INTRODUCTION
Research Background
This research examines the key factors influencing Vietnamese consumers' intention to use mobile content services (MCS), highlighting the significance of continuous information technology (IT) development and the acceptance and use of IT These themes are crucial for the business sector, as the ongoing IT revolution is reshaping business processes and practices (Mahabir & Geeta, 2013).
In recent years, advancements in information technology have significantly accelerated the growth of mobile technology, impacting daily life as the global number of mobile devices reached 4.6 billion (CBSnews, 2010) According to Gartner (2011), mobile applications were projected to generate $15.9 billion in end-user spending in 2012, while also stimulating related sectors like advertising, device sales, and innovation in mobile technology By delivering context-aware features that enhance user experiences, mobile content services have transformed the marketplace, attracting interest from various stakeholders including device manufacturers, merchants, mobile app developers, and marketing agencies.
In the second quarter of 2013, Vietnam saw the shipment of 5.8 million mobile phones, reflecting a strong market for smartphones and high mobile subscriber penetration at 174%, second only to Singapore and Thailand in Southeast Asia The Vietnamese market is in a mature state for laptops, while smartphones and tablets show significant growth potential Monthly, there are 2.1 billion advertisements served to mobile devices in Vietnam, translating to over 805 ads every second, with 62% of internet users accessing the web via mobile Additionally, the number of mobile internet users in Vietnam reached 19 million, highlighting the rapid expansion of mobile connectivity in the country.
The anticipated growth of mobile money services is set to be a major trend in the coming years, offering numerous benefits for users globally This evolution is expected to significantly influence the telecommunications, technology, and financial services sectors.
The growing importance of mobile devices and content services in Vietnam highlights the necessity of examining mobile content services.
Technology is significantly transforming service delivery and customer interactions, particularly in the complex landscape of consumer behavior in the technology market Understanding consumer acceptance of new technologies, such as mobile data content services, poses challenges for researchers The substantial financial investments in mobile technology infrastructure can lead to losses for companies if they fail to attract enough mobile clients While successful Vietnamese mobile content service providers like Mobifone, VinaPhone, and Viettel have emerged, others like Beeline and Hanoi Mobile have exited the market Despite notable advancements in Vietnam's mobile market over the past two years, mobile content services have not fully realized their potential, and research on consumer intention and behavior remains limited.
This study examines the factors influencing consumer behavioral intentions towards mobile content services, specifically within the context of the evolving mobile market in Vietnam.
Research Motivation
Mobile content services offer convenient access to information for users of various mobile devices, catering to their diverse needs Despite the proliferation of new mobile services, understanding their acceptance remains a critical issue Numerous studies have investigated the acceptance of mobile technology, including significant research on mobile banking Key studies in this area include those by Suoranta (2003), Cheong & Park (2005), and Sripalawat, Thongmak, and Ngramyarn (2011), highlighting the ongoing interest in how users embrace these technologies.
In 2009, the term "Mobile Content Services" emerged, encompassing various categories such as mobile communication, commercial, entertainment, and information services (Ying & Shieh, 2009) Despite its growing relevance, there is a lack of research exploring the factors influencing customer behavior towards mobile content services, particularly within the Vietnamese market.
This study aims to identify the key factors influencing Vietnamese consumers' behavioral intentions to use mobile content services Additionally, it offers insights and strategies for mobile service providers to enhance their opportunities and ensure sustainable growth in Vietnam's mobile content services market.
Research Objectives
This study aims to analyze the factors influencing customers' behavioral intentions to utilize mobile content services, thereby deepening our understanding of Vietnamese mobile consumers in the context of technology services The specific objectives include identifying key determinants that impact usage intentions and enhancing insights into consumer behavior in Vietnam's mobile market.
(1) To examine the causal factors that affect consumers’ behavioral intention to use mobile content services;
(2) To examine the important effects of cognitive and affective attitude on behavioral intention to use mobile content services.
Research Scopes
This study targets Vietnamese mobile users, specifically in Ho Chi Minh City (HCMC), the largest city in southern Vietnam and a key economic, technological, and financial hub HCMC attracts migrants from various provinces, contributing to its significant population growth, which was recorded at 7,990,100 in 2013 (General Statistics Office, 2013) Consequently, the city boasts a higher number of mobile users compared to other regions in Vietnam.
This study exclusively targets individual mobile clients in Vietnam, excluding institutions or groups that also utilize various mobile content service providers Furthermore, the research is concentrated solely on the factors influencing behavioral intentions to use mobile content services, intentionally omitting other dependent variables such as "use behavior" and "actual use."
Significance of the Research
This study enhances the understanding of the attitude change process, offering valuable insights for researchers analyzing this phenomenon It highlights the significant factors influencing behavioral intentions toward mobile content services in the Vietnamese market For managers, these findings provide guidance on managing attitude shifts, ultimately improving user acceptance of technology Additionally, the results offer practical implications for mobile operators, enabling them to boost the usage of mobile content services effectively.
Research Methodology
This study employs both qualitative and quantitative research methods, consisting of a pilot study and a main study The pilot study utilizes a combination of these methods to develop and refine a draft questionnaire, while the main study focuses solely on quantitative methods Final questionnaires are distributed to respondents through email, social networks, and hard copies, utilizing convenience sampling to streamline the sampling process Data analysis is conducted using tools such as IBM SPSS 22, IBM AMOS 22, and Microsoft Excel 2012, with further details of the research methodology provided in Chapter 3.
Structure of Research
The structure of research consists of five chapters:
This chapter provides an overview of mobile content services, detailing the motivation and objectives behind the research It also outlines the research scope, methodology, and structure, setting the stage for a comprehensive exploration of the topic.
Chapter 2: Literature Review and Research Model
In this chapter, the author provides the literature reviews that concern the mobile content services, theoretical models and competitive model Finally, the research model is presented
This part presents brief description of the research methodology, which includes the research process, construct measurements, measurement refinement, main study and data analysis methods
Chapter 4: Data Analysis and Results
Chapter 4 presents the sample specification and data analysis Based on the results, the author draws conclusions for the proposed hypotheses
Chapter 5: Conclusions, Implications and Limitations
This chapter presents the main conclusions, implications and limitations Moreover, the recommendations for future research are also provided.
Summary
This research explores the background of mobile content services, emphasizing the necessity of studying the factors influencing consumer behavioral intentions, particularly in Vietnam, where existing studies predominantly focus on mobile applications It identifies the gaps in current literature regarding mobile content services and targets individual mobile users in Ho Chi Minh City The article outlines the research objectives, methodology, and structure, leading into a literature review that will establish the research model and develop hypotheses.
LITERATURE REVIEW AND RESEARCH MODEL
Mobile Content Services
Mobile Content Services enable customers to engage with various types of content through their mobile devices These services are offered by providers and delivered directly to users' mobile devices or in other formats Accessed via public telecommunications networks, payment for these services can be made instantly using credit cards or through methods like post-payment on the user's phone bill or by deducting from a prepaid mobile account.
Mobile content services are a product of technological advancements stemming from the information technology revolution These services, closely intertwined with mobile applications, have emerged due to the convergence of the Internet, media, social media, information technology, and telecommunications Unlike traditional voice services, mobile content services encompass a range of digital offerings integrated into mobile networks and devices, providing a diverse array of benefits to users (Bouwman et al., 2012).
The convergence of technology has led to the interchangeable use of terms like Mobile Content Services, Mobile Data Services, and Mobile Applications in various studies According to Fang and colleagues (2006), mobile services can be categorized into three types based on their objectives.
(1) General tasks: These tasks do not involve transaction and gamming such as mobile email, mobile SMS, browse website, map and search location services
(2) Transaction tasks: These tasks include mobile banking, mobile money and online purchase via internet-store
Entertainment tasks encompass a variety of gaming and multimedia services, including mobile games, voting and contests through value-added services, polyphonic ringtones, logo downloads, wallpapers, and music streaming via mobile networks, along with standby background music options.
In Vietnam, there are three distinct types of tasks, each with its own objective General tasks focus on information retrieval and communication with others, while transaction tasks are aimed at executing financial transactions In contrast, entertainment tasks are designed to provide enjoyment to their participants.
The International Telecommunication Union (ITU) categorizes mobile data services into four main types: communication services, information content services, entertainment services, and commercial services Communication services, the most prevalent, encompass SMS, MMS, emails, and mobile chatting Entertainment services offer a variety of options, including ringtones, digital characters, horoscope readings, mobile gaming, videos, and music Information content services provide users with essential updates such as weather news, maps, sports news, traffic information, and news headlines Lastly, commercial services facilitate online financial transactions, bookings, shopping, and payments All these categories are available in Vietnam.
Theory of Reasoned Action
Figure 2.1 shows a model of the Theory of Reasoned Action (TRA), which is proposed by Fishbein and Ajzen (1975)
Figure 2.1 The Theory of Reasoned Action model (Ajzen & Fishbein, 1980)
The Theory of Reasoned Action (TRA) aims to predict and understand individual behavior by analyzing prior intentions and personal beliefs (Ajzen & Fishbein, 1980) According to TRA, a person's intention is influenced by two main factors: their attitude towards the behavior and the subjective norms surrounding it To quantify a person's attitude (A) towards a specific behavior, one can calculate the sum of the products of all salient beliefs—representing the perceived consequences of the behavior—and the evaluation of those consequences, leading to a comprehensive understanding of behavioral intentions.
The subjective norm (SN) is calculated by summing the products of an individual's normative beliefs (n), which reflect the perceived expectations of others, and their motivation to comply This formula effectively measures the subjective norm in relation to actual behavior.
SN = ∑ Hence, the individual behavior intention (BI) can be determined by one formula as below:
TRA provides a useful model that can explains and predicts the actual behavior of an individual fairy well.
Theory of Planned Behavior
Ajzen (1985) expanded the Theory of Reasoned Action (TRA) by introducing the Theory of Planned Behavior (TPB), which incorporates an additional construct: perceived behavioral control (PBC) PBC reflects an individual's perception of their ability to perform a specific behavior and is influenced by two key beliefs: control beliefs and perceived facilitation Control beliefs pertain to the perceived availability of necessary skills, resources, and opportunities, while perceived facilitation involves personal evaluations of the resources available to achieve desired outcomes (Mathieson, 1991) The TPB model is visually represented in Figure 2.2.
Figure 2.2 Theory of Planned Behavioral (Matheison, 1991)
Technology Acceptance Model
The Technology Acceptance Model (TAM) builds on Fishbein and Ajzen’s Theory of Reasoned Action (TRA) to elucidate the causal relationships among its variables, as outlined by Davis, Bagozzi, and Warshaw in 1989 For a visual representation, refer to Figure 2.3, which showcases the initial modified version of TAM.
Figure 2.3 First modified version of TAM (Davis et al, 1989)
This version posits that technology acceptance can be explicated by two variables:
Perceived usefulness and perceived ease of use are critical factors influencing technology acceptance Perceived usefulness refers to the extent to which an individual believes that a system can enhance their job performance, while perceived ease of use indicates the level of effortlessness associated with using the system According to Davis (1989), these two variables play a central role in shaping user acceptance Notably, there are instances where a person's strong intention to use a system is driven directly by its perceived usefulness, bypassing the need for a favorable attitude towards it.
Davis et al (1989) use above model to deploy a study with 107 users to measure their intention to use system after one-hour introduction about the system and repeat
Fourteen weeks later, the findings indicate that both "perceived usefulness" and "perceived ease of use" directly impact the behavioral intention to use technology, thereby excluding the attitude construct from the model Refer to Figure 2.4 for the finalized version of the Technology Acceptance Model (TAM).
Figure 2.4 Final version of TAM (Venkatesh & Davis, 1996)
2.4.1 Revised Original TAM with Separate Affective and Cognitive Attitude
The Technology Acceptance Model (TAM), initially based on Fishbein and Ajzen’s Theory of Reasoned Action (TRA), originally included attitude as a key construct to explain the relationships between variables (Davis et al., 1989) However, in the latest version of the model, Davis and his colleagues removed the attitude construct (Figure 2.4) Despite this omission, numerous studies, such as those by Agarwal and Prasad (1999), Lu, Yao, and Yu (2005), and Curran and Meuter, have continued to utilize the original TAM that incorporates attitude.
(2005) Thus, it is extremely difficult to compare these studies with contradictory findings about attitude since consistent measures of attitudes are not used across studies
The Theory of Reasoned Action (TRA) and the Technology Acceptance Model (TAM) both address the concept of attitude, with TAM viewing it as a unidimensional affective construct In contrast, Cacioppo, Petty, and Crites (1994) argue that attitudes are best classified through both affect and cognition The affective dimension reflects the emotional attraction a person feels towards an object, while the cognitive dimension encompasses the individual's beliefs, evaluations, and perceptions of that object based on their values.
Yang and Yoo (2004) argue that attitude significantly impacts the use of information systems, suggesting that it should be re-evaluated within the Technology Acceptance Model (TAM) They propose retaining the attitude construct, unlike the approach taken by Davis et al (1989), and recommend incorporating two dimensions of attitude: cognitive and affective.
The cognitive dimension of attitude significantly impacts an individual's use of information systems, while the affective dimension should be considered a distinct variable Specifically, cognitive attitudes reflect the anticipated performance of the system, whereas affective attitudes relate to the system's appeal and usability (Zaad & Allouch, 2008).
Figure 2.5 TAM with Affective and Cognitive Attitude (Yang & Yoo, 2004)
Wang and Liu (2009) developed a conceptual model that integrates the Technology Acceptance Model (TAM) from Davis et al (1989) and Yang and You’s (2004) research, incorporating behavioral intention as a mediator between attitude and usage Their case studies reveal that both cognitive and affective attitudes significantly positively influence the behavioral intention to use Taiwan's Railway Internet Ticket System.
Figure 2.6 Revised TAM with Behavioral Intention, Affective and Cognitive
Yang and Yoo (2004) highlight that the affective dimension of attitude is shaped by beliefs, which can be categorized as either evaluative or non-evaluative (true or false) They explain that cognitive attitudes arise from evaluative beliefs, which are in turn influenced by non-evaluative beliefs and values Consequently, evaluative beliefs evolve into affective attitudes, manifesting as feelings of like or dislike.
The relationship among the constructs of affective attitude, cognitive attitude, non-evaluative beliefs, and values is hierarchical Specifically, cognitive attitude positively influences affective attitude, as demonstrated by the empirical findings of Yang and Yoo (2004), which highlight the significance of values in shaping non-evaluative beliefs.
Research indicates that "attitude" can influence behavioral intention beyond its immediate effects Numerous studies, including the original Technology Acceptance Model (TAM) by Davis (1986) and subsequent models by Taylor and Todd (1995a) and Morris and Dillon (1997), have explored the positive relationship between attitude and behavioral intention.
2.4.2 Perceived Convenience – An External Variable of TAM
Convenience for consumers is primarily determined by the effort and time required to use a product or service (Berry, Seiders & Grewal, 2002) A product or service is deemed convenient if it reduces the emotional, cognitive, and physical burdens on the user (Chang, Yan & Tsen, 2012) Brown (1990) identifies five key elements of convenience: time, acquisition, use, execution, and place, while Yoo and Kim (2007) specifically measure the perceived convenience of wireless networks based on time, place, and execution They define perceived convenience as the level of ease users experience when utilizing wireless networks to accomplish tasks Furthermore, their research indicates that while perceived convenience does not directly influence the intention to use a service, perceived ease of use positively impacts perceived convenience, which in turn enhances perceived usefulness.
In a study by Cheolho and Sanghoon (2007), four key constructs—perceived usefulness, perceived ease of use, behavioral intention, and perceived convenience—were analyzed within a ubiquitous wireless LAN environment The findings revealed that perceived ease of use has a positive impact on perceived convenience, which in turn positively influences perceived usefulness.
Recent research by Chang et al (2012) highlights the positive impact of perceived ease of use on perceived convenience in English learning through personal digital assistants (PDAs) Additionally, it shows that perceived convenience enhances perceived usefulness and positively influences users' attitudes toward using PDAs for learning.
Perceived Mobility
The rapid growth of mobile technology has made mobile content services uniquely beneficial due to their inherent mobility, allowing users to access content anytime and anywhere This accessibility empowers individuals to manage their work and entertainment in diverse ways In this study, perceived mobility refers to the extent to which mobile content services are seen as providing timely and pervasive connections, which can either hinder or encourage their usage Research by Hong, Thong, Moon, and Tam (2008) suggests a positive relationship between perceived mobility and consumers' intention to continue using mobile content services.
2002) theorizes that mobility such a factor is likely to affect the formulation of behavior intention Amberg, Hirschmeier, and Wehrmann (2003) propose that perceived mobility is a construct specific to mobile services
Insufficient network signal coverage, weak device batteries, and a lack of mobile operators can hinder mobility and user satisfaction Krueger (2001) predicted a growing demand for "payment roaming," driven by users seeking cooperative solutions for processing payments while traveling outside their network or across different networks Buhan, Cheong, and Tan (2002) emphasized that effective solutions should integrate with existing systems to establish a global payment network.
Research Model and Hypothesis Development
This research proposes a theoretical model that integrates the revised Technology Acceptance Model (TAM) with distinct cognitive and affective attitudes, as well as perceived convenience and mobility The model highlights the specific relationships among the TAM constructs and the identified variables, as illustrated in Figure 2.7.
Figure 2.7 The proposed theoretical model
The Technology Acceptance Model (TAM) illustrates the connections among perceived ease of use, perceived usefulness, attitudes toward technology, and behavioral intention Specifically, it posits that perceived ease of use enhances perceived usefulness, while both perceived ease of use and perceived usefulness contribute positively to attitudes toward using technology Ultimately, a favorable attitude toward technology influences the intention to engage with it (Davis, 1986).
Regarding perceived ease of use positively affects perceived usefulness; there are many empirical tests, such as Davis (1986), Yang and Yoo (2004), Wang and Liu
(2009), which prove that users perform well in tasks when they do not need to pay much effort Therefore, hypothesis H1 is proposed as follows:
H1: Perceived ease of use positively affects perceived usefulness
Research by Yang and Yoo (2004) indicates that attitudes toward technology encompass both affective and cognitive components, a classification supported by Zaad and Allouch (2008) and Petty et al (1994) Numerous studies highlight the mediating roles of these two attitudes in relation to perceived usefulness, perceived ease of use, and behavioral intention, including works by Wang and Liu (2009) and Alhabahba et al (2012) This study aims to explore the causal relationships between the belief constructs in the Technology Acceptance Model (TAM)—perceived usefulness and perceived ease of use—and the attitude constructs of affective and cognitive attitudes, ultimately influencing behavioral intention Consequently, hypotheses are proposed to illustrate the relationships between these newly added attitude constructs and other variables within the TAM framework.
H2: Perceived usefulness positively influences cognitive attitude
H3: Perceived usefulness positively influences affective attitude
H4: Perceived ease of use positively influences cognitive attitude
H5: Perceived ease of use positively influences affective attitude
Numerous studies, including those by Taylor and Todd (1995a), Morris and Dillon (1997), and Davis (1986), have confirmed that attitudes significantly influence behavioral intentions Yang and You (2004) suggest that the dyadic view posits affective and cognitive factors as independent variables impacting behavioral intention Recent research on the Railway’s Internet Ticket System by Wang and Liu (2009) further supports this, indicating that both cognitive and affective attitudes positively affect behavioral intention, with a stronger beta coefficient from cognitive attitudes Consequently, employing both cognitive and affective attitude constructs rather than a single attitude leads to the formulation of new hypotheses.
H6: Cognitive attitude positively influences behavioral intention to use mobile content services
H7: Affective attitude positively influences behavioral intention to use mobile content services
Research indicates that cognitive attitudes positively impact affective attitudes (Yang & You, 2004) Empirical evidence from Wang and Liu (2009) supports this relationship, demonstrating that cognitive beliefs can shape clients' emotional responses Therefore, it is hypothesized that evaluative beliefs, or cognitive attitudes, contribute to the development of clients' affective attitudes.
H8: Cognitive attitude positively influences affective attitude
This chapter highlights that perceived convenience is an external variable within the Technology Acceptance Model (TAM) Research by Yoon and Kim (2007) and Chang et al (2012) indicates that perceived ease of use positively influences perceived convenience, which in turn enhances perceived usefulness Specifically, as mobile systems become easier to use, users perceive them as more convenient, leading to a greater perception of their usefulness Building on these findings, this study identifies three dimensions of perceived convenience: place, time, and execution, and proposes corresponding hypotheses.
H9: Perceived ease of use positively affects on perceived convenience
H10: Perceived convenience positively affects on perceived usefulness
In reviewing perceived mobility, it emerges as a significant variable within the Technology Acceptance Model (TAM), as it plays a crucial role in influencing user acceptance of systems (Hong et al., 2008; Amberg et al., 2003; Ajzen).
2002) Therefore, this study hypothesizes that perceived mobility might have a positive relationship with user’s behavioral intention to use mobile content services
H11: Perceived mobility has a positive influence on consumers’ behavioral intention to use mobile content services
The summary of supporting works for research proposition and theoretical model are presented on Table 2.1 and Figure 2.7, respectively
Summary of Supporting Works for Research Proposition
The degree to which a person believes that using a particular system would enhance his or her job performance
The degree to which a person believes that using a particular system would be free of effort
An individual's attitude is shaped by their specific beliefs about an object, which encompasses evaluation, judgment, and perception based on personal values.
The extent to which an individual likes the object of thought and measures the degree of emotional attraction toward the object
The level of convenience toward time, place and execution that one perceives when using the wireless network to complete a task
The extent to which "mobile services" are perceived as being able to provide pervasive and timely connections
Behavioral intention toward system usage Venkatesh et al
The competitive model plays a crucial role in the development of scientific theory, alongside the theoretical model (Nguyen Dinh Tho & Nguyen Thi Mai Trang, 2011b) Researchers are encouraged to evaluate both models rather than focusing solely on the theoretical aspect (Zaltman, Lemasters, & Heffering, 1982, as cited in Nguyen Dinh Tho & Nguyen Thi Mai Trang, 2011b) Bagozzi (1984, as cited in Nguyen Dinh Tho & Nguyen Thi Mai Trang, 2011b) emphasizes the necessity of comparing competitive and theoretical models within the same study Additionally, Bollen and Long (1993, as cited in Nguyen Dinh Tho & Nguyen Thi Mai Trang, 2011b) highlight the importance of assessing the competitive model, rather than relying on a single theoretical framework.
This study introduces a competitive model to complement existing theoretical frameworks, emphasizing the significant role of perceived convenience in influencing behavioral intentions to use services Research by Liao, Tsu, and Huang (2007) indicates that convenience value enhances shopping motivation, which is crucial for consumers' online shopping intentions Additionally, studies on Radio Frequency Identification (RFID) by Hossain and Prybutok (2008) and online shopping by Gupta and Kim (2007) demonstrate that perceived convenience serves as a key antecedent affecting the intention to adopt mobile technologies Furthermore, Gupta and Kim (2006) confirm that convenience positively influences online shopping intentions Thus, we propose a competitive hypothesis based on these findings.
Hc: Perceived convenience positively effect on behavioral intention to use mobile content services
Based on the competitive hypothesis, the competitive model is formed (Figure 2.8) This competitive model is used to compare with the theoretical model to select the final research model
RESEARCH METHODOLOGY
Research Process
This research consisted of two phases: a pilot study and a main study The pilot study utilized both qualitative and quantitative methods to explore key issues prior to the main study, which focused solely on quantitative methods The primary objective was to examine the factors influencing behavioral intentions toward mobile content services, with particular emphasis on the cognitive and affective components of user attitudes The analysis centered on individual mobile users, and the research was conducted in Ho Chi Minh City, Vietnam's leading business and technology hub Most participants had prior experience with mobile devices, providing valuable insights into user behavior.
The qualitative pilot phase utilized three focus group discussions with skilled mobile content service professionals fluent in English to explore issues related to the content, structure, and meaning of measurement scales This preliminary study aimed to refine the final draft before proceeding to the quantitative phase In the quantitative pilot phase, 150 questionnaires were distributed to mobile users in Districts 1 and 7 of Ho Chi Minh City, resulting in 130 valid responses Concurrently, Cronbach's alpha and exploratory factor analysis (EFA) were employed to evaluate the reliability and validity of the final draft measurement scales.
The primary research utilized a quantitative approach, selecting a sample size of 500 participants through convenience sampling All respondents were required to possess sufficient knowledge about the subject matter.
Before participating in the main survey, respondents were informed about "mobile content services." The primary objective of the study was to evaluate the measurement and structural models To ensure reliability, the measurement scales underwent thorough testing through composite reliability and confirmatory factor analysis (CFA) prior to the hypothesis testing of the theoretical model using structural equation modeling (SEM) The research process for the study is illustrated in Figure 3.1.
Construct Measurement
This study developed measurement scales for its model constructs by adapting items from previous research, as detailed in Table A.1 (Appendix A) Each measurement item utilized a 7-point Likert-type scale, ranging from 1 (completely disagree) to 7 (completely agree) The preliminary scales are presented in Table A.2 (Appendix A).
• Cognitive and affective attitudes were measured by three items for each construct, which were borrowed from the scales of Yang and You (2004):
- Mobile content services is a wise instrument for work/entertainment over mobile network (CA1)
- Mobile content services is a beneficial instrument for work/entertainment over mobile network (CA2)
- Mobile content services is a valuable instrument for work/entertainment over mobile network (CA3)
- Using mobile content services makes me feel happy (AA1)
- Using mobile content services makes me feel positive (AA2)
- Using mobile content services makes me feel good (AA3)
• Perceived usefulness and perceived ease of use were adapted from Davis (1989):
- I would find mobile content services to be useful in my daily life (PU1)
- Using mobile content services would be enable me to accomplish tasks more quickly (PU2)
- Using mobile content services in my tasks would increase my productivity (PU3)
- Using mobile content services would enhance my effectiveness on my tasks (PU4)
- Using mobile content services would make it easier to do my tasks (PU5)
- Using mobile content services in my task would improve my task performance (PU6)
Perceived ease of use (PEU)
- My interaction with mobile content services would be clear and understandable (PEU1)
- It would be easy for me to become skillful at using mobile content services (PEU2)
- I would find mobile content services easy to use (PEU3)
- Learning how to use mobile content services would be easy for me (PEU4)
- It would find it easy to get mobile content services to do what I want it to do (PEU5)
- It would find mobile content services to be flexible to interact with my tasks (PEU6)
• Perceived convenience (PC) was measured by four items borrowed from the scale of Yoo and Kim (2007):
- I can use mobile content services at anytime via the mobile network by my mobile devices (PC1)
- I can use mobile content services at any place via the mobile network by my mobile devices (PC2)
- Mobile content services provide a lot of convenience methods for me to engage in work/entertainment (PC3)
- In general, I feel that mobile content services is convenient for me to do work/entertainment (PC4)
• Perceived mobility (PM) was measured by five items from scale of Hong at al (2008):
- Mobility is an outstanding advantage of mobile phone with mobile content services (PM1)
- I would find mobile content services to be easily accessible and portable (PM2)
- I expect that mobile content service would be available for use whenever I need it (PM3)
- In general, I expect that I would have control over using mobile content services anytime and anywhere (PM4)
- I am able to use mobile content services anytime and anywhere (PM5)
• The behavioral intention to use mobile content services was measured by three items according to Venkatesh, Thong and Xu (2012):
- I intend to use mobile content services in the future (BI1)
- I will always try to use mobile content services in my daily life (BI2)
- I plan to keep using mobile content services as regularly as I do now (BI3).
Measurement Refinement
Based on the literature review, the author built the draft measurement scales for the study (See Table A.2, Appendix A) After that, the researcher started up the
A qualitative pilot study was conducted involving three focused groups, including two mini groups (Group A and Group B) comprised of skilled professionals in mobile content services with expertise in English Within this framework, five professional mobile application developers were selected for participation.
The study involved three groups: "Group A," "Group B," which consisted of five experienced mobile users, and "Group C," comprising eight participants who engaged in a full group discussion in Vietnamese Following a "mini group discussion" and "in-depth interviews" (Nguyen Dinh Tho & Nguyen Thi Mai Trang, 2011a), significant feedback and suggestions were gathered from the participants This led to a deeper investigation into the contents, structure, and meaning of the Vietnamese questionnaire As a result of the discussions, five items were removed from the preliminary scales.
Removed “Using mobile content services would make it easier to do my tasks (PU5)”, due to all of the participants selected “PU5” belong to
The statement "Using mobile content services in my tasks would improve my task performance" was removed to avoid confusion with related concepts such as "accomplishing tasks more quickly," "increasing my productivity," and "enhancing my effectiveness" within the context of this study.
The survey item "I would find it easy to get mobile content services to do what I want to do (PEU5)" was removed due to participant confusion Additionally, the statement "I would find mobile content services to be flexible to interact with my tasks (PEU6)" was excluded, as the concept of flexibility will be assessed within the perceived convenience (PC) scale.
Removed “I am able to use mobile content services anytime and anywhere (PM5)”, because “PC1” and “PC2” already measured the “time and place” characteristics
Lastly, the final draft questionnaire was formed and continued to use for quantitative pilot step (Table A3, Appendix A)
Cronbach's alpha is a key metric for assessing internal consistency, indicating how closely related a group of items is (UCLA, 2012) This study utilized Cronbach's alpha to evaluate reliability, which Nunnally (1967) defines as the degree to which measurements can be consistently repeated, highlighting that random influences causing variations in measurements contribute to measurement error.
In reliability analysis, a high coefficient indicates strong consistency, with Cronbach's Alpha values below 35 signifying low reliability and those above 70 indicating high reliability (Cuieford, 1965) Acceptable reliability is considered to be at least 50 (Nunnally, 1967), as noted by Connely (2011).
Cronbach’s alpha serves as a preliminary measure for evaluating the reliability of instruments or scales, indicating whether the items are consistent with one another However, it does not assess whether these items accurately measure the intended attribute Consequently, it is essential to also evaluate scales based on their content and construct validity.
Cronbach’s alpha is a key measure of reliability, with an acceptable value typically above 70, as noted by Leech, Barrett, and Morgan (2005) Values between 60 and 69 may be acceptable when the scale includes only a few items Conversely, a Cronbach’s alpha exceeding 90 could indicate redundancy among items or an excessive number of items for accurately measuring the intended concept.
Besides evaluating the value of the Cronbach’s alpha, the “Corrected Item–Total Correlation” needs to be considered Leech et al (2005) suggested that, if the
A high corrected item-total correlation (0.40 or above) indicates that the item is well correlated with other items, making it a valuable component of the summated rating scale Conversely, if the corrected item-total correlation is negative or below 0.30, it is essential to evaluate the item for potential wording issues or conceptual misalignment, which may require modification or removal of the item.
This pilot study adhered to the acceptable threshold for Cronbach's alpha, which should be equal to or above 6, and ensured that the "corrected item-total correlation" exceeded 30, as outlined by Nunnally and Bernstein (1994) The findings revealed that all seven scales achieved a Cronbach's alpha greater than 60, with the highest value at 928 and the lowest at 835 Additionally, each item's "corrected item-total correlation" surpassed 30, ranging from a maximum of 867 to a minimum of 647 Consequently, the reliability of all scales fell within a reasonable range, allowing for the use of all observation variables in the subsequent exploratory factor analysis (EFA) Detailed results of the Cronbach's alpha test are available in Table 3.1.
Scale Mean if Item Deleted
Scale Variance if Item Deleted
Cronbach's Alpha if Item Deleted
Perceived Ease of Use (PEU): Alpha = 883
Exploratory factor analysis (EFA) aims to identify the number of underlying factors influencing observed measures and assess the strength of the relationships between these factors and the measures themselves (DeCoster & Claypool, 2004; Byrne, 2010).
Exploratory Factor Analysis (EFA) is utilized when the relationships between observed and latent variables are uncertain This method allows researchers to explore the connections between observed variables and their underlying factors The primary goal is to identify the minimum number of factors that account for the covariation among these observed variables.
In this quantitative pilot study, exploratory factor analysis (EFA) was performed using IBM SPSS 22.0, which offers two primary methods for factor extraction: principal component analysis (PCA) and principal axis factoring (PAF) with promax rotation PAF with oblique rotation, specifically promax, is recommended as it provides a more accurate reflection of the underlying data structure compared to PCA, as noted by Hendrickson and White (1964) and supported by Gerbing and Anderson (1988).
The exploratory factor analysis (EFA) revealed the extraction of seven factors with an eigenvalue of 1.082, accounting for 70.77% of the variance (Table 3.2) EFA is deemed suitable for data when the Kaiser-Meyer-Olkin (KMO) value is 0.60 or higher, and Bartlett's test of sphericity is statistically significant at p < 0.05 (Pallant, 2005) In this study, the KMO value was 0.878, and Bartlett's test showed significance at p = 0.000 (refer to Table D1, Appendix D).
Table 3.2 indicates that all factor loadings after rotation surpassed the acceptable threshold of 50, confirming that no items needed to be removed from the scales Consequently, the finalized draft questionnaire was approved for use in the main study.
Main Study
Following the pilot study, the researcher finalized the questionnaire for the main research Subsequently, the main surveys were distributed to consumers in Ho Chi Minh City using a convenience sampling method, which was executed in four distinct phases.
Phase 1: The researcher finalized the main survey for the main study (Appendix B, C)
Phase 2: The researcher determined the sample size for the main study
According to DeCoster and Claypool (2004), Hair, Black, Babin, and Anderson
According to guidelines established in 2010, the minimum sample size (n) required for study analysis should be at least five times the number of observation variables, with a minimum threshold of 100 participants This means that n must satisfy both conditions: n ≥ 100 and n ≥ 5*k, where k represents the number of observation variables, to ensure the reliability of the results.
This research employed the Structural Equation Modeling (SEM) technique, as recommended by Raykov and Widaman (1995) in the work of Nguyen Dinh Tho & Nguyen Thi Mai Trang (2011b), highlighting the necessity of a large sample size for effective analysis However, the exact number defining a "large" sample size remains ambiguous, as noted by Hair et al.
According to 1998 research, a minimum sample size of 100-150 elements is recommended for Maximum Likelihood (ML) estimation, while Bollen (1989) proposed that a sample size of five elements is sufficient for each estimated parameter Therefore, based on these previous studies, the sample size for the main research was determined accordingly.
Phase 3: The researcher delivered the survey to public
Researcher issued questionnaires to respondents who live in Ho Chi Minh City by:
- Delivering paper forms (hard copy) directly to respondents
Researchers can easily distribute Google survey forms to the public through various channels, including social networks like Facebook and Twitter, as well as via email, instant messaging, and over-the-top (OTT) applications.
The survey form began with a clear definition of "mobile content services" to ensure all respondents understood the concept Data collection took place over a period of 45 days, during which the researcher gathered valuable insights.
A total of 160 paper questionnaires and 420 online surveys were distributed across various districts in Ho Chi Minh City, resulting in a response rate of 64% for the delivered forms For detailed statistics, refer to Table 3.3.
Summary of the Data Collection Process
Phase 4: Researcher received the survey results and double-checked for valid data Removing problem answers such as “missing cases”, “neutral answers (in this study is number 4)” for greater than 50% of the quantity of questions in the form After this step, the usable forms were 505, which achieved 87% of the total returned forms.
Data Analysis
This research utilized three software tools for data analysis: IBM SPSS version 22, Microsoft Excel 2010, and AMOS 22 The analysis phase began with descriptive statistics of measurement scales, followed by the application of Confirmatory Factor Analysis (CFA) and Structural Equation Modeling (SEM) to evaluate both measurement and structural models.
Confirmatory Factor Analysis (CFA) follows exploratory factor analysis (EFA) to validate the identified factor structure in research While EFA investigates potential factor relationships, CFA serves to confirm these structures In this study, CFA was employed to assess the measurement model, utilizing the maximum likelihood (ML) estimation method for analysis.
After validating the constructs using Cronbach’s alpha, exploratory factor analysis (EFA), and confirmatory factor analysis (CFA), the conceptual model was evaluated through structural equation modeling (SEM) with maximum likelihood (ML) estimation to test the hypotheses The competitive model was also assessed and compared to the theoretical model to determine the superior option Additionally, the bootstrap method was utilized to evaluate the stability of parameter estimates, especially when the assumptions of large sample size and multivariate normality are not met, thereby ensuring the reliability of the selected model's parameters (Byrne, 2010).
The bootstrap technique allows researchers to generate multiple subsamples from an original dataset, enabling the examination of parameter distributions for each subsample These distributions collectively form a bootstrap sampling distribution, functioning similarly to traditional sampling distributions used in parametric inferential statistics Unlike conventional methods, the bootstrap sampling distribution is tangible, facilitating the comparison of parametric values across repeated samples drawn with replacement from the original dataset.
Therefore, bootstrap procedure was applied to estimate the parameters of final selected model.
DATA ANALYSIS AND RESULTS
Sample Specification
A frequency analysis of 505 respondents revealed a gender distribution of 52% males (n = 264) and 48% females (n = 241) Age-wise, 35% were aged 20-29 years (n = 176), 30% were 30-40 years old (n = 151), 15% were over 40 years (n = 77), and 20% were under 20 years (n = 101) In terms of monthly income, 12% earned less than VND$5 million (n = 59), 45% earned between VND$5-10 million (n = 227), 15% earned VND$10-15 million (n = 75), and 29% earned over VND$15 million (n = 144).
The analysis focused exclusively on respondents utilizing mobile content services for personal tasks, with expenses covered by themselves or their families It also encompassed individuals without economic independence, such as students and elderly users, who represent a substantial segment of the mobile content services consumer base, with students comprising 20% of the sample (refer to Table 4.1 for detailed statistics).
Confirmatory Factor Analysis
Confirmatory factor analysis (CFA) was conducted using IBM AMOS 22 software, focusing on assessing the overall model fit indices Key indices from the structural equation modeling (SEM) literature, as suggested by Kline (2010), include the Chi-square statistic, degrees of freedom, and p-value; the Comparative Fit Index (CFI) which indicates the proportion of explained variance; the Tucker & Lewis Index (TLI) that adjusts for model complexity; and the Root Mean Square Error of Approximation (RMSEA) as highlighted by Browne.
& Cudeck, 1993) Based on previous researches from Nguyen Dinh Tho and Nguyen Thi Mai Trang (2011b), the model fit indices suggested: (1) /df < 2, (2)
The Tucker & Lewis Index (TLI) value should ideally range from 0.90 to 1, indicating a well-fitting model, while Comparative Fit Index (CFI) values above 0.90 are also associated with good model fit Additionally, the Root Mean Square Error of Approximation (RMSEA) should be less than 0.08, and if the significance test yields a p-value greater than 0.05, it suggests that the model adequately fits the data Key assessment indices include composite reliability, average variances extracted, unidimensionality, convergent validity, discriminant validity, and nomological validity.
In Chapter 3, the "Maximum Likelihood Estimates" method was utilized to estimate the model's parameters, as the assessment of normality for all observation variables met the criteria for kurtosis.
“skewnesses” – all of “skew value” and “kurtosis value” in this test dropped between [-1,+1] (Muthen & Kaplan, 1985), (Nguyen Dinh Tho & Nguyen Thi Mai Trang, 2011b) (See Table E1, Appendix E)
This research focuses on unidimensionality scales, prompting the author to assess measurement concepts within a saturated model A saturated model is characterized by the estimation of all parameters that connect the constructs to one another, as defined by Anderson and Gerbing (1988).
The saturated model in this study demonstrated 254 degrees of freedom, with a chi-square value of 328.726 The model's normalized chi-square value was 1.294, accompanied by a p-value of 001 Additionally, it achieved a TLI of 986, a CFI of 988, and an RMSEA of 024, indicating a strong fit with the market data, as illustrated in Figure 4.1.
Figure 4.1 Saturated model of the theoretical model
Discriminant validity analysis assesses how well two constructs are differentiated (Hair et al., 2010), while convergent validity evaluates the internal consistency of a single construct (Campbell & Fiske, 1959; Churchill, 1979).
Table 4.2 illustrates the results of the discriminant validity of constructs, showing that the correlations between constructs, along with their standard errors, were significantly different from unity Consequently, all constructs successfully met the discriminant validity criteria established by Steenkamp and van Trijp (1991).
PC ↔ BI 340 042 8.112 000 660 15.738 000 PEU ↔ CA 349 042 8.341 000 651 15.587 000 PEU ↔ AA 315 042 7.436 000 685 16.192 000 PEU ↔ BI 359 042 8.631 000 641 15.400 000
Note r: correlation coefficient; se: standard error; cr: critical ratio; p(r): p-value of r; p(1-r): p-value of (1-r)
The factor loadings for all items ranged from 70 to 85, with mean estimate weights between 75 and 81, and all estimates were significant at p-value < 001 (see Table E2, Appendix E) Additionally, the study met reliability and average variance extracted requirements, as Cronbach’s alpha and composite reliability were between 80 and 89, with the lowest average variance extracted at 56% These results confirm that all concepts demonstrated unidimensionality, convergent validity, and discriminant validity (Nguyen Dinh Tho & Nguyen Thi Mai Trang, 2011b) (see Table 4.3).
Concept Component Number of Items
Note ρ : average variance extracted; ρ : composite reliability; ̅ : mean of estimate weights
SEM Approach for Theoretical Model
The theoretical model was evaluated using Structural Equation Modeling (SEM) with AMOS 22.0 software, yielding an acceptable fit to the data set, indicated by the following statistics: /df = 1.418, p = 000, TLI = 980, CFI = 983, and RMSEA = 029 As shown in the SEM results, the only hypothesis that was not supported was the direct relationship between perceived usefulness and affective attitude, which had a p-value of 674, while all other hypotheses were confirmed.
The analysis revealed no residual variances below zero, indicating that the Heywood case (Heywood, 1931) did not occur during the maximum likelihood estimation However, there were eight standardized residuals with absolute values exceeding 2.58, as detailed in Table F10 of Appendix F This finding will be further explored in the subsequent section titled "Optimizing the Theoretical Model."
Figure 4.2 Standardized SEM results for theoretical model
Optimized the Theoretical Model
Although the theoretical model showed a good fit, several indices require improvement to identify a more accurate model AMOS provides two statistical tools—standardized residuals and modification indices—that are useful for detecting model misspecification.
Previous research by Nguyen Dinh Tho and Nguyen Thi Mai Trang (2011b) indicates that standardized residuals with absolute values exceeding 2.58 are deemed significant Additionally, Jureskog and Sorbom (1984) assert that the residual covariance between two variables in a correctly specified model should remain below an absolute value of two.
The second type of statistics concerning misspecification assesses how well the hypothesized model is represented Evidence of model misfit is indicated by modification indices (MIs), which can be understood as a statistic with one degree of freedom, as noted by Jűreskog and Sűrbom.
Hence, the author suggested optimizing the conceptual model based on
The study identified eight standardized residuals exceeding |2.58|, as shown in Table F10 (Appendix F), alongside several residual covariance pairs with high Modification Indices (MIs), notably the pair "ε05- ε23" (Table F1, Appendix F).
The process of optimizing was conducted in systematically as flowing:
• In the residual matrices, PU1 (column) consisted of five high standardized residual values (3.78, 3.99, 2.13, 2.16, 2.09) so it was noted
• PU1’s residual “ε05” also caused the pair “ε05- ε23” captured dramatic MI value (20.640, See Table F1, Appendix F)
• Decided to delete PU1 out of model
• Processed next PM1, PC4 variables
After completed optimizing, the theoretical model was represented schematically in Figure 4.3
Figure 4.3 The optimized theoretical model
Overall, the optimized model got better fit indices: /df = 1.179, CFI increased from 983 to 993, TLI increased from 980 to 992, RMSEA improved from 029 to .019 and p-value increased from 000 to 043
The model successfully met all criteria, including the absence of the Heywood case, standardized residuals below |2.58|, and low modification indices (MIs) as detailed in Table F11 in Appendix F Additionally, Table 4.5 presents the hypothesis testing results for the optimized model.
Competitive Model Test
Based on the important role of competitive model when testing a conceptual model (revised in chapter 2), this study suggested testing one competitive model with following hypothesis:
Hc: Perceived convenience positively effect on behavioral intention toward usage
Figure 4.4 The Standardized SEM results of Competitive Model
The estimations of competitive model yielded results in a good fit to data: /df 1.163, p = 059, TLI = 993, CF I= 994, RMSEA = 018 It indicated that competitive model was goodness-of-fit statistics (See Figure 4.4)
Next, using the “ – test” to determine whether significant difference between two models (optimized theoretical model and competitive model) The ” –Test” was presented in Table 4.6 below
The analysis reveals a p-value of 0.037, which is less than the significance level of 0.05, indicating a significant difference between the optimized theoretical model and the competitive model Consequently, this study aims to identify the most suitable model for research purposes (refer to Table 4.7 for further details).
Model df /df CFI TLI RMSEA P
The summary table revealed that the competitive model, which utilized one degree of freedom, significantly improved the model fit indices, particularly noted by the change in p-value from significant (.043) to non-significant (.059), indicating a better fit for the data Furthermore, the competitive hypothesis (Hc: Perceived convenience positively affects behavioral intention toward usage) was found to be significant, with all prior hypotheses from the theoretical model being supported and continuing to hold in the competitive model (Table 4.8).
Therefore, the competitive model was selected as the final research model
Competitive Model (Final Model) Relations of Constructs (Standardized)
Applying Bootstrap Procedure
The chosen model was utilized for estimation through a bootstrap procedure, employing AMOS to analyze N = 1000 samples The outcomes of the bootstrap analysis are clearly presented in Table 4.9.
Relations ML Estimate Bootstrap Estimate
Estimate SE SE-SE Mean Bias SE-Bias CR
Note: SE: standard error; Bias: mean – estimate; ML: estimate value; CR: critical value
The estimation results indicated that the bias and standard error of the bias estimate (SE-Bias) between Bootstrap and ML estimations were not significant, with |CR| values less than 2 Consequently, it can be concluded that all estimation results in the research model are reliable The final model for the research is illustrated in Figure 4.5.
Figure 4.5 The final research model
Hypotheses Testing
According to the SEM results as presented in Figure 4.5 (Final research model), the summary of hypotheses for study was presented in Table 4.10
The study reveals significant relationships among various constructs, with Perceived Ease of Use positively influencing Perceived Usefulness (β = 486, p < 001) Furthermore, Perceived Usefulness significantly affects Cognitive Attitude (β = 274, p < 001), while its influence on Affective Attitude is not supported (β = 032, p = 618) Perceived Ease of Use also impacts both Cognitive Attitude (β = 249, p < 001) and Affective Attitude (β = 175, p = 008) Additionally, Cognitive Attitude demonstrates a strong effect on Behavioral Intention (β = 301, p < 001), as does Affective Attitude (β = 244, p < 001) There is a notable connection between Cognitive Attitude and Affective Attitude (β = 416, p < 001) Perceived Ease of Use significantly enhances Perceived Convenience (β = 693, p < 001), which in turn positively influences Perceived Usefulness (β = 171, p = 016) Lastly, Perceived Mobility also contributes to Behavioral Intention (β = 327, p < 001), while Perceived Convenience has a positive but weaker effect on Behavioral Intention (β = 102, p = 032).
Note β: standardized coefficient regression weight; ***p
