection of 200 respondents in this study as the minimum sample size
can satisfy the requirement for factor analysis. This study had approached and
collected about 1,000 samples. Among those, there were 46 invalid samples,
12


accounting for 4.6%. Finally, the final number of observation is inputed and
used in analysis: n = 954 observations.
3.5 DATA ANALYSIS METHODS
In order to reach the most precise results for the research, questionnaires were
coded, computerized, and screened for errors before any statistical analysis.
Finally, the author would apply several statistic techniques using SPSS and
AMOS:
For SPSS: Frequency; Mean; Standard Deviation; Reliability test with
Cronbach’s alpha; Exploratory Factor Analysis (EFA).
For AMOS: Confirmatory Factor Analysis (CFA); Structural Equation
Model (SEM); Multigroup Structural Equation Modeling; Two ways
interaction analysis.

CHAPTER 4
RESULT & DISCUSSION
4.1 DESCRIPTIVE STATISTICS
Regarding gender, men accounting for 50.9%, and female accounting for
49.1%. Thus, the rate of male and female appearance in the sample is quite
balance.
In terms of age, respondents participated in the survey mainly aged 31-40
years old, and 21-30 years old with the proportion of 31.4% and 30.8%

respectively. These two age groups account for 61% of the sample. Next is the
customer group at 41-50 years old, accounting for 19.5%; the group of
respondents under the age of 30 accounts for 132%, the rest are over 50 years
old, accounting for only 5%.
In terms of education level, more than 50% of people in the survey sample
have University and College qualifications; of which 32.1% of respondents
have a college degree and 21.0% have a university degree. The number of
people with education at the universal level accounts for the majority of the
sample with about 40%. Only 5% of respondents have postgraduate
qualifications.
In terms of occupation, out of 954 respondents, civil servants, office workers
and business people had the highest rate of 26.2% respectively; 24.5%, 24.1%.
The interviewees are students who also account for a large proportion of the
sample with about 21%, while other types of occupations of respondents are
only about 4%.
13


Regarding the resident locality, the proportion of sample groups from Can
Tho, Kien Giang, Long An and Tien Giang is 32%, 16.9%, 30.4%, and 20.8%.
In terms of forms of payment, prepaid subscribers are the type most
commonly used by customers in the survey sample with the rate appearing in
the sample up to 86.4%. Meanwhile, only 13.6% of respondents said they used
postpaid subscribers. The results of this survey are consistent with the reality of
the difficulties in developing postpaid services of mobile service providers in
Vietnam.
4.2 RELIABILITY ANALYSIS
The most common method for evaluating the reliability of measurement
scale is based on Cronbach’s alpha index. The results of reliability analysis
show that most of the scales proposed for the study had met the statistical

requirements (Cronbach’s alpha> 0.7). This means that the measurement scales
are reliable enough according to statistical rules and to be able to conduct
further analysis steps.
4.3 EXPLORATORY FACTOR ANALYSIS (EFA)
The purpose of the EFA analysis step is to minimize the number of
explanatory variables, examine structures or relationships between variables,
contributing to reinforcing or rejecting the initial research hypotheses. In this
study, the author chose to perform EFA analysis using the extracting method
Principal Axis Factoring and Promax rotation based on the recommend of
Gerbring & Anderson (1988). These scolars stated that these methods will
reflect the data structure more accurate than the method of Principal
Components and Varimax (orthogonal).
The exploratory factor analysis was carried out for variable groups in turn:
(i) independent variable groups affecting customer satisfaction using mobile
phone services include: distributive justice, procedural justice, interactive
justice, and trust; (ii) perceived responsiveness; (iii) satisfaction; and (iv)
complaint intention. The only thing worth noting is that proposing the
combination of interactive justice and procedural justice.
Specifically, the results of exploratory factor analysis for independent
variables include: equity of distribution, fairness of procedures, fairness of
interaction, and belief for KMO coefficient of 0.929; and meaning Sig is 0.000;
There are 3 factors extracted in Eigenvalues value of 1,517 higher than the
minimum requirement (> 1); these observed variables also explain 62.907% of

14


the total observations higher than the minimum requirement (50%); moreover,
all factor loadings are greater than 0.5.
Bảng 4.1: Pattern Matrix of justice components and trust

Pattern Matrixa
Yếu tố
1
DISJUS01
DISJUS02
DISJUS03
DISJUS04
DISJUS05
PROJUS01
PROJUS02
PROJUS03
PROJUS04
INJUS01
INJUS02
INJUS03
INJUS04
INJUS05
TRUST01
TRUST02
TRUST03
TRUST04

2

3

0.677
0.789
0.801
0.865

0.858
0.539
0.577
0.491
0.561
0.742
0.765
0.920
0.861
0.873
0.786
0.837
0.842
0.908

Extraction method: Principal Axis Factoring.
Rotationl method: Promax with Kaiser Normalization.
The suggestion to combine these variables is unexpected but can be
explained by the origin of the interactive equity element which is separated
from the procedurl justice in the study published by Bies & Moag (1986). Based
on the results of the data analysis and the results of the review of literature, the
study proposed to group the procedural justice and interactive justice into a new
independent variable, still taking the name of procedural justice and denoted by
PROIN.

15


After grouping, the author had investigated Cronbach’s alpha coefficient for
this new construct. The result show that Cronbach’s alpha coefficient is 0.922,

which is greater than the Cronbach’s alpha level of each scale before.
4.4 FINAL CONSTRUCTS BASED ON EFA RESULTS
Based on the results of reliability analysis and exploratory factor analysis,
the final constructs in local language are presented as follows:
Distributive justice (DISJUS)
DISJUS01 Giá trị của các dịch vụ tôi nhận được tương xứng với mức giá
mà tôi đã trả cho các nhà mạng di động.
DISJUS02 Giá trị các dịch vụ mà tôi nhận được từ nhà mạng di động tỷ lệ
thuận với giá tôi đã trả.
DISJUS03 Chất lượng dịch vụ mạng phù hợp với giá của nó.
DISJUS04 Phí dịch vụ của nhà mạng di động này phù hợp với chất lượng.
DISJUS05 Nhìn chung, những dịch vụ được cung cấp bởi nhà mạng di
động này là hợp lý.
Procedural justice (PROIN)
PROJUS01 Những chính sách của nhà cung cấp dịch vụ di động được áp
dụng một cách công bằng đối với tất cả khách hàng.
PROJUS02 Những chính sách của nhà cung cấp dịch vụ di động được áp
dụng khá ổn định.
PROJUS03 Những chính sách của nhà cung cấp dịch vụ di động được áp
dụng dựa trên những thông tin chính xác.
PROJUS04 Nhà cung cấp dịch vụ di động sẽ làm rõ bất kỳ sự thay đổi nào
về chính sách và cung cấp thông tin khi khách hàng yêu cầu.
INJUS01
Tôi đã được đối xử công bằng trong suốt quá trình tương tác
với các nhân viên dịch vụ của nhà mạng.
INJUS02
Hệ thống hỗ trợ khách hàng của các nhà mạng cung cấp
những lời giải thích kịp thời và chính xác.
INJUS03
Hệ thống hỗ trợ khách hàng của các nhà mạng giải thích các

quy trình dễ hiểu.
INJUS04
Hệ thống hỗ trợ khách hàng của các nhà mạng hoạt động dễ
dàng và thú vị.
INJUS05
Hệ thống hỗ trợ khách hàng của các nhà mạng tương tác tốt
với tôi.

16


Trust (TRUST)
TRUST01 Dịch vụ của nhà mạng tôi đang dùng là đáng tin cậy.
TRUST02 Chính sách dịch vụ của nhà mạng tôi đang dùng là công
khai, minh bạch.
TRUST03 Các nhà mạng luôn thực hiện đúng cam kết của họ với khách
hàng.
TRUST04 Tôi tin vào những thông tin mà nhà mạng cung cấp cho tôi.
Perceived responsiveness (PERES)
PERES1
Khi có bất kỳ sự bất mãn của khách hàng xảy ra, tôi tin
tưởng rằng nhà mạng di động sẽ cố gắng khắc phục nó.
PERES2
Nhà cung cấp sẽ xử lý những khiếu nại của tôi để tôi tiếp tục
sử dụng dịch vụ của họ.
PERES3
Khi tôi chọn mua sim điện thoại, tôi đã chọn nhà mạng có
uy tín để đảm bảo chất lượng dịch vụ.
Satisfaction (SATIS)
SATIS01 Tôi rất thích sử dụng dịch vụ của nhà mạng di động tôi đang

dùng.
SATIS02 Tôi đã đúng đắn khi lựa chọn nhà cung cấp dịch vụ mạng
viễn thông này.
SATIS03 Tôi cảm thấy thoải mái với nhà cung cấp dịch vụ mạng viễn
thông này.
SATIS04 Trong tương lai, tôi sẽ tiếp tục sử dụng dịch vụ của nhà cung
cấp dịch vụ mạng hiện tại.
SATIS05 Nhìn chung tôi hài lòng với các dịch vụ được cung cấp bởi
nhà mạng di động hiện tại.
Complaint intention (COMPL)
COMPL01 Trong 3 tháng tới, tôi chắc chắn sẽ phàn nàn các vấn đề làm
tôi không hài lòng đến người quản lý của nhà mạng di động.
COMPL02 Trong 3 tháng tới, tôi sẽ yêu cầu nhà mạng di động xử lý sự
cố mà tôi gặp phải.
COMPL03 Trong 3 tháng tới, tôi sẽ thảo luận về các vấn đề tôi gặp phải
trong quá trình sử dụng dịch vụ với người quản lý hoặc nhân
viên của nhà mạng viễn thông.
COMPL04 Trong 3 tháng tới, tôi sẽ thông báo các sự cố tôi gặp cho
công ty viễn thông để họ làm tốt hơn trong tương lai.
17


4.5 CORRELATION ANALYSIS
Pearson's correlation analysis results show that all of the independent factors
in the model correlate with the dependent variable, which allows the SEM
analysis step in the next section.
4.6 CONFIRMATORY FACTOR ANALYSIS - CFA
Confirmatory Factor Analysis (CFA) “is used to confirm the measurement
theory that proposed in this study. According to Hair et al (2009), a
measurement theory identifies how measured variables correspond to

constructs in a theoretical model. Unlike EFA, CFA requires that constructs,
specific number of factors and variables loading on each factor are clearly
defined. Briefly, CFA was conducted to test whether the measurement model
fits with data.”

Hình 4.1: Kết quả phân tích CFA lần 2
18


Index Standard
ChiSquared/df
TLI
AGFI
CFI

≥ 0.90*
≥ 0.85**
≥ 0.95**

RMSEA

≤ 0.08**

< 5.00

Reference

Score

Wheaton & cộng sự (1997)

4.054

Hu & Benler (1999)
Hair & cộng sự (1995);
Hu & Benler (1999)
MacCallum & cộng sự
(1996)

0.944
0.919
0.954
0.057

4.7 STRUCTURAL EQUATION MODEL - SEM
Structural Equation Model (SEM) should be conducted to examine both the
measurement model and the structural regression model. The design of
Structural Equation Model was demonstrated as below:

Hình 4.2 Kết quả SEM
Estimate
H1
H2
H4

SATIS
SATIS
SATIS

<-<-<--

DISJUS
PROIN
TRUST

0.276
0.540
0.224
19

S.E.
0.052
0.054
0.039

C.R.
5.345
9.950
5.683

P
***
***
***


H5

COMPL

<--

SATIS

-0.634

0.041

-15.391

***

Estimate
(standardized)
H1
H2
H4
H5

SATIS
SATIS
SATIS
COMPL

<--<--<--<---

DISJUS
PROIN
TRUST
SATIS

0.253
0.540
0.195
-0.532

Index Standard
ChiSquared/df
TLI
AGFI
CFI

≥ 0.90*
≥ 0.85**
≥ 0.95**

RMSEA

≤ 0.08**

< 5.00

Reference

Score

Wheaton & cộng sự (1997)

4.054

Hu & Benler (1999)
Hair & cộng sự (1995);
Hu & Benler (1999)
MacCallum & cộng sự
(1996)

0.944
0.919
0.954
0.057

The analytical results reveal that justice components have positive impacts
on customer satisfaction, estimations are 0.276 and 0.540 respectively, with Pvalue satisfied statistical standards. This result means that hypotheses H1 and
H2 are accepted. Thus, procedural justice and distributive justice have positive
effects on satisfaction of customer using telecom services in Mekong Delta.
Trust also has a positive and statistically significant impact on customer
satisfaction with coefficient β = 0.224, hypothesis H4 is accepted. The result
supports that the more customers trust in telecom service provider, the higher
satisfied they fell.
The results also show that satisfaction play a role as a mediation factor which
shapes complaint intention. Coefficient of the relationship between satisfaction
and intention to complain is β = -0.634 (p-value = 0.000). This result shows that
the more satisfied customers are, the less they intention to complain about the
service.
Finally, the model of complaint intention of customer using telecom services
in Mekong Delta region is presented as follows.

20

Distributive justice

0.253
0.540

Procedural justice

Satisfaction

-0.532

0.195

Complaint
intention

Trust

Hình 4.3: Mô hình các yếu tố ảnh hưởng đến ý định phàn của khách hàng sử
dụng dịch vụ ĐTDĐ tại khu vực ĐBSCL (chuẩn hóa)
4.8 BOOSTRAP ANALYSIS
In order to test the sustainability of the theoretical model, boostrap analysis
is conducted. The number of boostrap sample was B = 3,000 times. The results
of bias standard deviation shows that the values are small and stable. The CR
index is less than 1.96 (1.96 is the value of the normal distribution at 0.9750),
p-value deduction is> 5%, so it can be concluded that SEM results are reliable.
4.9 MODERATOR ANALYSIS – PERCEIVED RESPONSIVENESS
The study uses technique: multigroup structural equation modeling to
examine the moderator level of perceived responsiveness. The observations
were divided into 2 groups using median-split method: Group 1 is a collection

of observation who had a high perceived responsiveness and Group 2 is a
collection of observation who had a low perceived responsiveness. Chi-square
differential test was conducted to assess the difference between the two sample
groups mentioned above:
Chidf
square

PInvarian
value
t?

Overall
Model
Unconstrained
Fully
constrain
Number of
groups
Difference

2,084.2
37
2,182.4
46

57
6
60
2
2

98.209

26 0,000

21

No

Groups are different at
the model level. Check
path differences.


The results of Chi-square test confirm the difference at the model level
between the group with low perceived responsiveness and high perceived
responsiveness. Then, path differences analysis should be proceeded.

Finally, multigroup structural equation modeling (SEM) is carried out to
determine the degree of influence of the moderator variable. The results showed
that the effect of satisfaction on the complaint of low perceived responsiveness
group was -0.59 while the effect of high perceived responsiveness group was 0.70. This result supports to the expectations that when customers feel the
provider is responsible, they tend to complain directly if they are not satisfied.
Hypothesis H6 is accepted.
4.10 CONTROL VARIABLE - GENDER
Gender often plays a role as a control variable in the research model. Chisquare testing was conducted to examine the difference between male and
female groups. The results of Chi-square testing show that there is no evidence
to confirm the role of gender in the research model. Path analysis was also
conducted, but still did not provide any better evidence. This means the role of
gender can not be supported in the research model about customers’ complaints

in the context of telecom service in Mekong Delta.

22


CHAPTER 5
CONCLUSION AND IMPLICATION
This research contributes to an overall conceptual understanding of the
nature and important antecedents of complaint intentions, from the justice, trust,
and responsiveness perspective. Consumers’ perception of justice is an
important and unique consideration in telecom service. It provides important
insight for service providers, in terms of understanding the major concerns of
users in a virtual social exchange context. Customer satisfaction is also a key
mediator in the reduction of complaint intentions.
Several practical and theoretical implications arise from these findings.
Firstly, justice perception indicates a fair exchange proces in the various forms
of contact-based experience with telecom service providers. Providers should
try to improve communication channel with customers from marketing effort,
in particular, relationship-based marketing, to allow a rich interactive
experience with customers. The communication channel is the principal
mechanism and facilitates a mutual understanding between customers and
providers. A new marketing approach, namely relationship marketing, intends
to analyze customers’ requirements in depth, via frequent and personal
interactions with customers, in order to build a long-term relationship (Tax et
al., 1998). It is critical to effectively help communicate with consumers to
maintain a high quality of products/services. Moreover, it also results in better
customer services to improve justice perception in the trading process. Finally,
trust is also important for customer satisfaction and complaint intentions. This
implies that communication channel can be improved if consumers are able to
first form a psychological state of positive feelings toward providers. The effort

for building trust belief may involve many activities. Some activities are
relevant to e-marketing to build trustworthiness. A reliable technological
infrastructure includes good quality of service information, security, privacy,
and quick response time, which allow consumers to feel comfortable about
providing personal information to providers. Other activities consider pre- and
pos-purchase factors, such as reputation, satisfaction guarantee policies,
reliable quality, and after-sale services. Subsequent research in telecom service
context could use this study as a foundation. Firstly, to the best of the authors’
knowledge, few studies relate to complaint intentions in telecom service,
especially in Mekong Delta of Vietnam. This integration of three important
factors provides important foundation for future work. Future research might
consider other behavior-based factors, such as personal traits and flow
23


experience, to enable a deeper understanding of complaint behaviors. In
addition, other technology-based frameworks, such as TAM and the IS success
model (DeLone & McLean, 2003), could be also used to extend the scope of
this study.
Finally, although this research has produced some interesting results, a
number of limitations may be inherent in it.
1-Samples are collected only in Can Tho, Long An, Tien Giang and Kien
Giang, so the research results could be limited when applied to other localities;
2-Due to the deployment conditions, the non-probability sampling method is
chosen, so there are some natural weaknesses in generalizing other cases;
3-potential respondents were asked to recall a recent service failure, before
they answered the survey, to provide a solid basis for completing the survey in
a consistent manner. However, this may have a potential effect of the treatment
on respondent’s perception of justice perception and trust belief, since this study
investigates a priori experience of the antecedents.

The author expects that future studies will understand the limitations of this
study and improve to change the negative perspectives on complaining
behavior. Complaints should be treateded as a gift, a great opportunity to
enhance customer loyalty (Voorhees & Brady, 2006), not a problem for
companies (Kendall & Russ, 1975).

24