Factors affecting loyalty of internet customers, evidence form vietnam

1.4 The research questions Based on the research objectives, we accordingly give two main questions as follow: - What are the factors that influence customer loyalty in Internet services

Nguyen Thi Kim Quyen

FACTORS AFFECTING LOYALTY OF INTERNET CUSTOMERS: EVIDENCE FROM VIETNAM

MASTER OF BUSINESS (Honours)

Ho Chi Minh City – Year 2012

Nguyen Thi Kim Quyen

FACTORS AFFECTING LOYALTY OF INTERNET CUSTOMERS: EVIDENCE FROM VIETNAM

I would not have been able to complete my thesis without the advice, assistance,understanding and encouragement of a number of people I would like to gratefullyand sincerely acknowledge their assistance and support First of all, the mostimportant person I would like to thank is my supervisor, Dr Dinh Thai Hoang.Thanks to his valuable advice and generous guidance throughout the process ofcompleting this thesis Secondly, I would like to express my gratefulness to Dr.Nguyen Dinh Tho for his kindness in supporting me to deal with analyzing data andother issues of the thesis

Furthermore, I would like to thank all of the staff at International School of Business– University of Economics Ho Chi Minh City who are always enthusiastic and give

me a lot of administrative support, assistance and understanding

Finally, I would like to give my gratefulness to all classmates who have been sharingwith me all their knowledge and experience as well as happiness and difficultiesduring MBA course I would not complete my thesis without their encouragementand support

In brief, it has been a long journey and I have grown in so many ways I would like tosay “thank you” again for the involvement of everyone who has helped me tocomplete this thesis

For Internet service providers (ISP), it is essential for them to understand the criticalfactors that influence customer loyalty These issues are critical to the long-termbusiness success of any service provider Based on the literature reviewed in thethesis, this study empirically investigates the impact of factors on customer loyalty

A quantitative method via questionnaire survey was used to obtain the views ofapproximately 300 Internet users in Ho Chi Minh city regarding their opinions andperceptions about their current ISP and their intention to stay with them Multiplelinear regression analysis was used to investigate the relationship between customerloyalty and influencing factors mentioned above

According to the findings of this research, in order to retain internet servicecustomers, providers need to take into consideration with factors such as networkquality, price perception, perceived value and trust The results of this research areimportant for ISP to enhance and improve their competitiveness, in other words, tobecome successful and have more loyal customers, which will ultimately result inmore profitability

Keywords: customer retention, customer loyalty.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS 3

ABSTRACT 4

CHAPTER 1: INTRODUCTION 5

CHAPTER 1: INTRODUCTION 5

1.1 Introduction 5

1.2 Research background 5

1.3 The research objectives 7

1.4 The research questions 7

1.5 The research model 7

1.6 The research methodology 7

1.7 The research structure 8

1.8 Conclusion 8

CHAPTER 2: LITERATURE REVIEW 10

2.1 Introduction 10

2.2 Customer loyalty with internet service provider 10

2.2.1 Network quality and customer loyalty 11

2.2.2 Price perception and customer loyalty 11

2.2.3 Perceived value and customer loyalty 12

2.2.4 Trust and customer loyalty 13

2.3 Controlling factors 13

2.4 The research model 13

2.5 Conclusion 15

CHAPTER 3: RESEARCH METHODOLOGY 16

3.1 Introduction 16

3.2 The research approach 16

3.3 Measurement 16

3.4 Sample and Data collection 19

3.5 Data analysis 20

3.6 Conclusion 22

CHAPTER 4: DATA ANALYSIS AND RESULT 23

4.1 Introduction 23

4.2 Preliminarily data analysis 23

4.2.1 Cleaning and screening data 23

4.2.2 Profile of respondents 23

4.3 Evaluating the measuring scales 25

4.4 Multiple Linear Regression analysis 32

4.4.1 Introduction 32

4.4.2 Computing variables 32

4.4.3 Assumption for Multiple Regression 32

4.4.4 Testing hypotheses 35

4.4.5 Controlling variables 37

4.5 Conclusion 39

CHAPTER 5: CONCLUSION AND IMPLICATIONS 40

1.1 Introduction 40

1.2 Conclusion 40

1.3 Implications of the study 41

1.4 Limitations and suggestions for further research 42

REFERENCES 43

APPENDIX A: QUESTIONAIRE 47

APPENDIX B: DATA ANALYSIS 53

LIST OF FIGURES

Figure 1: The theoretical model 14

Figure 2: The Histogram 33

Figure 3: The Nomal P-P Plot of Regression Standardized Residual 34

Figure 4: Scatterplot 34

LIST OF TABLES

Table 1: Research questions and hypotheses 15

Table 2: Measuring scale for variables 19

Table 3: The profile of respondents 25

Table 4: The reliability statistics 27

Table 5: KMO and Barlett’s test 29

Table 6a: The pattern matrix of EFA result 30

Table 7b: The pattern matrix of EFA result 31

Table 8: The Cronbach’s alpha result - retest 31

Table 9: Collinearity statistics 35

Table 10: Hypotheses testing 36

Table 11: Regression with controlling variable of gender 37

Table 12: Regression result with controlling variable of income level 38

Table 13: Regression result with controlling variable of using experience 38

CHAPTER 1: INTRODUCTION 1.1 Introduction

This research examines the driving forces of internet users’ loyalty and their intention

to stay with their current provider among existing Internet Service Providers (ISP) inthe context of Vietnam telecommunication sector This chapter will provide generalinformation of research problem as well as research background, identify researchobjectives and research questions This chapter also briefly outlines researchstructure

1.2 Research background

High-speed internet services are becoming more and more popular and the rapiddevelopment of information technologies affects every aspect of our lives Theimportance of information and communication technology is undeniable as it hasbeen applied in various fields for the purpose of service enhancement Although basicInternet services had existed in Vietnam since the early nineties, until 1997, the firstcommercial Internet Service Provider (ISP) opened for business – Vietnam DataCommunication Company (VDC) which belongs to Vietnam Posts &Telecommunication Groups (VNPT) Since then, Vietnam has been connected by twogateways: one in Hanoi connecting with Hong Kong and Australia, and the other in

Ho Chi Minh City, which connects with the United States by Sprint Internet usageremains low in comparison with that of other Asian nations, yet connectivity hasincreased rapidly over the past few years, thanks to an initiative by the Ministry ofPost and Telecommunications According to the internet statistic report of VNIC(Vietnam Internet Center), up to June 2012, the number of internet users in Vietnam ismore than 31 million which is equivalent to the scale of 35.4 percent users per capita.Over more than fifteen years of formation and development, together with thedevelopment of internet services, the number of ISPs has also remarkably increased,which leads to severe competition among ISPs, especially in a developing countrylike Vietnam Until now, there are more than ten ISPs operating in Vietnam includingVDC/VNPT, Viettel, FPT, Netnam, SPT,

Hanoi Telecom, VTC, G-Tel, Vishiped and CMC TI in which three of the leadingcompanies: VDC/VNPT, Viettel, FPT have obtained more than 85 percent of totalInternet service market share in Vietnam As a result, ISPs must try their best todevelop their marketing strategy and improve their business performance to increasetheir level of competitiveness From the above statistics, we can understand that thecompetition of Internet services in Vietnam is a severe war among three big ISPs:VDC/VNPT, Viettel and FPT Beside the diversification of services, enhancing thecustomer satisfaction and customer loyalty is considered to be the most effective way

to compete in the market, especially in telecommunication industry which is on theverge of saturation point in some years’ time Customer loyalty are believed to be themost effective weapons for providers to complete with other rivals because thesefactors are the primary variables that impact customers’ intention to stay with theircurrent providers (Khatibi, Ismail & Thyagarajan, 2002) The providers which satisfycustomers can expect higher market shares and greater profitability and, in turn,encourage their customers to become loyal (Khatibi, Ismail & Thyagarajan, 2002).Aaker (1991) has discussed the role of loyalty in the brand equity process and hasspecifically noted that brand loyalty leads to certain marketing advantages such asreduced marketing costs, more new customers and greater trade leverage Expandingloyal customers can assist in generating a business’s profitability because the cost ofretaining existing customers is less than that of acquiring new ones (Reichheld, Teal,1996) Therefore, understanding how and why a sense of loyalty develops incustomers remains one of the crucial management issues of our day By assessing theInternet service users in Ho Chi Minh city, this research will give useful information

of a range of factors impact on Vietnamese internet users’ perceptions of their currentISPs, from which factors effecting customer loyalty, especially in telecommunicationsector, will be derived In other words, this research provides ISPs the direction andvaluable understandings of what factors should be taken into account to becomesuccessful

and have more loyal customers, which ultimately result in more profitability andleads to their strong competitiveness and stable growth in the future.

1.3 The research objectives

Arising from the research problem, this research aims to explore the main attributes ininternet service provisioning - customer loyalty in the context of Vietnamtelecommunication market More specifically, the main objective of this research is toidentify the factors influencing customer loyalty in Internet services in Vietnam Wealso examine the affecting level of each factor on Vietnamese internet users’ loyalty

1.4 The research questions

Based on the research objectives, we accordingly give two main questions as follow:

- What are the factors that influence customer loyalty in Internet services in Vietnam?

- How do these factors influence on loyalty of Internet customers?

1.5 The research model

Based on the literature review, a model with four hypotheses are established in order

to explore the research problems and develop the hypotheses Figure 1 – The researchmodel – in Chapter 2 shows the proposed model for this research regardingcustomers’ intentions to stay with their current ISP in Vietnam This model leads tofour hypotheses tested that show the relationship between network quality, priceperception, perceived value and trust and customer loyalty

1.6 The research methodology

An overview of the research methods used to collect and analyze the data are brieflydiscussed In this research, a survey with questionnaire in was used to collect data.The questionnaire was made in English and then translated into Vietnamese becauseall of respondents are Vietnamese Random sampling was chosen to reduce bias in

the sample The data collected from the survey was coded, analyzed and interpreted

by SPSS software More details of the research methodology are provided in Chapter

3 and the result of the research is presented in Chapter 4

1.7 The research structure

This research begins with Chapter 1 – Introduction, which provides generalintroduction and brief background of internet services in general and in the context ofVietnam in particular to the research including research background, researchobjectives, research questions, research model, research methodology and structure ofthe research

Chapter 2 – Literature review - screens the literature of customer loyalty fromprevious researches Basing on these literature reviews, a model with four hypotheses

is developed We examine the relationship between customer loyalty and its affectingfactors including network quality, price perception, perceived value and trust.Besides, demographic factors such as gender, income, and experience of service usingare also tested their impact on customer loyalty

Chapter 3 – Research methodology - presents methods to conduct this research,consisting of three sections Firstly, measuring scale with items for each factor isprovided Next, sample and data collection are explained in detail The last onedescribes the process and analytical techniques used for data analysis

Chapter 4 – Data analysis - reports the results from data analysis process andconfirms the value of hypotheses mentioned in Chapter 2 and relationship amongfactors

Chapter 5 – Conclusion, implications and limitations – repeats the result derived fromthis research Moreover, this chapter also discusses limitations of the research andsuggestions for future researches

1.8 Conclusion

This chapter presents the background to the research in order to introduce the keyresearch problem, research objectives, research questions, research model and

research methodology By reviewing the existing literature, this research aims to testthe relationship between customer loyalty and its affecting factors in VietnameseInternet service market Four factors are assumed to have positive effect on loyalty ofInternet customers Data are collected by questionnaire and analyzed by SPSSsoftware.

CHAPTER 2: LITERATURE REVIEW

2.1 Introduction

This chapter presents a review of the literature in the main areas of customer retentionwhich concentrates on factors influencing customer loyalty for internet serviceproviders, especially in the context of Vietnam telecommunication industry At first,from the tested relationship between customer loyalty and relating factors includingnetwork quality, price perception, perceived value and trust, four hypotheses derivedfrom which (H1, H2, H3, H4) are proposed Then, the controlling variables such asgender, income and service using experience are also tested their effect on customerloyalty

2.2 Customer loyalty with internet service provider

Customer loyalty is an outcome of a customer’s purchase and use resulting from thecomparison of the reward and the costs of the purchase in relation to the anticipatedconsequences (Churchill & Surprenant 1982) Jacoby (1975) advises that loyalty is abiased behavioral purchase process caused by the customer’s emotional processes.Loyalty can be measured by several factors including proportion of purchase,purchase sequence, and probability of purchase (Dick & Basu 1994) The meaning ofcustomer loyalty can be explained as the customer’s intention to stay with the currentservice provider and, at the same time, their behavior to make frequent repeatpurchases of their preferred brands As a result, most businesses prefer to retain theirvalued customers and persuade them to become more loyal because this can result ingreater income and profit Without customer loyalty, service providers may not beable to retain their competitive advantage This can be explained as loyal customersare easier to serve and make fewer demands on employees since they already knowand understand the company’s working processes (Chow & Holden 1997) Loyalcustomers appear to provide more repeat patronage and are less likely to shop aroundwith competitors for better deals (Bowen & Chen 2001) In order words, companiescreate profit from their long-term relationship The

longer loyal and profitable customers stay with a business, the more profit that thebusiness can gain (Reichheld & Teal 1996).

In general, a range of advantages arise when businesses have loyal customersincluding increasing profitability, increasing price premiums, increasing long-termrelationship and commitment, increasing purchase intention and positive word-of-mouth, reducing marketing and operating costs and increasing switching barriers toprevent customers from defecting their current service provider (Reichheld & Teal1996) Therefore, customer loyalty cannot be neglected and most companies must trytheir best to satisfy their customers in order to develop long-term relationships withthem and to ensure their customers become loyal

2.2.1 Network quality and customer loyalty

Network quality is one of the most important drivers of overall service quality whichleads to customer satisfaction in the context of telecommunication (Chun & Hahn,2007; Wang, Lo & Yang, 2004) Previous literature suggested that stability,transmission speed and network coverage are the core attributes of network quality(Woo & Fock, 1999; Mohd R.B Yaacob, 2011) The stability and transmission speed

of internet service were important to users and they will consider changing to otherproviders due to stability and speed factor (Mohd R.B Yaacob, 2011)

Moreover, the uptime of service was tested to have impact on customer satisfaction aswell as customer loyalty (Wang, Lo & Yang, 2004) The results of Wang, Lo &Yang’s (2004) study suggested that, in order for service providers to be successfullycompetitive, they must try to improve the availability of service We, therefore,hypothesize that:

H1: Network quality has a positive impact on customer loyalty.

2.2.2 Price perception and customer loyalty

With regards to price perception, although Internet broadband users are willing to paymore for better service, they will consider changing to other provider because of theprice factor (Mohd R.B Yaacob, 2011) Therefore, we believe that customers

of Internet services are sensitive to price and the higher price level could lead to lowdemand, accordingly Ranaweera & Neely (2003) showed that price perception has adirect linear relationship with customer loyalty in telecommunications sector Wetherefore believe that such relationship may be more explicit in the fierce pricecompetitive in telecommunication market like Vietnam Hence, we formulate thefollowing hypothesis:

H2: Service price is positively related to customer loyalty.

2.2.3 Perceived value and customer loyalty

Perceived value can be defined as the customers’ overall assessment of the utility ofproducts or services based on perceptions of what is received and what is sacrificed(Monroe 1991, Parasuraman, Zeithaml & Berry 1988) Simply stated, perceived value

is trade-off between benefits and sacrifices The perceived sacrifice includes all thecosts the customer faces when they are in the process of purchasing such as purchaseprice, acquisition costs, installation, repair and maintenance, risk of failure or poorperformance (Monroe 1991, Parasuraman, Zeithaml & Berry 1988) The perceivedbenefits include physical and service attributes, technical support, and other indicators

of perceived service quality Monroe 1991, Parasuraman, Zeithaml

& Berry (1988) point out that perceived value is subjective and individual, andtherefore varies among customers Research evidence suggests that customers whoperceive that they receive value for money are more satisfied and tend to becomeloyal customers more often than ones who do not (Zeithaml, Berry & Parasuraman1996) Customers who remain with a service provider for a longer period of timebecause they are pleased with the perceived quality and value of the services tend tobuy more additional services and spread favorable word-of-mouth messages and have

a willingness to recommend the firm to others (Boulding 1993, Zeithaml & Berry1988) Perceived value, from the service providers’ standpoint, should be viewed as amajor competitive weapon and can be used to create and improve their sustainablecompetitive advantage (Khatibi, Ismail & Thyagarajan 2002) By adding

more perceived value to the service, customer loyalty can be increased We, therefore,hypothesize that:

H3: Perceived value has a positive effect on customer loyalty

2.2.4 Trust and customer loyalty

Trust can be viewed as an antecedent of customer loyalty (Dick & Basu, 1994) Tocreate customer loyalty toward a brand, building long-term relationships withcustomers in order to win their trust is considered advantageous (Morgan & Hunt,1994) When customers have trust in the products or services of their preferred brand,they are willing to purchase more of the brands without the need to be heavilypersuaded, and this can result in lower advertising and other related costs (Murphy,2001) Trust can be viewed as being complementary to satisfaction, in ensuring long-term relationships, and strengthening customer loyalty (Hart & Johnson, 1999) Inaddition, trust is a willingness to rely on other parties in whom one has confidence,when the buyers trust the sellers, it affects their purchasing behavior (Chow &Holden, 1997) This leads to the following hypothesis:

H4: Trust positively influences customer loyalty.

2.3 Controlling factors

Choudrie, Jyoti (2005) find there are strong influences of demographic variables such

as gender, income on customer loyalty In this research, the relationship betweengender, income also were tested Besides, the customer’s experience with serviceshould be examined its impact on customer loyalty

2.4 The research model

The research model is the foundation that the entire research is based on and a goodmodel should help in identifying and labeling the underlying variables in thecircumstances that are related to the research problem (Sekaran 2000) This section isbuilt basing on the existing literature to propose a model linking the correlationsbetween customer loyalty and relating variables To appropriately answer the

research questions, we develop the model with five main variables including networkquality, price perception, perceived value and trust, which are considered to bedependent variables affecting customer loyalty Furthermore, controlling variable ofdemographic is tested its influence on the relationship among these variables.

The proposed model and all hypotheses tested in the context of Vietnamtelecommunication market are outlined as follow:

Network quality H1

H2Price perception

Perceived value

Trust

Customer loyaltyH3

H4

Controlling variables (Gender, Experience,

Income)

Figure 1: The theoretical modelThe following table provides a summary of the research hypotheses and questions to

be tested in the research:

1 What are the factors that influence H1: Network quality has a positive impactcustomer loyalty in Internet on customer satisfaction

services in Vietnam?

2 How do the factors influence

customer loyalty in Internet

Table 1 - Research questions and hypotheses

2.5 Conclusion

In summary, we develop four hypotheses in the theoretical model in the context ofVietnam telecommunication industry The model proposes that customer loyalty ispositively affected by network quality, price perception, perceived value and trust Inaddition, we also examine the effect of demographic factors on customer loyalty

CHAPTER 3: RESEARCH METHODOLOGY 3.1 Introduction

The purpose of this chapter is to introduce the research methods and approaches used

to investigate customers’ perceptions of Internet service users in Ho Chi Minh city.Specifically, this chapter will give you some general information about the way weconducted the research including the research approach, measurement, sampling, datacollection, and data analysis technique

3.2 The research approach

Quantitative approach primarily involves statistical analysis and relies on numericaldata to ensure objective and accurate results are obtained in order to draw conclusions

or to test hypotheses (Ticehurst & Veal, 2000) This research involves quantifying therelationship between variables and will use quantitative research to test the researchassumptions and hypotheses in order to solve the research problem This is aquantitative research because it seeks to answer set hypotheses from reviews ofprevious literature rather than attempting to develop new theory

3.3 Measurement

This section presents the items for measurement of all relating factors Items of each factor are developed from previous literature review mentioned in Chapter 2

Measuring scale of network quality from Mohd R.B Yaacob (2011) and Wang, Lo

& Yang (2004) were used:

- The upload and download speed of network is always strong

- The uptime of network is always available without interruption

- The connection quality is always reliable

- I can access the network at anytime without delay

- The price charged by X is reasonable

- The service’s price of X is cheaper than others

- Service supplied by X is equivalent to its price

- Service offered by X is better value for money than what I would pay for the same service of others.

- I’m willing to pay more for better service quality

Based on Wang, Lo & Yang (1988), five items for measuring the effect of perceived value on customer loyalty are adopted:

- The customer service staff of my ISP gives me adequate support

- Whenever I have problem with services, X takes corrective action without delay

- X keeps me informed of things that I need to get the best use of the service

- I feel comfortable with the willingness of assistance and support from X

- Overall, the chosen offerings are worth for my money, effort and time

- I believe that X will be ready and willing to offer me assistance and support

- I believe that X always fulfills the promises that it makes to its customers

- I believe that X has necessary ability and knowledge to fulfill its tasks

- I believe that X always puts interests of its customers first of all

- I will recommend X as the best ISP in HCMC

- I will continue to do business with X

- X is my first choice in my future internet service needs

- I will purchase other services offered by X in the future

- I will encourage friends and relatives to do business with X

Items for demographic of respondents:

No Variables Literature review Measuring items

1 The upload and download speed of network isalways strong

Network Mohd R.B Yaacob 2 The uptime of network is always available

quality

Lo & Yang (2004) 3 The connection quality is always reliable

4 I can access the network at anytime withoutdelay

1 The price charged by X is reasonable

2 The service’s price of X is cheaper than others

3 Service supplied by X is equivalent to its price

2 Price Ranaweera & 4 Service offered by X is better value for money

perception Neely (2003): than what I would pay for the same service of

3 Perceived Wang, Lo & Yang 3 X keeps me informed of things that I need to

value (1988) get the best use of the service

4 I feel comfortable with the willingness ofassistance and support from X

5 Overall, the chosen offerings are worth for mymoney, effort and time

4 Trust Dick & Basu 1 I believe that X will be ready and willing to

(1994) offer me assistance and support

2 I believe that X always fulfills the promises that

it makes to its customers

3 I believe that X has necessary ability andknowledge to fulfill its tasks

4 I believe that X always puts interests of itscustomers first of all

1 I will recommend X as the best ISP in HCMC

2 I will continue to do business with X

3 X is my first choice in my future internet

5 Customer Nguyen and service needs

loyalty Leblanc (1991) 4 I will purchase other services offered by X in

the future

5 I will encourage friends and relatives to dobusiness with X

Table 2: Measuring scale for variables

3.4 Sample and Data collection

Sampling can be described as the process of selecting a sufficient proportion ofpeople from the entire population (Sekaran, 2000) The process of random samplingwas used in order to reduce bias in the sample of Vietnamese Internet users.According to DeCoster, J (2004), minimum sample size used in statistics analysisshould be equal to or greater than five times of the number of independent variables,but not less than 100 to generate reliable results: n ≥ 100 and n ≥ 5k (where k is thenumber of items) This research has 23 items, as a result, the minimum sample sizerequired to run EFA in this research is: n = 5 x 23 = 115 Otherwise, Tabachnick &Fidell (2001) stated that the minimum sample size in case of multiple regressionshould be: n = 50+8m (where m is the number of independent variables) Apply thisformula for 4 independent variables of this research, we have the minimum samplesize for multiple regression: n = 50 + 8 x 4 = 82 In summary, this research needs

115 samples at least to take EFA and multiple regression analysis With quantitativeapproach, in order to ensure the reliability and validity, the adequate sample size forthis research was about three hundreds of internet service users The survey wasconducted from August to October with Internet service users in Ho Chi Minh city.The questionnaire was sent to more than 500 potential respondents to get the targetsample size of around 300 respondents.

Data was collected by questionnaire survey which consists of three sections The firstsection was questions about customers’ experience with ISPs; the questions in thesecond section focused on respondents’ opinions on service and their future decisionwith the current ISP; and the last one were demographic questions Basically, thequestionnaire contained statements that could be answered through ranking and open-ended questions as well Respondents were asked some fundamental questions abouttheir experiences with ISP and which companies were their current ISPs In thesecond section, a seven-point Likert scale ranging from “Strongly disagree” (1) to

“Strongly agree” (7) was used to measure respondents’ perceptions about their currentISP’s service quality and intention to stay with or leave for another one Thequestionnaire was made in English at first and translated into Vietnamese because therespondents were all Vietnamese

3.5 Data analysis

Following data collection, the next step is the process of data analysis in order toaddress the research questions The procedure of converting raw data into informationconsists of several processes including editing, coding, data entry and data analysis(Zikmund 2000) In this research, the SPSS software was chosen to analyze the data

in terms of descriptive statistics Data processing procedures and analyticaltechniques used in this research are discussed as below:

Firstly, the Cronbach alpha was used to examine the reliability of measuring scale.Cronbach alpha is a statistical test if correlation of the items in the scale relative toeach other (Hair 2006) So, this method of analysis can remove inappropriatevariables and limit junk variables in the research process and assess the reliability of

the scale through the Cronbach alpha coefficient (Hair 2006) The scales are reliablewhen Cronbach’s alpha coefficient of each scale is equal to or bigger than 0.7 (Pallant2005).

Secondly, Effecting Factor Analysis (EFA) was conducted to test the number offactors extracted Factor analysis is a multivariate statistical technique which definesthe underlying structure among a large number of variables (Hair 2006) The twofundamental purposes of factor analysis are to summarise the information contained

in a large number of variables and condense the data into smaller number of factors(Hair, 2006) In this research, there were five factors that were assumed to haveimpact on customer loyalty of Internet services Factor analysis was used to createfactors for each of the five measuring scale related In addition, factor analysis helpedthe researcher to determine which factors were highly correlated to customer’sloyalty The result is considered to be acceptable when following conditions are met(Pallant, 2005):

 The sample size should be equal to or greater than five cases for each variable

 Factor analysis is appropriate to data if:

 The Kaiser-Meyer-Olkin value (KMO) is 0.6 or greater

 The Bartlett’s test of sphericity is statistically significant: p < 0.05

 The number of factors is determined when:

 The total variance explained by these components should be above 50%

 Factor loading criteria should be 0.5 or more to ensure a practicalsignificance

Thirdly, Pearson’s correlation coefficient was used to examine the relationshipsbetween two or more research variables (Veal 2005) According to Veal (2005), therelationship between the variables can take different forms If the value of correlationcoefficient is 1.0, then there is a perfect positive correlation between two variables Incontrast, if the value of correlation coefficient is -1.0, it can be

concluded that there is a perfect negative correlation between two variables Inaddition, there is no relationship between two variables of the value of correlationcoefficient is zero.

Finally, multiple regression analysis was conducted to analyze the relationship amongvariables Multiple regression analysis is a statistical technique that is used to analyzethe relationship between several independent variables and a single dependentvariable (Hair et al.2006) It was used to examine the simultaneous effects of severalindependent variables on a dependent variable In this research, Multiple LinearRegression method was used to test the research model and hypotheses Pallant(2005) explains the conditions to accept the result are:

 The sample size is: n > 50 + 8m (where m is the number of independent variables)

 No multicollinearity is found

 Normality and linearity should exist

 We also use R-square value to express how much of the variance in the

dependent variable was explained by the model

3.6 Conclusion

This chapter introduces the research methods and general information about the way

we conducted the research including the research approach, measurement, sampling,data collection, and data analysis technique More detailed discussion of the analyticmethods used and results are provided in the Chapter 4

CHAPTER 4: DATA ANALYSIS AND RESULT 4.1 Introduction

This chapter provides the results of the data analysis step by step and interpretations

of research findings The purpose of this chapter is to answer the research questionsand test the study hypotheses This chapter consists four main parts: (1) preliminarilydata analysis, (2) evaluating the measuring scales and (3) testing the hypotheses andthe research model

4.2 Preliminarily data analysis

The quantitative research was conducted by using a survey questionnaire which wascompleted by Internet service users in Ho Chi Minh city At first, 100 samples wereused for pilot test which was specifically mentioned above Then, 286 completedquestionnaires were collected and analyzed in the main study Preliminarily dataanalysis consists of cleaning and screening data and summarizing profile ofrespondents

4.2.1 Cleaning and screening data

This step helps us to detect inappropriate samples which were completed by carelessrespondents For example, the respondent chose the same range of scale for allstatements in the questionnaire Moreover, missing data were also considered to havenegative impact on the result In this research, missing data was not a problembecause the questionnaire survey was designed and completed online Respondentscould not submit their answers with any missing data in the questionnaire So, in thisstep, we only checked inappropriate ones After cleaning and screening data, 11samples were distorted by the first error – out of range values

4.2.2 Profile of respondents

The sample size should be large enough to make sure coverage of collected data Ingeneral, the ratio of observations to independent variables should never fall belowfive to one Nevertheless, the desired level is between 15 and 20 observations for

each independent variable (Hair et al.1998) This research has 5 variables with 23measuring items Therefore, according to this standard, the minimum sample size forthis research is about 194 However, in order to make sure the sample size wasproportionally distributed among ISP’s market share, the desired sample size for thisstudy was around 300 samples The link of the online designed questionnaire was sent

to more than 500 Internet service users Then, 297 samples were recorded by googletools After cleaning and screening data, 286 samples were used for official analysis

in which (1) VNPT: 148 samples equal to 52%, (2) Viettel: 72 samples, equal to 25%,(3) FPT: 58 sample equal to 20% and (4) remaining 3% for other providers such asSCTV, SPT, The profiles of respondents are outlined as following table:

Less than 1 year 42 15%

Table 3: The profile of respondents

In summary, gender of respondents was equal between male and female Most ofrespondents belonged to the age group from 25 to 35 years old and average monthlyincome from more than 10M to 15M In addition, there were 47% of respondents hadbeen experienced with their current ISP for more than 3 years

4.3 Evaluating the measuring scales

Measuring scales were preliminarily evaluated the reliability and validity byCronbach alpha and Exploratory Factor Analysis (EFA) Firstly, Cronbach alpha wasused to test scales’ reliability Reliability can be defined as the dependability orconsistency of the variable’s measurement that is the degree to which measures arefree from error and therefore provide consistent results (Neuman 2006) TheCronbach alpha coefficient is a commonly used method to examine the reliability formultipoint-scaled items It can be considered a perfectly adequate index to test theconsistency reliability (Sekaran 2000) Therefore, the Cronbach alpha coefficient wasused in this research as the key test to measure scale reliability The alpha coefficientranges from 0 to 1 and where the alpha coefficient is greater than 0.7, this can beconsidered an acceptable reliability (Hair et al 1998) In this research, items withalpha coefficient less than 0.6 and corrected item-total correlation lower than 0.30would be deleted In this research, cronbach alpha coefficient of all factors wereacceptable because they are greater than 0.8, ranging from .882 to .901 TheCronbach’s alpha analysis was summarized as the table 4 below:

Item-Total Statistics

Items Scale Mean Scale Variance Corrected Item- Squared Cronbach's

if Item if Item Deleted Total Multiple Alpha if

Table 4: The reliability statistics

Network quality was measured by 4 items with good overall alpha coefficient by

901 Cronbach alpha if item deleted of all items were less than alpha coefficient offactors (.901), which means that deletion of any item from questionnaire would notimprove the factor’s reliability Besides, Corrected Item-Total correlation of allshowed good indication, that was much more than 0.3 We, therefore, concluded thatall items for network quality variable were reliable

Cronbach’s alpha of price perception variable indicated reliable coefficient by 888.Moreover, the alpha coefficient of each item showed its useful contribution to overallcoefficient alpha and Corrected Item-Total Correction column of this factor consisted

of high number Hence, price perception variable and five of its measuring scaleswere accepted and used for further analysis

The alpha coefficient of perceived value was 879 which presented an acceptablereliability as mentioned above Furthermore, Cronbach’s alpha if Item Deletedshowed that no improvement in the overall reliability of the scale occurred when any

of these questions were deleted and the Corrected Item-Total Correlation of all itemsabove 0.3 Thus, all of items represented for this factor and the factor was included insubsequent analyses

The general alpha coefficient of trust variable was 879, which was higher than therequired standard (.70), which indicated for its reliability All items also reflected