INTRODUCTION
Research background
In today's market, where consumer trust in traditional advertising is declining, word of mouth has emerged as a powerful competitive advantage for companies According to the International Word of Mouth Marketing Conference (2005), personal recommendations significantly influence consumer behavior Furthermore, a 2009 Nielsen Global Online Consumer Survey, which analyzed responses from over 25,000 online consumers across 50 countries, revealed that information from friends and online reviews are considered the most trustworthy forms of advertising.
Figure 1.1 Degree of trust in forms of advertising
According to the report, ninety percent of consumers trust recommendations from individuals, while seventy percent rely on information shared by online users This indicates a significant increase in consumers' dependence on word-of-mouth influence, whether from friends or unfamiliar online sources, during their decision-making process.
The influence of word of mouth (WOM) is well-documented in research, demonstrating its significant impact on consumer behavior As a natural and everyday form of communication, WOM occurs spontaneously and independently of sellers, which explains its powerful effect on individuals' purchasing decisions Furthermore, WOM serves as both a vital source of information and an effective persuasive communication tool, reinforcing its role in shaping consumer choices.
Word of mouth (WOM) plays a crucial role in attracting new customers and ensuring the long-term economic success of a company (Hennig-Thurau et al., 2002) It helps to alleviate cognitive dissonance among existing customers, which arises from the fear of making a wrong decision (Wangenheim, 2005; Festinger, 1957) By sharing their experiences, customers seek to reassure themselves about their purchase choices, effectively using WOM as a strategy to mitigate post-decision dissonance (Wangenheim, 2005).
Word of mouth (WOM) serves as an effective risk reduction strategy, as noted by Buttle (1998) Research by Dichter (1966) demonstrates that positive WOM enhances customers' purchase intentions for innovative products by alleviating the perceived risks associated with their acquisition Similarly, Garbarino and Strahilevitz (2004) highlight that receiving recommendations from friends can significantly lower perceived risks when making online purchases.
Customers place significant trust in the advice and recommendations from those who have firsthand experience with a service Research indicates that consumers often rely more on peer opinions than on direct communications from companies, highlighting the critical role of word-of-mouth (WOM) in influencing purchasing decisions.
Despite extensive literature highlighting the significance of word-of-mouth (WOM), research focusing on its antecedents remains limited, particularly regarding the influence of customer-employee relationships on WOM intentions.
Vietnamese aviation market
The aviation industry plays a crucial role in driving economic development in Vietnam by boosting foreign currency income, improving the balance of payments, and facilitating cultural exchange Recognizing its importance, the Government has identified aviation as a key economic sector and is actively promoting investments to modernize and enhance service quality Additionally, Vietnam's strategic location in the Asia-Pacific region positions it advantageously, as this area accounts for 50% of global air transport volume, underscoring the potential for growth and international connectivity.
According International Civil Aviation Organization (ICAO) at Montreal in July
In 2012, civil aviation in the Asia-Pacific region emerged as the most profitable, generating nearly $11 billion in net income, while European airlines reported less than $1 billion and African airlines faced a loss of $100 million The International Civil Aviation Organization (ICAO) projected a continued global aviation growth rate of 5.4% in 2012, with anticipated increases of 6% in 2013 and 6.4% in 2014 This positive trend presents significant opportunities for the development of the global aviation industry, particularly benefiting Vietnam.
In the first quarter of 2012, the Vietnamese aviation sector experienced a notable increase in air passenger numbers, rising approximately 4.4% compared to the same period in the previous year, according to a report by the Vietnamese General Statistics Office.
According to the Investment newspaper, the International Air Transport Association (IATA) predicts that Vietnam will emerge as the third fastest-growing market for passenger and international cargo transport in 2014, trailing only China and Brazil.
By 2015, the Vietnamese aviation market will have 34 - 36 million passengers; by
2019, it will reach 52-59 million passengers Besides, the freight will increase to 850,000 - 930,000 tons in 2015 and from 1.4 to 1.6 million tons in 2019
With all above information, Vietnamese aviation promises a huge potential for development In other words, it is a big chance for airline ticket agents in their business.
Problem statement
Airline ticket agents serve as the primary representatives for airlines, responsible for making reservations and selling tickets Their role involves managing high-stress situations as they directly assist customers in resolving travel-related issues Often, they face complaints that are beyond their control, such as flight delays and cancellations Due to the unpredictable nature of air travel, airline ticket agents frequently work during nights, weekends, and holidays to ensure customer satisfaction.
Airline ticket agents face significant work pressure and intense competition from both other agents and airline head offices To differentiate themselves, agencies engage in aggressive marketing campaigns, enhance customer service, and implement innovative sales strategies They often offer attractive deals, increase online advertising, make cold calls, provide 24/7 support, and offer convenient payment options However, airline head offices, traditionally seen as suppliers, have become formidable competitors by reducing agents' commissions and promoting direct bookings through their websites This shift allows airlines to offer flight information and itineraries directly to customers, further diminishing agents' roles in the booking process.
To address these challenges, airline ticket agents are increasingly seeking effective and cost-efficient advertising methods that not only maintain existing customer relationships but also attract new clients.
The positive word-of-mouth (WOM) method is highly effective in influencing customers to switch brands or change their attitudes Research shows that WOM is seven times more impactful than print advertising, four times more effective than direct sales, and twice as effective as broadcast channels (Katz and Lazarsfeld, 1955) Additionally, WOM significantly enhances customer attitudes, with Day (1971) noting it is nine times more productive, while Mangold (1987) estimates its effectiveness at five times greater than other methods.
Word of mouth is recognized as the most effective method of conveying information, particularly in the services sector, as highlighted by Zeithaml et al (1985) Furthermore, Philip Kotler (2007) supports this notion, asserting that service buyers often place greater trust in recommendations from others rather than in traditional advertising when choosing a provider.
Analyzing and identifying the factors that influence positive word of mouth is crucial for airline ticket agents To the best of my knowledge, there has yet to be a study conducted on this topic in Vietnam.
Research Objectives
The overall objective of this thesis is to examine the mediating role of relationship quality in the relations between relationship benefits and word of mouth Specifically:
1 The relationship between confidence benefits and relationship quality
2 The relationship between special treatment benefits and relationship quality
3 The relationship between social benefits and relationship quality
4 The relations between relationship Quality and word of mouth
Research question
This research aims to answer the following questions:
- Do confidence benefits affect to relationship quality?
- Do special treatment benefits impact on relationship quality?
- Do social benefits impact on relationship quality?
- Does relationship quality impact on word of mouth?
Scope and methodology of study
This study will examine the impact of positive word of mouth among customers who have purchased airline tickets directly To gather relevant data, participants will be selected from individuals who have used airline ticket services within the last six months, including their family members and excluding those employed by airlines or ticket agencies The research will take place in Binh Duong and Ho Chi Minh City.
The study will be carried out in two primary phases: first, qualitative research will be utilized to develop and enhance the Vietnamese questionnaire, followed by quantitative research to gather and analyze survey data while estimating and testing the model Specifically, the research will employ SPSS 11.5 software to assess the reliability and validity of measurements using Cronbach’s alpha coefficient and exploratory factor analysis (EFA), as well as to test the model and hypotheses through regression analysis.
Practical significances of the study
The research highlights the importance of airline ticket service providers improving customer relationships by carefully selecting key benefits for their clients This approach not only helps reduce costs but also enhances business efficiency Additionally, the study emphasizes the significant impact of word-of-mouth marketing in driving business success.
Specifically, the study indicates the relations between relationship benefits constructs and relationship quality and between relationship quality and word of mouth
The study also contributes the theories related to word of mouth in the airline ticket service in Vietnamese context
In short, the study has a practical significance for individuals or enterprises that care for or related to airline ticket service.
Structure of the thesis
This chapter outlines the significance of word of mouth in the Vietnamese aviation industry, detailing the problem statement, research questions, and objectives of the study It also defines the scope and methodology employed, highlighting the practical implications of the research.
Chapter 2: Literature review and proposed research model
This chapter offers a comprehensive literature review that evaluates various measurements of relationship quality and word of mouth It will also outline the research framework and propose relevant hypotheses.
Chapter 3: Research methodology This chapter presents the methodology of empirical study
Chapter 4: Data analysis and results
This chapter presents the analysis of gathered data which to test the hypotheses and answer the research questions
Chapter 5: Conclusions and implications This chapter will present implications of this study and will end the research with limitation and suggestions for future research.
LITERATURE REVIEW
Relationship quality & relationship benefits
Numerous studies have explored the concept of relationship quality, identifying key components such as satisfaction, commitment, and trust (Baker et al., 1999; Dorsch et al., 1998; Smith, 1998; De Wulf et al., 2001; Hennig-Thurau et al., 2002) While earlier research, like that of Baker et al (1999), highlighted cooperation as a factor, and Dorschar et al (1998) suggested dimensions like orientation and opportunism, there is a consensus that satisfaction, commitment, and trust are more specific and essential indicators of relationship quality Despite their academic distinction, consumers often struggle to differentiate these constructs, leading them to group them together (Crosby, Evans, & Cowles, 1990; De Wulf et al.).
In service contexts, customers evaluate relationship quality primarily through the lenses of trust, satisfaction, and commitment Research indicates that these three variables are essential for accurately measuring the nature of relationships between customers and service providers Therefore, this study identifies trust, satisfaction, and commitment as the key factors in assessing relationship quality.
According to Zineldin (2006), relationship benefits deliver value that exceeds consumer expectations, leading to significant impressions (Kinard and Capella, 2006; Reynolds and Beatty, 1999) These benefits can influence customer behaviors, including beliefs, loyalties, and commitments, as demonstrated by research from Morgan and Hunt (1994) and Hennig-Thurau et al (2002) Their findings show that increased benefits for users correlate with heightened commitments from providers, highlighting the advantages of relationship benefits in customer-provider interactions.
Relationship benefits significantly enhance relationship quality by improving value perception, which is essential for fostering further connections This enhanced value leads to greater satisfaction and commitment, both critical components of relationship quality Consequently, relationship quality mediates the connection between relationship benefits and word-of-mouth communication Ultimately, the findings suggest that strong relationships are vital for improving overall relationship quality.
According to Gwinner et al (1998), relationship benefits encompass three key dimensions: confidence, special treatment, and social benefits Each of these dimensions significantly contributes to delivering core value to customers.
Confidence benefits are crucial for fostering strong relationships between customers and service providers, as highlighted by various studies (Berry, 1995; Bitner, 1995; Kinard and Capella, 2006) These benefits contribute to a customer's sense of security and comfort, reducing anxiety and clarifying expectations when utilizing services (Gwinner et al., 1998) As customers perceive higher confidence benefits, their commitment to the relationship with service providers strengthens, enhancing the overall perception of relationship quality (Hennig-Thurau et al., 2002; Kim et al., 2006) Furthermore, confidence benefits reflect customers' desires to maintain sustainable relationships and trust in the core services offered (Patterson and Smith, 2001).
H1 Confidence benefits are positively related to relationship quality
Special treatment benefits are crucial in enhancing relationship quality between customers and service providers, encompassing both monetary and non-monetary advantages that exceed customer expectations (Gwinner et al., 1998; Reynolds and Beatty, 1999; Patterson and Smith, 2001; Hennig-Thurau et al., 2002) When service providers engage in high levels of face-to-face interaction and treat each customer as a unique segment, it fosters improved relationship quality (Dolen et al., 2004) Previous studies highlight that these special treatment benefits significantly contribute to customer satisfaction (Reynolds and Beatty, 1999) and commitment (Hennig-Thurau et al., 2002) Additionally, tangible elements such as price reductions or expedited services further enhance relationship quality, as evidenced by Hyun's research (2010) Therefore, it is evident that there is a positive correlation between special treatment benefits and relationship quality.
H2 Special treatment benefits are positively related to relationship quality
Social benefits play a crucial role in enhancing customer satisfaction by fostering long-term relationships with service providers These benefits include recognition by employees, familiarity, and the development of friendships, which are rooted in the emotional connections between customers and providers Strengthening these relationships involves knowing customers' names and preferences, ultimately leading to improved cooperation and enriching the overall customer experience By prioritizing social benefits, service providers can significantly enhance customer loyalty and satisfaction.
Commitment plays a crucial role in mediating the relationship between social benefits and word-of-mouth (WOM) communication, as highlighted by Hennig-Thurau et al (2002) Additionally, personal interactions and friendships between customers and employees have been shown to significantly enhance customer satisfaction, according to research by Price and Arnould (1999) and Gwinner and Gremler (2000).
Previous research indicates a positive correlation between social benefits and relationship quality constructs, including commitment and customer satisfaction with service providers Customers tend to enhance their trust levels when they establish social bonds with employees Consequently, social benefits have a direct and positive impact on relationship quality.
Research by Kim et al (2001) highlights that stronger social relationships with customers enhance relationship quality, equating such bonds to business relationships (Crepiel, 1990) Gremler (2001) advocates for companies to foster tight connections with their customers, while Kinard and Capella (2006) emphasize the essential role of communication in influencing relationship quality These insights lead to the hypothesis that social benefits are positively related to relationship quality.
Relationship quality and word of mouth
Word of mouth (WOM) has been extensively studied, with early research by Arndt (1967) defining it as noncommercial communication between individuals regarding brands, products, or services Later, Westbrook (1987) further elaborated on this concept, highlighting its significance in consumer interactions.
Word-of-mouth (WOM) is defined as informal communication among consumers regarding the ownership, usage, or characteristics of specific goods and services, as well as their sellers According to Anderson et al (1998), WOM involves private parties evaluating products and services, while East et al (2008) describe it as informal advice exchanged between consumers WOM is characterized by its interactive nature, speed, and lack of commercial bias, making it a non-commercial exchange primarily driven by spontaneous senders rather than sponsors (Stern et al., 1994).
Listeners perceive word-of-mouth (WOM) as a trustworthy and valuable source of information due to its non-commercial nature (Day, 1971; Murray, 1991) Research indicates that WOM can be up to nine times more effective than traditional advertising in transforming negative perceptions into positive ones, further supporting this notion (Murray, 1991).
Research from 1991 indicates that recommendations from family, friends, or colleagues are more positively received than traditional advertising methods Shed (1971) highlighted the significance of word-of-mouth (WOM), demonstrating that it is more effective than advertising in capturing customers' attention and influencing their purchasing decisions.
Word of Mouth (WOM) significantly impacts the buying cycle, as highlighted by Britt's research in Sociology since 1966 In the initial stage, WOM influences attention, interest, and expectations (Lynn, 1987; Stock and Zinsner, 1987; Woodside et al., 1992; Webster, 1991; Zeithaml et al., 1993) As consumers progress to the evaluation and conviction phases, they rely on WOM as a trusted information source (Murray, 1991) Ultimately, in the final stage of the buying decision, Magold (1987) demonstrated that WOM's influence surpasses that of other factors, underscoring its critical role in consumer behavior.
Word of Mouth (WOM) significantly influences business outcomes, as noted by File et al (1994) This influence can be a double-edged sword, largely determined by customer satisfaction levels (Bolfing, 1989; Engel et al., 1969; Hartline and Jones, 1996; Richins, 1983; Tybout et al., 1981; Westbrook, 1987) The degree of satisfaction, along with product performance, plays a crucial role in shaping the nature of WOM (Singh and Pandya, 1991; Tanner, 1996; Hartline and Jones, 1996) When customer experiences exceed expectations, they tend to generate positive WOM; conversely, if expectations are not met, negative WOM is likely to arise (Singh and Pandya).
Research has confirmed the presence of two distinct types of relationships across various sectors, including hospitality (Cadotte and Turgeon, 1988), purchasing (Swan and Oliver, 1989), non-profit organizations (Cermak et al., 1991), and health services (Headley and Miller, 1993).
Listeners are often more influenced by negative word-of-mouth (WOM) than positive WOM, as highlighted by Arndt (1967) Notably, negative WOM tends to spread more quickly and widely than its positive counterpart This phenomenon underscores the significant impact of negative feedback in communication.
2:1 (Technical Assistance Research Program, 1986) Only one customer is not content, he could retell it to 9 – 20 people (Desatnick, 1987)
In case of having negative WOM, a business future can be in danger (File et al.,
1994) It is one of three kinds of unhappy customers’ penalty In the research of
In 1970, Hirschman identified the consequences of dissatisfied customers as a form of feedback to providers or a cessation of transactions However, this concept did not account for word-of-mouth (WOM) communication, as the feedback mechanism was limited to a two-way interaction between the provider and the customer.
In 1988, Singh identified the phenomenon of negative word-of-mouth (WOM) by analyzing customer feedback within the context of provider-customer relationships and social networks This concept has since been reinforced by contemporary theorists, who define customers' punitive actions as including providing negative feedback to the provider, severing relationships, or sharing unfavorable opinions within their social networks.
According to Desatnick (1987), more than 90% of dissatisfied customers are unlikely to return to a business, and 13% of these individuals will share their negative experiences with 9 to 20 people This highlights the significant impact of negative word-of-mouth (WOM), which spreads rapidly and can be more influential than traditional advertisements, likening its effect to a destructive fire.
The significance of word of mouth, particularly the impact of negative word of mouth, has been well-documented in numerous studies However, there remains a gap in research regarding the factors that contribute to positive word of mouth.
2.2.2 Relationship quality and word of mouth
Enhancing positive word of mouth (WOM) is crucial for businesses, as research by Gremler et al (2001) shows that fostering strong relationships between customers and employees can significantly boost WOM Customers who perceive high relationship quality with service providers tend to become advocates for the company, sharing their positive experiences (Griffin, 1995; Reynolds and Beatty, 1999) The influence of relationship quality on WOM communication is substantial, with various studies highlighting that key elements such as satisfaction, commitment, and trust play a critical role (Anderson and Sullivan, 1993; Anderson and Weitz, 1989; Dwyer et al., 1987; Hennig-Thurau et al., 2002; Sui and Baloglu, 2003) Additionally, research in the hospitality sector, particularly in hotels and luxury restaurants, confirms that strong relationship quality positively affects WOM (Kim et al., 2001) Therefore, we hypothesize that fostering relationship quality will enhance positive word of mouth.
H4 Relationship quality is positively related to word-of-mouth.
Proposed conceptual model
Based on the ideas derived from the previous researches of Sandy et al (2011) in
In their study on enhancing the service experience, the authors emphasize the importance of generating positive word-of-mouth through key elements such as Confidence Benefits, Special Treatment Benefits, Social Benefits, Functional Quality, Technical Quality, and Relationship Quality These factors collectively contribute to creating a memorable service encounter that encourages customers to share their positive experiences.
Within their framework, technical benefits and functional benefits are two dimensions of service quality (Choi et al., 2004; Ferguson et al., 1999; Gronroos,
The most noticeable advantages for customers are the tangible benefits derived from products or services In contrast, relationship quality encompasses the intangible aspects of customer-provider interactions, reflecting feelings of trust, commitment, and satisfaction.
According to Roberts et al (2003), relationship quality was a distinct construct that significantly differed from service quality, and was a better predictor of behavioral intentions than service quality
In the context of Vietnam, the author has focused exclusively on the relationship quality factor to investigate the positive word of mouth regarding airline ticket services This choice aligns with the constraints of time and resources faced during the study.
H1 Confidence benefits are positively related to relationship quality
H2 Special treatment benefits are positively related to relationship quality
H3 Social benefits are positively related to relationship quality
H4 Relationship quality is positively related to word-of-mouth.
Summary
This chapter reviews previous research on the connections between word of mouth, relationship quality, and relationship benefits These studies provide essential background for the current research Findings indicate that various factors influence relationship quality and word of mouth at multiple levels Notably, Sandy et al (2011) introduced a model highlighting three key factors that affect word of mouth: functional quality and technical quality.
Social Benefits quality and relationship quality which are affected by 3 components of relationship benefits, namely confidence benefits, special treatment benefits and social benefits
This study builds upon the theoretical framework for generating positive word of mouth established by Sandy et al (2011) However, due to constraints related to time and resources, the author has chosen not to include the service quality factors of functional quality and technical quality in this research.
METHODOLOGY
Research design
This study was carried out in two primary phases: first, qualitative research was utilized to develop the foundational questionnaire, and second, quantitative research focused on gathering and analyzing survey data to estimate and validate the research model.
The initial draft of the questionnaire was developed using various theories from prior research on word of mouth, which have been evaluated in the international market This process led to the creation of a set of observed variables for the first draft questionnaire in this study.
To ensure clarity for Vietnamese respondents, the English questionnaire was translated into Vietnamese This translation aimed to facilitate easy and accurate comprehension of the questions The original meaning of the English version was verified by peers from my eMBA class and two close friends proficient in English.
To ensure the questionnaire was appropriate for the respondents, the author organized a focus group consisting of seven clients who had purchased airline ticket services in the past six months This step aimed to verify that the language and content of the measurement scales were relevant and aligned with the Vietnamese market, particularly within the airline ticket agency sector.
Following the implementation of a focus group to draft the questionnaire in the Vietnamese context, a direct survey method was employed, interviewing 100 respondents (n0) who were the subjects of the study This pilot quantitative test aimed to assess the questionnaire's quality and the respondents' understanding and willingness to provide information Additionally, it evaluated the consistency of the measurement scales for potential adjustments Analysis using Cronbach’s alpha and exploratory factor analysis confirmed that the measurements were reliable and valid, with no items omitted Consequently, the final questionnaire was constructed based on these initial scales for the subsequent steps.
Following the pilot study, the main research was carried out through face-to-face interviews with participants from Binh Duong and Ho Chi Minh City who had purchased airline tickets within the past six months The objectives of this study included evaluating the measurement scales, testing the hypotheses, and validating the proposed research model.
Questionnaire design
This study used measurements from existing scales:
The Relationship Benefits Scale consists of three key components assessed through 13 observed variables The first component, confidence benefits, is evaluated using four variables that gauge customer confidence, reduced anxiety, and fulfillment of expectations The second component, special treatment benefits, encompasses five variables that highlight advantages such as discounts, exclusive deals, better pricing, and expedited service Lastly, social benefits are measured through four variables that reflect the relationship between customers and providers, including customer recognition and the development of friendships This scale is grounded in the research of Gwinner et al (1998) and Reynolds and Beatty (1999).
Table 3.1 The measurement scale of relationship benefits
CB1 I have more confidence the service will be performed correctly
CB2 I have less anxiety when I buy the service
CB3 I know what to expect when I go in
CB4 I get the service provider’s highest level of service
ST1 I get discounts or special deals that most customers do not get
ST2 I get better prices than most customers
ST3 The service provider does services for me that they don’t do for most customers
ST4 I am placed higher on the priority list when there is a line or queue for this service
ST5 I get faster service than most customers
SB1 I am recognized by certain employees of the service provider
SB2 I am familiar with the employee(s) who perform(s) the service
SB3 I have developed a friendship with the service provider
SB4 The service provider knows my name
The quality of a relationship is indicative of customers' satisfaction, trust, and commitment levels In this study, the relationship quality scale is based on the measurement frameworks established by Doney and Cannon (1997), Hennig-Thurau et al (2002), and Morgan and Hunt.
(1994) and Oliver (1997) which included six observation variables
Table 3.2: The measurement scale of relationship quality
RQ1 Overall I am satisfied with this service provider
RQ2 My feelings toward this service provider are very positive
RQ3 This service provider can be trusted
RQ4 This service provider is trustworthy
RQ5 My relationship with the service provider is something I am very committed to
RQ6 I believe the service provider and I are both committed to the relationship
Word of mouth reflects customers' positive feedback towards service providers, characterized by their tendency to share favorable comments and make recommendations to others This concept is measured through three observational variables, as outlined by Zeithaml et al (1996).
Table 3.3: The measurement scale of word of mouth
WOM1 I say positive things about the service provider to other
WOM2 I recommend the service provider to someone who seeks my advice
WOM3 I encourage friends and relatives to do business with the service provider
The questionnaire in this study includes 3 main sectors
The initial section of the survey aims to filter out respondents who may not be suitable for participation Specifically, it targets customers who have purchased and used airline ticket services within the last six months Individuals who have never bought an airline ticket cannot provide relevant insights regarding the agent’s services Additionally, if respondents purchased tickets a long time ago, they may struggle to recall their experiences with service employees Moreover, individuals or their family members employed in airlines or ticket agencies are excluded from the survey to prevent bias, as they may provide overly positive feedback about their own or their relatives' businesses, even in cases of poor service.
The second section of the study focused on measurement scales assessing various factors, including confidence benefits, special treatment benefits, social benefits, relationship quality, and word of mouth Each factor was evaluated using a 7-point Likert scale, ranging from "totally disagree" to "totally agree."
The third sector related to private questions of respondents such as gender, age, monthly income and educational level.
Sample Design
According to Hair et al (1998), the exploratory factor analysis (EFA) needs to collect data sample size at least 5 respondents per item
To conduct regression analysis, Tabachnick & Fidell (1996) established the sample size formula n ≥ 8m + 50, where n represents the sample size and m denotes the number of observation variables Based on this calculation and the author's research capabilities, a sample size of 220 was deemed appropriate for this study.
To achieve the sample size (n= 220), over 500 questionnaires were interviewed with non-probability sampling methods Unsuitable questionnaire was gradually eliminated because of the inappropriate respondents or the blank answers.
Methods of data analysis
After collecting, the unqualified questionnaires were eliminated, otherwise kept, encoded, entered, cleaned and analyzed by SPSS 11.5 software
The analysis involved assessing the reliability of the measurement scale using Cronbach’s Alpha and evaluating its validity through exploratory factor analysis Additionally, regression analysis was employed to test the model and hypotheses effectively.
Research Process
The research process showed in Figure 3.1
Main study (n"0) The formal questionnaire
Summary
The study employed a mixed-methods approach, utilizing both qualitative and quantitative research techniques Qualitative insights were gathered through a focus group consisting of seven customers, while quantitative data was collected through an initial survey of 100 customers to refine the formal questionnaire The formal quantitative research was conducted with a larger sample size to ensure robust findings.
The study surveyed individuals in Ho Chi Minh and Binh Duong City who purchased airline tickets within the last six months A questionnaire featuring 22 observation variables was employed, utilizing a 7-point Likert scale Following data cleaning, the analysis focused on assessing the reliability and validity of the measurement scales through Cronbach's alpha coefficients and exploratory factor analysis, as well as testing the proposed model and hypotheses All analyses were conducted using SPSS 11.5 software.
DATA ANALYSIS AND FINDINGS
Descriptions of sample
The respondents in this study were Vietnamese, who have experienced airline ticket service in the past six month
The 220 qualified respondents included 76 male respondents (36.8%), and 144 female respondents (63.2%) Most of them aged from 15 to 34 years old (96.4%)
The income distribution among respondents revealed that 43.6% earned below VND 5 million, while 39% earned between VND 5 million and VND 10 million Additionally, 17.7% fell within the VND 10 million to VND 15 million range, with only 2.7% earning between VND 15 million and VND 20 million, and none exceeding VND 20 million In terms of education, the majority of respondents, 67.7% (149 individuals), held a college or graduate degree, followed by 28.6% (63 individuals) with postgraduate qualifications, and a small percentage, 3.7% (8 individuals), with only high school or vocational school education.
Table 4.1 Statistical report of demographic variables
Reliability and validity of the measurement scale
In this study, the reliability and validity of the measurement scale were evaluated by two methods: Cronbach alpha coefficient analysis and Exploratory Factor Analysis (EFA)
The reliability of scales was evaluated using Cronbach’s Alpha, a tool designed to determine whether the questions effectively measure theoretical concepts This assessment aimed to eliminate irrelevant items or scales, ensuring the integrity of the measurement process.
To ensure coherence among questions measuring the same concept, the Cronbach's alpha coefficient serves as a statistical test to assess the correlation level among items within a scale A measurement scale is considered adequately reliable if the Cronbach's alpha coefficient ranges from 0.70 to 0.80, while a reliability threshold is accepted at 0.60 or higher, as outlined by Nunnally & Bernstein.
1994), cited in Nguyen Dinh Tho 2011)
While the correlation coefficient among the scale's items was satisfactory, the contribution of each question varied, with some performing better than others Cronbach's alpha coefficients effectively measured the association among variables within the same scale, but fell short in providing insight into which questions should be retained or removed.
To address measurement issues, it is essential to focus on item-total correlations; questions that contribute less to the overall measurement concept typically exhibit weak correlations with the total score It is advisable to retain items that show strong correlations with the total According to Nguyen Dinh Tho (2011), any variables with an item-total correlation below 0.3 are considered to have poor evaluation quality.
The analysis using Cronbach’s Alpha on five components—Confident Benefit (CB), Social Benefits (SB), Special Treatment Benefits (ST), Relationship Quality (RQ), and Word of Mouth (WOM)—demonstrated adequate reliability, as shown in Table 4.2 Each component's Cronbach’s alpha coefficient exceeded 0.6, with Corrected Item-Total correlations greater than 0.3 The overall results indicated strong measurement scale reliability, with the lowest alpha being 0.7811 for Confidence Benefits Furthermore, the alpha values remained consistent even when individual items were considered for deletion, leading to the conclusion that no items were excluded from the analysis.
The study assessed various benefits through specific measurement items, revealing that Confidence Benefits (CB) were evaluated using four items (CB1 - CB4) with a Cronbach alpha of 0.7811 Special Treatment Benefits (ST) were measured by five items (ST1 - ST5), achieving a high Cronbach alpha of 0.9288 Additionally, Social Benefits (SB) were gauged through four items (SB1 - SB4), resulting in a Cronbach alpha of 0.9032 Finally, Relationship Quality (RQ) was assessed with six items (RQ1 - RQ6), further emphasizing the reliability of the measurements.
Cronbach alpha of 0.8708 Word of Mouth (WOM) was measured by 3 items (WOM1 - WOM3) and had a Cronbach alpha of 0.8753
Scale Variance If Item deleted
Cronbach’s Alpha If Item Deleted Confidence Benefits (CB): Cronbach Alpha = 0.7811
Special Treatment Benefits (ST): Cronbach Alpha = 0.9288
Social Benefits (SB): Cronbach Alpha = 0.9032
Relationship Quality (RQ): Cronbach Alpha = 0.8708
Word of Mouth (WOM): Cronbach Alpha = 0.8753
The Cronbach alpha coefficients indicate that the measurement scales, which include all 22 observed variables, meet the necessary reliability standards Therefore, they are suitable for exploratory factor analysis in the subsequent phase to assess the validity of the scale.
To ensure the accuracy of the measurement scales, it was essential to evaluate their validity in addition to the reliability assessed by Cronbach's alpha coefficients The author employed the Exploratory Factor Analysis (EFA) method to verify the validity of the scale.
Exploratory Factor Analysis (EFA) is a statistical technique that simplifies a large set of variables by extracting a smaller number of factors composed of homogeneous variables This method retains the essential information from the original data while enhancing interpretability EFA is also employed to evaluate the convergence level of these components, ensuring a coherent understanding of the underlying relationships within the data.
When analyzing exploraroty factors, researchers concerned about some standards:
Barlett’s Test of sphericity is a statistical test designed to evaluate the hypothesis that there are no correlations among the variables within a population (Hoang Trong and Chu Nguyen Mong Ngoc, 2008).
The Keiser-Meyer-Olkin (KMO) measure of sampling adequacy is an important index used to assess the suitability of factor analysis A high KMO value, ranging from 0.5 to 1, indicates that factor analysis is valid, while a lower value suggests that it may not be appropriate for the data.
About analysis procedure, the extraction method applied Principal Components with Varimax rotation method, accepting all factors that had Eigenvalues over 1
4.2.2.1 EFA implementation for relationship quality measurements:
The EFA analysis demonstrated adequacy, evidenced by a KMO value of 0.847, exceeding the threshold of 0.5, and a Chi-Square value of Bartlett’s test at 3180.693 with a significant p-value of 0.000, which is below the 0.05 level.
The exploratory factor analysis (EFA) revealed the extraction of four factors with an eigenvalue of 1.123, accounting for a total variance of 70.992% (see Table 4.3) Consequently, these four factors collectively explain 70.992% of the data's variation.
The Rotated Component Matrix revealed a clear grouping of 19 variables into four distinct components: special treatment benefit, relationship quality, social benefits, and confidence benefits, with high concentration factors aligning with the original plan Notably, the observation variances correlated strongly with the initial factors, and all factor loadings exceeded 0.5, indicating adequate correlations between variables and factors, thus confirming the retention of all items.
Table 4.3 EFA results of relationship quality scale:
Extraction Method: Principal Component Analysis
Rotation Method: Varimax with Kaiser Normalization a Rotation converged in 7 iterations
In short, based on the criteria of Exploratory Factor Analysis, the scales were valid
22 observation variables were grouped into 4 components as origin scales that were Special Treatment Benefit (ST), Relationship Quality (RQ), Social Benefits (SB) and Confidence Benefits (CB)
4.2.2.2 EFA implementation for WOM scale:
The exploratory factor analysis (EFA) of the Word of Mouth (WOM) scale demonstrated its appropriateness, as indicated by a KMO value of 0.743, exceeding the 0.5 threshold, and a significant Chi-square value of 344.145 (p < 0.05) from Bartlett’s Test The EFA revealed a single factor with a total variance extraction of 80.531% and an eigenvalue of 2.416 All three variables (WOM1, WOM2, and WOM3) showed factor loading coefficients above 0.5, confirming their grouping into one component without any variable rejection This indicates that the WOM scale is suitable for the study and can be utilized for subsequent analyses.
Table 4.4 EFA results of word of mouth scale:
Bartlett's Test of Sphericity Approx Chi-Square: 337.889
Extraction Method: Principal Component Analysis
Testing the research model and the hypotheses
The initial phase of multiple linear regression (MLR) analysis involved assessing the linear correlations among all variables This process examined the relationships between independent variables and the dependent variable, as well as the correlations among the independent variables themselves To quantify these linear relationships, the Pearson’s correlation coefficient was utilized, indicating that a coefficient value approaching ±1 signifies a strong correlation between the two quantitative variables.
The correlation coefficients could help to guess the linear correlation between variables
The results of the Pearson correlation analysis of this study were presented in Table 4.5
** Correlation is significant at the 0.01 level (2-tailed)
The correlation analysis results presented in Table 4.5 indicate that the dependent variable of relationship quality exhibits a significant linear correlation with three independent variables: Confidence Benefits, Social Benefits, and Special Treatment Benefits, all at a significance level of 0.000, which is below 0.01 Notably, the correlation coefficient between relationship quality and Confidence Benefits is substantial at 0.502, while the correlations with Social Benefits and Special Treatment Benefits are lower, at 0.358 and 0.292, respectively Despite these varying levels of correlation, it can be inferred that all independent variables are relevant for inclusion in the model to further investigate relationship quality.
There is a significant correlation of 0.670 between Special Treatment Benefits and Social Benefits, indicating potential multicollinearity Consequently, these two variables must be analyzed with caution in the multiple linear regression examining the relationship quality as the dependent variable and confidence benefits, social benefits, and special treatment benefits as independent variables.
A strong correlation exists between relationship quality and word of mouth, with a coefficient of 0.695 This indicates that relationship quality serves as a significant independent variable in a simple linear regression model, effectively explaining the dependent variable of word of mouth.
4.3.2 Multiple linear regression analysis of relationship quality scale:
Multiple linear regression was utilized to examine the relationship between two or more independent variables and a single quantitative dependent variable Additionally, this analysis facilitated the testing of the study's hypotheses.
The relationship quality (RQ) is significantly influenced by three independent factors: confidence benefits (CB), special treatment benefits (ST), and social benefits (SB), as identified through a multiple regression analysis.
The model's significance was assessed to determine if the deviations in the dependent variable were adequately explained by the independent variables Using an F-test, the hypothesis β1 = β2 = β3 = 0 was evaluated, revealing that a p-value less than 0.05 indicated a significant model at a 95% confidence interval Conversely, a p-value greater than 0.05 suggested that the model was not significant, implying that the independent variables did not support predictions of the dependent variable As shown in Table 4.5, the hypothesis was rejected with a significance value of 0.000, which is below 0.05, confirming that the model fit the data with a 95% confidence interval.
Squares df Mean Square F Sig
Total 132.281 219 a Predictors: (Constant), ST, CB, SB b Dependent Variable: RQ
Second, the result in table 4.7 showed the value of R 2 = 0.350 (#0) and adjusted
The adjusted R² value is 0.341, indicating that the Special Treatment Benefits (ST) variable does not significantly enhance the explanation of the independent variable, Relationship Quality (RQ) This suggests that approximately 34.1% of the variance in Relationship Quality is accounted for by the independent variables analyzed However, it also implies that there may be other undiscovered factors influencing Relationship Quality that were not included in the multiple regression analysis Using the adjusted R² helps prevent overestimating the model's explanatory power.
Std Error of the Estimate
1 592(a) 350 341 63084 a Predictors: (Constant), ST, CB, SB
In multiple linear regression models, it is essential that independent variables are not perfectly correlated, as this can lead to multicollinearity issues Multicollinearity occurs when independent variables are closely related, making it challenging to assess their individual impacts on the dependent variable Increased multicollinearity can result in unstable coefficient estimates and inflated standard deviations, rendering the coefficients less significant (Hoang Trong & Chu Nguyen Mong Ngoc, 2008) To evaluate multicollinearity, the variance inflation factor (VIF) is employed; a VIF above 10 indicates that the variable contributes little to explaining the variance in the dependent variable (Hair et al., 2006) However, a VIF greater than 2 should still be considered significant in the context of multiple linear regression coefficients (Nguyen Dinh Tho, 2011).
The result in table 4.8, satisfied requirements with all VIFs below 2 (VIF of CB 1.012, VIF of ST = 1.819, VIF of SB = 1.819)
Standardized Coefficients t Sig Collinearity Statistics
To ensure the reliability of the regression results, key assumptions of linear regression were examined, including the linearity of the relationship between dependent and independent variables, independence of errors (absence of serial correlation), homoscedasticity (constant variance of errors), and normality of the error distribution Violations of these assumptions, such as nonlinearity, serial correlation, heteroscedasticity, or non-normality, could lead to inefficient, biased, or misleading forecasts and economic insights However, as indicated in Appendix 7.2, all assumptions were satisfied, confirming the reliability of the regression results.
A multiple regression analysis established a linear relationship between the dependent variable of relationship quality and independent variables, including confidence benefits, special treatment benefits, and social benefits With a 95% confidence interval, factors that demonstrated a statistically significant level below 0.05 were found to impact relationship quality Conversely, certain variables may not adequately explain the dependent variable of relationship quality.
Table 4.8 reveals that the variables CB and SB exhibited significant values below 0.5, indicating their statistical significance in the regression model and their impact on RQ In contrast, the Special Treatment Benefits variable was found to be inappropriate for explaining the model, as its significance value of 0.376 exceeded the 0.05 threshold, suggesting that relationship benefits were not accounted for by special treatment benefits.
Thus, the relationship between “Relationship Quality” and the independent factors was demonstrated by the following equation:
In short, the regression results indicated that relationship quality was influenced by
2 factors, namely Confidence benefits (CB) and Social Benefits (SB), and not be influenced by Special Treatment Benefits (ST), specifically:
- Confidence benefits had a positive and significant impact on relationship quality (β=0.469, p