Journal of Retailing and Consumer Services 29 (2016) 104–113 Contents lists available at ScienceDirect Journal of Retailing and Consumer Services journal homepage: www.elsevier.com/locate/jretconser Internet service providers' service quality and its effect on customer loyalty of different usage patterns Thu Nguyen Quach a, Paramaporn Thaichon b,n, Charles Jebarajakirthy a a b Swinburne University of Technology, Victoria, Australia S P Jain School of Global Management, 5 Figtree Drive, Sydney Olympic Park, Sydney, New South Wales 2127, Australia art ic l e i nf o a b s t r a c t Article history: Received 22 May 2015 Received in revised form 18 November 2015 Accepted 21 November 2015 Available online 5 December 2015 This study attempts to investigate the dimensions of an ISP's service quality, and their effects on customer loyalty in high-tech services. Data was obtained from 1231 internet users. The analyses include segmenting ISPs' customers on the basis of their usage pattern and evaluating their perceptions of Internet service quality dimensions. Through the use of structural equation modelling and bias correct bootstrapping techniques, the study confirms that service quality dimensions can influence both attitudinal and behavioural loyalty. These effects, however, are different across different groups of ISP customers. The contribution of the present paper stems from the modelling of mediation effects and the incorporation of Internet usage that can help better explain the impact of service quality dimensions on customers' loyalty in high-tech service settings. & 2015 Elsevier Ltd. All rights reserved. Keywords: Internet usage pattern Attitudinal loyalty Behavioural intentions Services quality Internet service provider (ISP) High-tech services 1. Introduction Service quality is an important differentiator in a competitive business environment, and a driver of service-based businesses (Zhao and Benedetto, 2013). By enhancing service quality, businesses can influence customers' retention and loyalty (Thaichon et al., 2012). However, very few studies have assessed how different aspects of Internet service providers' (ISP) service quality would influence their customers' loyalty (Vlachos and Vrechopoulos, 2008). ISPs may benefit from obtaining accurate information regarding their customers' assessments of their brand's delivered service quality; such information may enable service brand managers to formulate appropriate marketing strategies in order to achieve competitive advantage and long term profitability. This paper attempts to fill this important research gap by investigating the effects of ISPs' service quality on their customers' loyalty in the high-tech Internet services. In addition, this paper segmenting ISPs' customers on the basis of their usage pattern and evaluating their perceptions of Internet service quality dimensions. With the rise of technology-enabled services, the attention of the services literature has shifted to measurement and operationalisation issues in service quality (Carlson and O’Cass, 2011; Ganguli and Roy, 2010). The earliest service quality model was n Corresponding author. E-mail addresses: [email protected] (T.N. Quach), [email protected] (P. Thaichon), [email protected] (C. Jebarajakirthy). http://dx.doi.org/10.1016/j.jretconser.2015.11.012 0969-6989/& 2015 Elsevier Ltd. All rights reserved. introduced by Parasuraman et al. (1985), and was referred as SERVQUAL, including (1) tangibles; (2) reliability; (3) responsiveness; (4) assurance; and (5) empathy. In addition to SERVQUAL, E-S-QUAL has been developed by Parasuraman et al. (2005) as an attempt to fully capture service quality in the new information age. However, telecommunications service quality cannot effectively be measured by SERVQUAL or E-S-QUAL (He and Li, 2010) as these scales lack the ability of addressing specific issues relevant to this particular context, especially high-tech ISPs. In particular, while SERVQUAL applies to general service, E-S-QUAL focuses on service providers who operate via the internet platform (Vlachos and Vrechopoulos, 2008) and not those who provide the internet connection and platform for online business-to-business and business-to-customer activities. On the other hand, segmentation can help better leverage a service provider's resources and capabilities to fully take advantage of existing opportunities (Weinstein, 1987). As the needs of consumers are not homogenous, it is essential to divide the market into various segments (Mazzoni et al., 2007). Although the concept of market segmentation has been discussed extensively in the literature (Mazzoni et al., 2007; Wedel and Kamakura, 2003), there is scarce empirical evidence of how ISPs can effectively segment their target audience. In this study customers are segmented based on their usage pattern, which is one of the most logical basis of segmentation in similar types of services (Mazzoni et al., 2007; Wedel and Kamakura, 2003). More specifically perceptions of service quality dimensions and their relationships with loyalty of light, medium and heavy users will be evaluated. Based T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 on the foregoing discussion, the objectives of this research study are: firstly, to establish the relationships between specific service quality dimensions of residential internet services and customers' behavioural and attitudinal loyalty. Secondly, to investigate the differences between light, medium and heavy users of the hightech residential internet services. Lastly, to provide managerial implications to high-tech residential internet service providers. 2. Literature review and development of hypotheses 2.1. ISPs' service quality dimensions Previous research indicates that judgment of overall service quality in the telecommunications industry comes from customers' perceptions of a stable and strong network quality (Lai et al., 2009), ready-to-serve customer support team (Aydin and Özer, 2005), informative website support (Thaichon et al., 2012) and a high level of security and privacy that is trusted by customers (Roca et al., 2009). Network quality is one of the core service drivers in the telecommunications industry (Lai et al., 2009). Network quality in the internet services industry involves the quality and strength of the network signals (Wang et al., 2004), number of errors, downloading and uploading speed (Vlachos and Vrechopoulos, 2008). Any break in the internet connectivity may lead to low perceptions of service quality. Moreover, when customers face problems in high-tech internet services, they often seek help and support from technical and customer service staff. For this reason, customer service teams are under constant pressure to perform their work reliably, dependably, and according to set protocols in order to meet their productivity goals and deliver quality customer service (Rod and Ashill, 2013). A study in the Turkish telecommunications industry demonstrates that handling customers' complaint efficiently contributes to overall service quality (Aydin and Özer, 2005). Information technology tools are utilised to increase efficiency and effectiveness of information delivery (Ganguli and Roy, 2010). Relevant, timely, and reliable information helps customers to obtain information and enable effective decision making (Hsieh, 2013). Moreover, information quality plays an important role in building customers' overall positive attitude towards the company (Jaiswal et al., 2010). In fact, a service provider facilitating high levels of information quality and website support is often perceived to have better service quality. There have been considerable concerns regarding safety and ethical behaviour in e-commerce (Limbu et al., 2011). Customers are prone to attribute low risks in purchasing from service providers who are reputable in relation to their security practices (Roca et al., 2009). Security refers to the extent a customer perceives the entire transaction as being safe, which includes payment methods and transmitting confidential information (Chang and Chen, 2009; Thaichon et al., 2014). Privacy is often a concern of customers of high-tech services, and this dimension relates to the customer perception of the quality of processes used for personal information transmission and storage (Özgüven, 2011). Several studies report that security and privacy are related to service quality (Wolfinbarge and Gilly, 2003). 2.2. Behavioural and attitudinal loyalty The concept of customer loyalty has received considerable attention in the marketing literature. There are many approaches to measuring customer loyalty and several studies have attempted to define the “true nature” of loyalty. Basically, several researchers explain loyalty purely from the behavioural point of view (Jaiswal and Niraj, 2011) whilst some argue that an attitudinal perspective is more reflective of customer loyalty (Flint et al., 2011; Jacoby and 105 Chestnut, 1978). This research embraces an integrated theory, which suggests that customer loyalty is a combination of both behavioural and attitudinal loyalty (Dick and Basu, 1994; Oliver, 1999). In this respect, Flint et al. (2011) consider customers' loyalty as a concept with multiple aspects including repurchase intention and corresponding preferences and attitudes towards the brand. While behavioural loyalty is defined as repeat purchase (Zeithaml et al., 1996), this study considers attitudinal loyalty as customers' inner thoughts of attachment, word-of-mouth, and recommendations (Zeithaml et al., 1996). 2.3. Relationships between service quality dimensions, behavioural and attitudinal loyalty It is commonly acknowledged that service quality drives customer loyalty and company profitability (Prentice, 2014). This study has endeavoured to study the effects of each service quality dimension on behavioural and attitudinal loyalty. Network quality including connectivity quality, clarity of signals, and speed of internet is deemed to be the fundamental quality characteristics in high-tech services which affect customer retention (Seo et al., 2008). Other scholars also confirm that network quality is one of the most important drivers of customer loyalty when dealing with prepaid cell phone (Miranda-Gumucio et al., 2012). In the context of the US mobile phone services, Cassab (2009) demonstrated that network quality has the largest coefficient values in the regression analysis of experimental data and therefore, has stronger influence on customers' loyalty intention. Similar results are reported in the US wireless services by Seo et al. (2008), who state that the connectivity quality of wireless is positively and significant related to customers' repurchase intention. Based on the foregoing discussion, we propose the following hypotheses: H1a : Network quality is positively associated with customers' attitudinal loyalty. H1b : Network quality is positively associated with customers' behavioural loyalty. It has been suggested that customer service has a positive and significant impact on customer loyalty in technology-based banking services (Ganguli and Roy, 2011). Responsiveness of technical and customer service staff has significant positive influence on behavioural loyalty in the Greek mobile telephony service (Santouridis and Trivellas, 2010). Customer care is the major determinant of repeat purchase intention and customer loyalty, while deficiencies in the quality of customer service are the main reasons for customers switching (Miranda-Gumucio et al., 2012). Hence, the following has been hypothesised: H2a : Customer service and technical support influence customers' attitudinal loyalty. H2b : Customer service and technical support influence customers' behavioural loyalty. A service provider facilitating high levels of information quality and website support is able to maintain a long term relationship with customers (Canhoto and Clark, 2013). Likewise, researchers report that website design is found to be a determinant of loyalty of online customers in South Africa and Australia (Caruana and Ewing, 2010). Other website characteristics such as ease of use and information are significant influencers of customer loyalty in e-commerce (Toufaily et al., 2013) and content websites (Jaiswal et al., 2010). Based on the above discussion, the following are hypothesised: 106 T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 H3a : Information quality influence customers' attitudinal loyalty. H3b : Information loyalty. quality influence customers' behavioural Previous research reports that security and privacy are related to customer loyalty, especially in the e-commerce contexts (Ratnasingham, 1998). Recently, Limbu et al. (2011) reported a positive link between customer privacy and website loyalty in the USA. In online financial services, privacy is shown to have a direct effect on intention to recommend (Finn et al., 2009). Researchers believe that customer privacy plays a significant part in determining loyalty (Jaiswal et al., 2010). Jin and Kim (2010) state that the contribution of security and privacy towards customer loyalty is greater in online multichannel retailers than in offline retail settings in Korea. Hence, the following have been hypothesised: H4a : Security and privacy influence customers' attitudinal loyalty. H4b : Security and privacy influence customers' behavioural loyalty. There exists a strong relationship between attitudinal loyalty and customer repurchase intentions as suggested by Bandyopadhyay and Martell (2007). Han and Hyun (2012) demonstrate that conative loyalty has a positive effect on action loyalty in hotel services in the United States. In the Korean online marketplaces, attitudinal loyalty is positively related to purchase intentions (Hong and Cho, 2011). A research in the Chinese mobile phone service reported that attitudinal loyalty has a significant positive effect on behavioural loyalty (Zhang et al., 2010). Therefore, we propose that an ISP's service quality dimensions have both a direct and indirect effect on its customers' behavioural loyalty. The indirect effect is channelled via its customers' attitudinal loyalty. This is articulated in the proposed conceptual model depicted in Fig. 1, and the following is hypothesised: Transactions Development Agency (ETDA, 2013) discloses that in general Internet users in Thailand who are online for 11 h per week make up 35.7 per cent; those who use Internet from 11 to 20 h weekly constitute 25.8 per cent; 10.7 per cent spend from 21 to 41 h on the Internet per week; and 27.8 per cent spend more than 41 h every week. Based on the usage segmentation of previous research, this study determines three main groups of Internet users, namely light (i.e. less than 9 hours per week), medium (i.e. 9–29 h per week), and heavy users (i.e. more than 29 h per week) (Thaichon et al., 2014). Heavy internet users, who are the likes of online game players or frequent internet surfers, would possibly perceive network quality as the pre-dominant dimension which influences their perception of an ISP, and might consider network quality as the priority to choose and remain with a service provider. Light users would possibly perceive customer services and technical support as their priority for recommending an internet service or in actual behaviour of staying with an ISP, since they are usually unfamiliar with technical issues. Therefore, it is most likely that each specific service quality dimension distinctively impacts customer loyalty (Prentice, 2014), which refers to both attitudinal and behavioural perspectives, depending on differently segmented groups of customers. Hence, the following have been hypothesised: H6 : ISPs' service quality dimensions differently influence attitudinal loyalty across the three segments of light, medium and heavy users. H7 : ISPs' service quality dimensions differently influence behavioural loyalty across the three segments of light, medium and heavy users. 3. Method H5 : Attitudinal loyalty is a mediator in the relationship between specific ISPs' service quality dimensions and behavioural loyalty. 3.1. The study sample Apart from identifying the four ISP service quality dimensions in high-tech residential internet services, it is of interest that, different customers have distinctive needs and require tailored approaches (Mazzoni et al., 2007). In this process, customers are generally segmented as heavy, medium and light users (Thaichon et al., 2014). Heavy users are those who spend more than 29 h on the internet every week (Assael, 2005). The average time spent on Internet of a normal person is from 9 hours to 20 h per week (ACMA, 2012). Heavy users often spend more than 29 h on the Internet weekly, while light users only use the Internet for less than 9 h every week (Assael, 2005). A study by Electronic To test the hypotheses, an online survey was designed and conducted in all regions of Thailand. Thailand is endowed with a wide variety of natural resources, a substantial population and a relatively strong economy. The enhanced investment on education has resulted in knowledge improvement and higher educational qualification of the Thai people. Thailand is fast becoming an information society. This is part of the reason for the considerable development of the telecommunications industry in Thailand. Overall contribution to GDP from the Internet in Thailand is expected to be 3.8 per cent p.a. in 2020 (Telenor, 2015). As the Internet is a capital good that enables increased production across all economic activities, ISP industry plays a very important part in the Thai economy (Telenor, 2015). Internet in Thailand has passed its infancy and is currently in the growth stage, poised to take off (Telenor, 2015). Thailand is ranked third in South East Asia by way of residential internet usage with an estimated 17 million internet users in 2009 (CIA, 2013) and over 27 million internet users in 2014 (NECTEC, 2015), representing a penetration rate of approximately 40% (NECTEC, 2015). The Internet penetration in Thailand has seen a rapid increase since 2014 and the average internet user growth rate is 30% per year (NECTEC, 2015). Therefore, this provides a suitable scenario to study the behaviour of Internet users. However limited research on customers' buyer behaviour of home internet services has been conducted in Thailand. On the other hand, the competition in Thailand among residential internet service providers is intense (Kim, 2015). Currently there are three majors ISPs and 16 smaller ones across the country (Thaichon and Quach, 2013). In this highly competitive market, the churn rate of internet users Fig. 1. Conceptual framework. T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 was approximately 12% in 2009 (Thaichon et al., 2012). This scenario, therefore, poses huge challenges to ISPs especially in the area of customers' repurchase intention. 107 Table 1 Sample structure. Demographic profile Current study (%) 2013 (%) Chi square statistics 3.2. Data collection χ2 (1) ¼2.388, p ¼ 0.877 Gender Data was collected from residential internet users in all regions of Thailand in 2013. Customer databases of well-established ISPs who account for 95% of the Thai home Internet services (True, 2013) in Thailand were utilised as the sampling frame. The selected comprehensive customer databases incorporates diverse customer profiles, including those who have switched from other ISPs or those who wish to change to other service providers. Hence, the sampling frame is representative of the entire population of Thai home Internet service customers. Simple randomisation was chosen to achieve freedom from human bias and to avoid classification errors (Black, 1999). This means that each individual in the population had an equal chance of being included in the sample (Teddlie and Yu, 2007). As such, the participants for the survey were randomly selected from the customer databases by computer software. This sampling technique has several advantages as it is simple and easy to apply, especially when a comprehensive list of ISP customers was obtained (Rossi et al., 2013). The corporations did not have any control in determining who participated in the survey, and did not benefit from influencing or creating bias during the data collection process. The survey responses were stored in the Opinio database. The companies did not have access to this data. They were only interested in an independent and academically rigorous process, the findings of which would assist them in developing and designing long-term customer retention strategies. It was calculated that the representative sample of Thailand's population would be a number exceeding 700 (using a confidence level of 95 per cent and a sampling error of 72.5). Hence we emailed a total of 4000 surveys in two stages, i.e. 2000 surveys were distributed in all geographical regions of Thailand and the other 2000 were similarly emailed to participants a week later. The online survey was made available via the Opinio platform. The web link was relayed for the online survey and emailed to ISP customers who were randomly selected from the databases. Responses to the online survey were automatically returned to the researchers through the Opinio platform. Opinio software enables the production and reporting of a survey and assures the anonymity, confidentiality and privacy of the respondents. In order to achieve accurate results, and in particular, to prevent multiple attempts by the same respondent, the default in Opinio was set as follows: ‘not allow multiple submissions', and ‘prevent with cookies and IP-address check'. This means that the Opinio software recognised every respondent's IP address (computer ID), and only one completed survey was accepted from a particular computer. The Opinio platform was kept live for a period of three months. 3.3. Respondent profiles The final usable sample size was 1231. In terms of the profile of the respondents, as illustrated in the Table 1, the sample structure is very similar to the overall structure of Internet users as specified in previous research in Thailand (ETDA, 2013). For instance, 45.6 per cent of the total respondents were males, while 55.4 per cent were females. This is consistent with the results of previous study of Internet users in Thailand in 2013 with 47.8 per cent males and 52.2 per cent females (ETDA, 2013). Similarly, Generation Y made up more than 50 per cent of the respondents, consistent with the findings of the earlier research (ETDA, 2013). In addition, respondents with income from 10,001 to 30,000 baths were the largest group while Bachelor's degree dominated the categories of Male Female Age 45.6 55.4 o 19 20–29 30–39 40–49 50þ Monthly household income 7.5 29.1 34.3 17.2 11.9 Under 10,000 10,001–30,000 30,001–50,000 50,001–100,000 Over 100,000 Education 7.2 34.9 20.6 23.9 13.4 Secondary or below qualification 2 Years college or associate degree Bachelor's degree Postgraduate degree or higher Number of Internet users in a household 1 2 3 4 5 or more Location 6.1 5.5 8.3 6.6 59.7 25.9 61.1 26.8 Bangkok and surrounding suburbs Others 47.8 52.2 χ2 (4) ¼8.405, p ¼ 0.078 6.6 30.8 32.7 19.3 10.6 χ2 (4) ¼7.344, p ¼0.119 6.4 36.3 21.5 24.5 11.3 χ2 (3) ¼6.963, p ¼ 0.073 χ2 (4) ¼5.327, p ¼ 0.255 13.2 14.7 18.9 25.9 27.3 14.7 15.9 19.2 24.2 26 χ2 (1) ¼2.855, p ¼0.091 48.3 45.9 51.7 54.1 education level. Most of respondents stayed in a household with 3 or more Internet users. These findings were in line with the 2013 study of ISP customers in Thailand (ETDA, 2013). Almost half of the respondents in this study came from Bangkok and its surrounding suburbs. This can be explained by the fact that the penetration rate of Internet in Thailand is only approximately 40% (NECTEC, 2015) and Internet is mainly available in cities and urban areas. In order to further compare the respondents' structure in the sample and the overall population structure, Chi-square statistics were obtained for each of the socio-demographic variables. The results presented in Table 1 indicate that there is no significant difference between the sample and population distributions. In terms of Internet usage, in this study 21.2 per cent of the respondents were light users, 33.1 per cent were medium users and 45.7 per cent were heavy users. In terms of Internet usage groups, a majority of light users were between 29 and 38 years (i.e. 41.2 per cent) and 39 and 49 years (i.e. 35.7 per cent). Interestingly 68.9 per cent of pensioners fell in the category of light users. This could be explained by the fact elder people are usually not very conversant with technology, especially in an Asian context. In terms of gender, females were more likely to be light users than males. In addition, approximately 70 per cent of full time workers were medium users. There were no noticeably uncommon patterns in the age groups and the distribution of gender among medium users. Heavy users were more distinctive when compared to the other two segments. Young Internet users made up the largest percentage of this group; 108 T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 Table 2 Online activitiesn. Online activity Heavy user (%) Medium user Light (%) user (%) Difference test Emails 30.4 44.3 67.4 Search for information Social media 54.2 15.1 40.2 38.1 33.5 20.1 χ2 (2)¼ 5.823, p ¼.054 χ2 (2)¼ 10.621, p ¼.006 χ2 (2)¼ 2.669, p ¼.259 n Respondents could give more than one answer. more than 61.3 per cent were in the 18–28 age bracket and 49.7 per cent were in the 29–38 age bracket. Consequently, more than half of the students in this study were classified as heavy users. In relations to online activities, heavy users spent most time on searching for information (i.e. 54.2 per cent), 38.1 per cent selected social media and 33.9 per cent picked emails. In contrast, 44.3 per cent of medium users chose emails as their primary activity followed by social media (i.e. 33.5 per cent) and information search (i.e. 15.1 per cent). Similarly, emails were the most popular among light users with 67.4 per cent; information search and social media were next with 40.2 per cent and 20.1 per cent respectively. Chi square difference test was conducted and reported in Table 2. The results indicate a significant difference between these groups of customers in terms of information search behaviour. survey are shown in Table 3, which is depicted in the Section 4. Vlachos and Vrechopoulos (2008) connection quality scale was used to measure network quality. The scale examines connection quality for mobile phone services which is very similar to the nature of an ISP's network quality. The customer service scale was taken from Wolfinbarge and Gilly (2003) which addresses both customer service and technical support in the ISP's offerings. Four different scales relating to information and website support from Chae et al. (2002), Lin (2007), Kim and Niehm (2009) and Vlachos and Vrechopoulos (2008) were considered. After a thoughtful analysis, the Kim and Niehm's (2009) information quality scale was selected as this scale has stronger factor loadings (.80–.83), and Cronbach's alpha rate (α ¼.96). Vlachos and Vrehopoulos's (2008) privacy scale was selected. The scale's measurement items investigate whether customers feel safe parting with information during transactions, and also seek their opinions on security features of an ISP. The loyalty scale from Kim and Niehm (2009) was chosen to measure attitudinal loyalty since it has strong factor loadings (.71–.95) and Cronbach's alpha (α ¼ .93). Additionally, this scale covers all the aspects of attitudinal loyalty and is used to determine if customers consider themselves to be loyal patrons of a particular ISP. Zeithaml et al. (1996) behavioural loyalty scale and complaining behaviour scale were selected to measure another aspect of loyalty in the ISP market. They have reasonable factor loadings (.74–.94) for behavioural loyalty scale; and (.76–.99) for complaining behaviour scale. They aim to investigate whether the customer wishes to stay with a particular ISP in the next few years or to switch. 3.4. Measures Respondents were required to rate their perceptions for every item using a likert scale which was anchored at 1 for strongly disagree and 5 for strongly agree. These statements originally in English were translated into Thai language by a professional translator. Subsequently, the translated Thai version was back translated into English. Significant misunderstanding or confusion caused by a cross-cultural transformation was detected through the back-translation process. Discrepancies in the translation were carefully inspected and corrected to ensure that the items reflected the original meaning, and did not contain any social judgments. To confirm the error-free translation, the translated versions were then crosschecked by three other bilingual researchers to ensure face and content validity. The items of the 4. Analysis and results 4.1. Factor analysis and validity testing The multi-scale nature of the data comprising ordinal scales requires the use of polychoric correlation matrices of software programmes (Hair et al., 1998). Hence, AMOS version 20 (Analysis of Moment Structures) was employed to analyse the data. The items used to assess the ISP's service quality were factor analysed to confirm dimensionalities. Confirmatory factor analysis (CFA) was performed to examine whether theoretical relationship between items and their hypothesised factors were supported by the data (Cunningham, 2010). Table 3 depicts 6 constructs in the Table 3 Instrument items and reliability indices. NQ CS IW SP AL BL Items FL α CR AVE I do not experience any Internet disconnection from this ISP The Internet downloading and uploading speed meet my expectations The Internet speed does not reduce regardless peak or off-peak hours Customer service personnel are knowledgeable Customer service personnel are willing to respond to my enquiries My technical problems are solved promptly This ISP provides sufficient information This ISP provides up-to-date information This ISP provides relevant information I feel that my personal information is protected at this ISP I feel that my financial information is protected at this ISP I feel that the transactions with this ISP are secured I consider myself to be a loyal patron of this ISP I would say positive things about this ISP to other people I would recommend this ISP to someone who seeks my advice I would consider this ISP as my first choice to buy services I would do more business with this ISP in the next few years I would do less business with this ISP in the next few years (-) .714 .845 .809 .872 .886 .798 .832 .822 .818 .734 .765 .911 .822 .898 .924 .897 .800 .532 .828 .833 .625 .887 .889 .727 .876 .864 .679 .855 .847 .651 .917 .913 .779 .797 .796 .576 Notes: FL ¼ factor loadings, α ¼Cronbach's alpha, CR¼ Construct reliability, AVE ¼Average variance extracted, NQ¼ network quality; CS ¼ customer service and technical support; IW¼ information and website support; PS ¼privacy and security; AL ¼attitudinal loyalty; BL ¼ behavioural loyalty. T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 Table 4 Correlations between variables. NQ CS IW PS AL BL Table 6 Regression results for the mediation of the effect of attitudinal loyalty on behavioural loyalty. NQ CS IW PS AL BL .625 .607 .679 .583 .615 .619 .368 .727 .746 .603 .569 .568 .461 .557 .679 .761 .700 .677 .340 .364 .579 .651 .657 .647 .378 .324 .490 .432 .779 .748 .383 .323 .458 .419 .560 .576 Notes: The diagonal elements are the AVEs (italicised and bolded). The lower-left triangle elements (italicised) are correlations and the upper-right triangle elements are the squared correlations between constructs. All correlations are significant at the.01 level (2-tailed). NQ¼ network quality; CS ¼customer service and technical support; IW¼ information and website support; PS ¼ privacy and security; AL¼ attitudinal loyalty; BL ¼ behavioural loyalty. conceptual model, associated indicators, and their standardised coefficients, reliability indices and average variance extracted. Factor loadings of all indicators were positive and statistically significant. Reliability estimates for each of the construct, i.e. Cronbach's alpha, and composite reliabilities exceeded the threshold .70 (Nunnally, 1978). Table 4 presents the correlations and squared correlations between constructs. The correlations between constructs were significant, ranging from .583 to .761 which were well below the .90 cut-off (Tabachinick and Fidell, 2001). Therefore, there was no redundancy or violation of multicollinearity. In addition, the average variance extracted (AVE) for each factor was over .50, indicative of adequate convergence (Fornell and Larcker, 1981). The construct AVE estimates were larger than the corresponding squared inter-construct correlation estimates (SIC), thereby supporting discriminant validity. 4.2. Hypotheses testing Structural equation modelling (SEM) was conducted to examine the research model. SEM is suitable as it allows testing of structural models, specifically those containing latent constructs (Anderson and Gerbing, 1988). SEM enables the estimation of multiple and crossed relationships between dependent and independent variables, and is able to denote constructs unobserved in these relationships as well as dealing with measurement error in the estimation process (Beerli et al. (2004). The results are presented in Table 5. Whereas Chi-square statistic known to be highly sensitive to sample size was significant (χ2 (120) ¼563.180, p ¼.000), the other fit indices (CMIN/DF ¼4.693, GFI.970, Table 5 Results for the relationships between service quality dimensions, attitudinal and behavioural loyalty, coefficients. AL AL AL AL BL BL BL BL BL o— o— o— o— o— o— o— o— o— NQ CS IW SP NQ CS IW SP AL Estimate S.E. C.R. .234 .045 .400 .297 .045 .038 -.062 .029 .941 .030 .038 .053 .037 .023 .028 .041 .028 .024 7.885 1.193 7.566 8.048 1.958 1.358 -1.515 1.036 39.805 P nnn .233 nnn nnn .050 .174 .130 .300 nnn Notes: 1. NQ ¼network quality; CS ¼customer service and technical support; IW¼ information and website support; PS ¼ privacy and security; AL ¼attitudinal loyalty; BL ¼behavioural loyalty; CFI ¼ comparative fit index; GFI goodness of fit index; AGFI¼ adjusted goodness of fit index; RMSEA ¼root mean square error of approximation; TLI¼ The Tucker-Lewis coefficient; 2. Model fit indices: χ2 (120) ¼ 563.180, p ¼ .000, χ2/df ¼4.693, GFI¼ .970, AGFI ¼.957, TLI¼ .978, CFI¼ .983, RMSEA ¼.042; 3. R2 (BL) ¼ .913; R2 (AL) ¼.53. nnn 109 p Values are statistical significant at .001 levels; Structural path Direct model NQ-BL CS-BL IW-BL PS-BL R2 BL1 ¼ .524 Mediation model NQ-AL CS-AL IW-AL PS-AL AL-BL NQ-BL CS-BL IW-BL PS-BL R2BL2 ¼ .913 Standardised coefficients Standardised indirect effects .242nnn .063 .268nnn .267nnn – – – – .222nnn .036 .326nnn .258nnn .938nnn .043 .030 .050 .025 – – – – – .208nnn .033 .306nnn .242nnn Notes: NQ¼ Network quality; CS ¼Customer service and technical support; IW¼ Information and website support; PS ¼ Privacy and security; AL¼ Attitudinal loyalty; BL ¼ Behavioural loyalty R2BL1 is the proportion of variance in BL explained by the direct model R2 BL2 is the proportion of variance in BL explained by the mediation model nnn pr .001. AGFI¼.957, TLI¼ .978, CFI¼ .983, RMSEA ¼.042) indicate that the model was a good fit to the data. The ISP's service quality explained 53.0% of variance in attitudinal loyalty (R2 ¼53.0%), and the whole model explained 91.3% of variance in behavioural loyalty (R2 ¼91.3%). The direct effects of network quality, information and website support, privacy and security were found to be significant on attitudinal loyalty. However, none of service quality dimensions directly influenced behavioural loyalty. Attitudinal loyalty was the only direct determinant of behavioural loyalty. In order to test the mediation relationships and indirect effects of service quality dimensions on behavioural loyalty, two competing models, namely direct effect model and mediation model (i.e. the research model) were examined in line with the approach of Singh et al. (1994). The mediation relationship was tested on the basis of standardised residual co-variances and modification index values. Additionally, the bias corrected bootstrap was used in order to aid the confirmation of the mediation effect as recommended by Preacher and Kelley (2011). The results are provided in Table 6. All service quality dimensions, except for customer service and technical support, significantly influenced behavioural loyalty in the direct model. The mediation model test demonstrates that the explained variance in behavioural loyalty increased from 52.4% to 91.3% by incorporating attitudinal loyalty as a mediator. The mediation effect of attitudinal loyalty was confirmed in the mediation model which demonstrated the following: (1) higher variances, (2) a significant effect of the ISP's service quality dimensions (except for customer service and website support) on attitudinal loyalty, (3) insignificant relationship between service quality dimensions and behavioural loyalty, and (4) significantly influence of attitudinal loyalty on behavioural loyalty. In other words, attitudinal loyalty fully mediated the relationships between network quality, information and website support, privacy and security, and behavioural loyalty. Drawing upon the results of obtained by the bias-corrected bootstrap with 10,000 resamples, significant indirect effects on behavioural loyalty were found for all dimensions, except for customer service and website support. James and Brett (1984) assert that complete mediation happens when the effect of the independent variable(s) on the dependent variable completely disappears once the mediator is added as a 110 T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 Table 7 Results for the relationships between service quality dimensions, attitudinal and behavioural loyalty, coefficients among light, medium and heavy users groups. Path Light users Medium users Direct effects AL AL AL AL BL BL BL BL BL Indirect effects Total effects Direct effects Indirect effects Total effects Direct effects Indirect effects Total effects .213nnn .044 .458nnn .308nnn .033 .001 .023 .001 .945nnn o — NQ o — CS o — IW o — SP o— NQ o— CS o— IW o— SP o— AL Goodness of fit indices Heavy users .201nn .042 .433nn .291nn .166n .034 .398nn .386nnn .169nn .119n .323nn .084 .894nnn .234nn .04 .41nn .291nn .945nnn χ2(120) ¼ 261.439, CMIN/DF ¼2.179, GFI¼ .953, AGFI ¼.933, TLI ¼.976, CFI¼ .981, RMSEA ¼ .045, 90% CI ¼ .037:.052 .149n .031 .356nn .345nn .318nn .149n .034 .429nn .894nnn χ2(120)¼ 23.768, CMIN/DF ¼1.923, GFI ¼.951, AGFI¼ .930, TLI¼ .976, CFI¼ .982, RMSEA ¼ .044, 90% CI ¼ .036:.053 .279nnn .054 .364nnn .257nnn .019 .018 .003 .021 .959nnn .268nn .052 .349nn .246nn .287nn .069 .352nn .267nn .959nnn χ2(120) ¼362.058, CMIN/DF ¼3.017, GFI ¼ .960, AGFI ¼.943, TLI ¼.975, CFI¼ .980, RMSEA ¼ .045, 90% CI ¼.040:.050 Chi Square difference test Δχ2(24)¼39.765, p¼ .023 Notes: NQ¼ network quality; CS¼ customer service and technical support; IW¼ information and website support; PS ¼ privacy and security; AL¼ attitudinal loyalty; BL ¼behavioural loyalty; CFI ¼comparative fit index; GFI goodness of fit index; AGFI ¼adjusted goodness of fit index; RMSEA ¼root mean square error of approximation; TLI¼ The Tucker-Lewis coefficient; n pr .05 p r.01 nnn p r.001. nn Table 8 Chi square difference tests. Path AL AL AL AL BL BL BL BL o— o— o— o— o— o— o— o— NQ CS IW SP NQ CS IW SP Δχ2 df p 2.189 .047 .635 1.762 5.516 2.724 7.715 2.189 2 2 2 2 2 2 2 2 .335 .977 .728 .414 .063 .256 .021 .579 Notes: CS¼ Customer Service; NQ¼ Network Quality; IW ¼Information and Website support; SP ¼ Security and Privacy; AL¼ attitudinal loyalty; BL ¼ behavioural loyalty. predictor of the dependent variable. On this basis, it can be concluded that the relationship between network quality, information and website support, privacy and security, and behavioural loyalty was fully mediated by attitudinal loyalty. Hence, H1a, H1b, H3a, H3b, H4a, H4b, and H5 found support. In order to test H6 and H7 structural equation modelling with multigroup invariance testing was conducted. Results are shown in Table 7. Firstly, the possibility that a fully constrained model was invariant across groups was tested. This means specification of a model in which all factor loadings, all factor variances and all factor covariances were constrained equal across light, medium, and heavy users (Byrne, 2004). As indicated in Table 7, the difference of overall Chi square test (Δχ2(24) ¼39.765, p ¼.023) was statistically significant, indicating that some equality constraints did not hold across the light, medium and heavy users. To further verify the differences in each structural path between the three groups of users, it is recommended to separately examine structural models for the three segments. The models of Internet usage showed reasonable fit to the data (Table 7). A preliminary examination of the structural models of the three usage groups illustrates that customer service and technical support's effects on both behavioural and attitudinal loyalty remained insignificant among the three internet user groups. Network quality, privacy and security, and information and website support were significant predictors of attitudinal loyalty among the three groups. Noticeably, information and website support's total effects on behavioural loyalty was only found to be significant among light users and heavy users. In fact, though positively affecting attitudinal loyalty, information and website support had a negative direct impact on behavioural loyalty of medium users, which resulted in insignificant total effect. This signifies some potential differences in the relationships between four service quality dimensions and behavioural loyalty across the light, medium and heavy groups. In order to confirm differences in the structural paths of interest among three groups of Internet usage, a test for invariance of factor loadings was conducted. In this process, the unconstrained model was run, and eight paths (from four quality dimensions to attitudinal and behavioural loyalty) were fixed to be invariant in all groups to arrive at a constrained model (Cunningham, 2010). In the eight models, only one model testing the path from information and website support towards behavioural loyalty resulted in significant difference in the chi-square test (Table 8). This result confirms that the effects of information and website support on behavioural loyalty was not the same for people from different internet usage groups (Δχ2 (2) ¼7.715, p¼ .021). Therefore, H6 was rejected and H7 was partially supported. 5. Discussion This study advances the literature related to service quality as a key determinant of customer loyalty among different groups of customers in high tech services context. The contribution to extant research can be considered relatively robust as the research model explained a considerable proportion of the variance in the criterion variables. 5.1. ISPs' service quality dimensions, attitudinal and behavioural loyalty The results show that all service quality dimensions except for T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 customer service and technical support were positively related to attitudinal and behavioural loyalty. Information and website support was the predominant predictor of attitudinal loyalty, and behavioural loyalty of ISP customers. The role of privacy and security was also confirmed. These results support previous findings on the relationship between loyalty and information and website support (Toufaily et al., 2013), and privacy and security (Limbu et al., 2011). Interestingly, customer service and technical support was neither a significant predictor of attitudinal nor behavioural loyalty, contrasting with the finding of Santouridis and Trivellas (2010), which claimed that this dimension had the most significant impact on loyalty. This can be explained by the fact that customers in Asian culture typically often seek help from their friends or family for technical issues and tend to avoid communicating directly with the service providers due to ego concerns and value orientations (Neuliep, 2012). Another possible explanation could be that this dimension manifests its effect via other variables not included in this study, for example, satisfaction (Caruana and Ewing, 2010). Network quality significantly contributed to attitudinal and behavioural loyalty mirroring prior research by Miranda-Gumucio et al. (2012). This finding confirms that the basic need of an ISP customer is still the core service performance, i.e. internet downloading/uploading speed, signal stability and consistency. However, unlike findings of previous research, this dimension was not the most significant predictor of attitudinal and behavioural loyalty in this study. This phenomenon can be justified as network quality has become more stable and reliable among the ISPs, especially in big cities such as Bangkok, which no longer makes it a differentiating tool. 5.2. ISPs' service quality dimensions, loyalty among light, medium and heavy users In considering the usage profile of respondent, information quality had different effects on behavioural loyalty across different groups of internet users. Information and website support manifested its positive and significant impact on behavioural loyalty of light and heavy users through attitudinal loyalty. The current study was conducted in high-tech services in which customers are more likely to communicate with the company through websites. When customers feel that they have access to correct and adequate information, they are most likely to remain with the company. In contrast, information and website support directly had a significant negative effect on behavioural loyalty among medium users. Medium users are often more conversant with the internet than light users, which makes information support less necessary. In addition, they do not exhibit a need for considerable amount of information like heavy users or light users. In fact, Table 2 illustrates that medium users spent more time on emails and social media than on information search, in contrast to the other two groups of users. Also, whereas many pensioners were light users and a majority of heavy users were students, most of medium users were full time employees who are usually “money rich time poor”. Therefore, too much information might result in information overload for medium users and become dysfunctional (Jacoby 1984). In this case, it lessens the desire of customers to stay with their current service provider. These reasons together might explain why medium users do not attach much importance to information support in their quest towards behavioural loyalty. 5.3. The mediation role of attitudinal loyalty in the relationship between ISPs' service quality dimensions and behavioural loyalty Results from the testing of the mediation effects confirmed that 111 the relationship between the ISP's service quality dimensions, i.e. network quality, information and website support, privacy and security, and behavioural loyalty were fully mediated by attitudinal loyalty. The model which incorporated attitudinal loyalty as a mediator explained 92.9% of the variance in behavioural loyalty of the ISP's customers in this study. This finding reveals a salient and active role of attitudes of customers as a significant determinant for their repeat purchase behaviour. Fostering attitudinal loyalty increases the likelihood of the ISP becoming a supplier of choice, winning new customers through positive word of mouth and recommendations, and improving customer retention. For this reason, it is essential for service providers to investigate the factors that influence customer attitudes towards the firm. 6. Implications This study has both theoretical and practical implications. Overall service quality is widely considered as one of key factors determining customer attitudinal and behavioural loyalty. However the results in this study highlight that in fact different ISP's service quality dimensions generate different effects on the two perspectives of loyalty. It is recommended that research on service quality should not just focus primarily on network quality but on other service quality dimensions, as prominence has been given to this dimension in extant literature (He and Li, 2010; Kim and Yoon, 2004). This study revealed that service quality dimensions exert different effects on customer loyalty among the various groups of customers segmented based on their usage pattern, confirming the importance of customer segmentation. Moreover, this study is original in that it is the first of its kind which attempts to investigate the dimensions of an ISP's service quality, and their effects on loyalty in high-tech services. This study has evaluated ISPs' service quality dimensions which are identified as (1) network quality, (2) customer service and technical support, (3) information quality and website information support, (4) security and privacy. By enhancing service quality, firms can influence customers' behavioural and attitudinal loyalty, which are critical for an ISP's success and long term sustainability. As a result, ISPs will be able to reduce the current issues relating to customer switch and churn rate in the residential internet services market. Managers of ISPs should also be aware that, although the concept of service quality is multidimensional, not all dimensions contribute equally to loyalty. Different service quality dimensions exert different effects on customer loyalty. Information and website support was found to be the predominant predictor of both types of loyalty instead of the presumed network quality, signifying a potential change of behaviour and perception of ISP customers. This underscores a need of improving company websites and information support to the service subscribers, such as website accessibility and user-friendly interface, information accuracy, and timeliness. Although customer service was not found to be a direct antecedent of loyalty in this study, companies still need to maintain a satisfactory level of their customer service and technical support performance and keep up with the industry standards. In addition ISPs can promote a diversity of support forms, such as online and offline support. Particularly online support is convenient for customers experiencing simple issues or those perceiving direct face-to-face communication as being too intimate and personal. Furthermore, the finding confirms that internet customers are heterogeneous. In order to effectively and efficiently retain customers, an ISP needs to understand its customers' characteristics. It is critical to look into aspects that are relevant to the attitudinal and behavioural loyalty of specific customer groups in order to leverage its key resources and maximise 112 T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 profitability. By highlighting features of customers' interest, service providers can better enhance customer loyalty. 7. Limitations and future research directions There are several limitations of this study. Firstly it is possible that customers may have no choice but to stay with the service provider as the switching costs are high (Shirin and Puth, 2011). This explains why seemingly loyal customers, who have large volumes and high frequency of purchases, can quickly switch to other alternatives as switching costs decrease (Dick ad Basu, 1994; Shirin and Puth, 2011). Switching costs can be considered as the costs (including monetary, psychological and emotional costs) involved in moving from a service provider to another (Porter, 1980). Switching costs help companies to deal with inevitable variations in service quality (Jones et al., 2000). In other words, despite a decrease in service quality, customers might stay with the company because the perceived costs of changing to a new service provider exceed the potential benefits of switching (Lam et al., 2004). De Ruyter, Wetzels and Bloemer (1998) confirm the moderating effects of switching costs on the relationship between service quality and loyalty. As such, incorporating switching cost into future research could open up more opportunities to understand the relationship between service quality and loyalty. In addition, the choice of context (home internet services in Thailand) where the proposed conceptual model in this study was empirically tested may restrict the generalisability of the findings. In the interest of generalisation, replication in other countries and sectors to further evaluate the model would be a step towards addressing this limitation. Finally, this study examines the effects of the cognitive evaluation, i.e. service quality dimensions, on customer loyalty. An investigation on customers' affective evaluations, for instance trust, satisfaction, and commitment, as antecedents of loyalty and their interrelationships with service quality dimensions might be a fruitful area of research and also valuable in overcoming the limitation of this study. References ACMA, 2012. Internet use and social networking by young people Australian Communications and Media Authority. Available at: ⟨http://www.acma.gov.au/ webwr/_assets/main/lib310665/no1_internet_use_social_networking_young_ people.pdf⟩ (accessed 01.04.14). Anderson, J., Gerbing, W., 1988. Structural equation modelling in practice: a review and recommended two stage approach. Psychol. Bull. 27 (1), 5–24. Assael, H., 2005. A demographic and psychographic profile of heavy internet users and users by type of internet usage. J. Advert. Res. 45 (1), 93–123. Aydin, S., Özer, G., 2005. The analysis of antecedents of customer loyalty in the Turkish mobile telecommunication market. Eur. J. Mark. 39 (7/8), 910–925. Bandyopadhyay, S., Martell, M., 2007. Does attitudinal loyalty influence behavioural loyalty? A theoretical and empirical study. J. Retail. Consum. Serv. 14 (1), 35–44. Beerli, A., Martin, J.D., Quintana, A., 2004. A model of customer loyalty in the retail banking market. Eur. J. Mark. 38 (1/2), 253–275. Black, T.R., 1999. Doing quantitative research in the social sciences: An integrated approach to research design, measurement, and statistics. SAGE Publications, Thousand Oaks, CA. Byrne, B.M., 2004. Testing for multigroup invariance using AMOS graphics: a road less traveled. Struct. Equ. Model. 11 (2), 272–300. Canhoto, A.I., Clark, M., 2013. Customer service 140 characters at a time: the users’ perspective. J. Mark. Manag. 29 (5/6), 522–544. Carlson, J., O’Cass, A., 2011. Developing a framework for understanding e-service quality, its antecedents, consequences, and mediators. Manag. Serv. Qual.: Int. J. 21 (3), 264–286. Caruana, A., Ewing, M.T., 2010. How corporate reputation, quality, and value influence online loyalty. J. Bus. Res. 63, 1103–1110. Cassab, H., 2009. Investigating the dynamics of service attributes in multi-channel environments. J. Retail. Consum. Serv. 16, 25–30. Chae, M., Kim, J., Kim, H., Ryu, H., 2002. Information quality for mobile internet services: a theoretical model with empirical validation. Electron. Mark. 12 (1), 38–46. Chang, H., Chen, S., 2009. Consumer perception of interface quality, security, and loyalty in electronic commerce. Inf. Manag. 46 (7), 411–417. CIA, 2013. The World Factbook. Retrieved from: ⟨https://www.cia.gov/library/pub lications/the-world-factbook/geos/th.html⟩ (accessed 01.08.13). Cunningham, E., 2010. A practical guide to structural equation modelling using Amos, 1 ed. Statsline, Melbourne. De Ruyter, K., Wetzels, M., Bloemer, J., 1998. On the relationship between perceived service quality, service loyalty and switching costs. Int. J. Serv. Ind. Manag. 9 (5), 436–453. Dick, S., Basu, K., 1994. Customer loyalty: toward an integrated conceptual framework. J. Acad. Mark. Sci. 22 (2), 99–113. ETDA, 2013. Thailand Internet User Profile 2013 Electronic Transactions Development Agency. Available at: ⟨http://www.etda.or.th/etda_website/files/system/ InternetUserProfile_2013_14.11.13.pdf⟩ (accessed 15.05.2014). Finn, A., Wang, L., Frank, T., 2009. Attribute perceptions, customer satisfaction and intention to recommend e-services. J. Interact. Mark. 23, 209–220. Flint, J., Blocker, C., Boutin, P., 2011. Customer value anticipation, customer satisfaction and loyalty: an empirical examination. Ind. Mark. Manag. 40, 219–230. Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with unobservable variables and measurement error. J. Market. Res., 39–50. Ganguli, S., Roy, S.K., 2010. Service quality dimensions of hybrid services. Manag. Serv. Qual. 20 (5), 404–424. Ganguli, S., Roy, S.K., 2011. Generic technology-based service quality dimensions in banking. Impact on customer satisfaction and loyalty. Int. J. Bank. Mark. 29 (2), 168–189. Hair, J., Anderson, R., Tatham, R., Black, C., 1998. Multivariate Data Analysis, 5 ed. Upper Saddle River, NJ, Prentice Hall. Han, H., Hyun, S.S., 2012. An extension of the four-stage loyalty model: the critical role of positive switching barriers. J. Travel Tour Mark. 29 (1), 40–56. He, H., Li, Y., 2010. Key service drivers for high-tech service brand equity: the mediating role of overall service quality and perceived value. J. Mark. Manag. 27 (1/2), 77–99. Hong, P., Cho, H., 2011. The impact of consumer trust on attitudinal loyalty and purchase intentions in B2C e-marketplaces: Intermediary trust vs. seller trust. Int. J. Inf. Manag. 31 (5), 469–479. Hsieh, P., 2013. Perceived opportunism (PO) in e-return service encounters. Manag. Serv. Qual. 23 (2), 96–110. Jacoby, J., 1984. Perspectives on information overload. J. Consum. Res. 10 (4), 432–435. Jacoby, J., Chestnut, R.W., 1978. Brand Loyalty: Measurement and Management. John Wiley & Sons, New York. Jaiswal, A.K., Niraj, R., 2011. Examining mediating role of attitudinal loyalty and nonlinear effects in satisfaction-behavioral intentions relationship. J. Serv. Mark. 25 (3), 165–175. Jaiswal, A.K., Niraj, R., Venugopal, P., 2010. Context-general and context-specific determinants of online satisfaction and loyalty for commerce and content sites. J. Interact. Mark. 24, 222–238. James, L.R., Brett, J.M., 1984. Mediators, moderators, and tests for mediation. J. Appl. Psychol. 69, 307–321. http://dx.doi.org/10.1037/ 0021-9010.69.2.307. Jin, B., Kim, J., 2010. Multichannel versus pure e-tailers in Korea: evaluation of online store attributes and their impacts on e-loyalty. Int. Rev. Retail., Distrib. Consum. Res. 20 (2), 217–236. Jones, Michael A., Mothersbaugh, David L., Beatty, Sharon E., 2000. Switching barriers and repurchase intentions in services. J. Retail. 76, 259–274 (Summer). Kim, G., 2015. South Asia ISPs Have To Move Faster As Competition Builds, Spectrum Futures. ⟨http://spectrumfutures.org/south-asia-isps-have-to-move-fas ter-as-competition-builds/⟩ (accessed 18.08.15). Kim, H., Yoon, C., 2004. Determinants of subscriber churn and customer loyalty in the Korean mobile telephony market. Telecommun. Policy 28 (9), 751–765. Kim, H., Niehm, L.S., 2009. The impact of website quality on information quality, value, and loyalty intentions in apparel retailing. J. Interact. Market. 23 (3), 221–233. Lai, F., Griffin, M., Babin, B., 2009. How quality, value, image, and satisfaction create loyalty at a Chinese telecom. J. Bus. Res. 62, 980–986. Lam, S.Y., Shankar, V., Erramilli, M.K., Murthy, B., 2004. Customer value, satisfaction, loyalty, and switching costs: an illustration from a business-to-business service context. J. Acad. Mark. Sci. 32 (3), 293–311. Limbu, Y.B., Wolf, M., Lunsford, D.L., 2011. Consumers’ perceptions of online ethics and its effects on satisfaction and loyalty. J. Res. Interact. Mark. 5 (1), 71–89. Lin, J., 2007. Online stickiness: its antecedents and effect on purchasing intention. Behav. Inf. Technol. 26 (6), 507–516. Mazzoni, C., Castaldi, L., Addeo, F., 2007. Consumer behavior in the Italian mobile telecommunication market. Telecommun. Policy 31 (10/11), 632–647. Miranda-Gumucio, L., Gil-Pechuan, I., Palacios-Marques, D., 2012. An Exploratory Study of the Determinants of Switching and Loyalty in Prepaid Cell Phone Users. An Application of Concept Mapping. Springer-Verlag, Berlin Heidelberg http://dx.doi.org/10.1007/s11628-012-0177-3. NECTEC, 2015. Internet User and Statistics in Thailand, National Electronics and Computer Technology Centre. Available at: ⟨http://internet.nectec.or.th/web stats/home.iir?Sec ¼ home⟩ (accessed 01.08.15). Neuliep, J.W., 2012. Intercultural Communication: A Contextual Approach. SAGE Publications, Inc., St. Norbert College. Nunnally, J., 1978. Psychometric Theory. Mcgraw Hill, New York. Oliver, L., 1999. Whence consumer loyalty. J. Mark. 63 (1), 33–44. Özgüven, N., 2011. Analysis of the relationship between perceived security and customer trust and loyalty in online shopping. Chin. Bus. Rev. 10 (11), 990–997. T.N. Quach et al. / Journal of Retailing and Consumer Services 29 (2016) 104–113 Parasuraman, A., Zeithaml, V.A., Berry, L.L., 1985. A conceptual model of service quality and its implications for future research? J. Mark. Manag. Vol. 49 (No. 1), 41–50. Parasuraman, A., Zeithaml, V.A., Malhotra, A., 2005. ES-QUAL a multiple-item scale for assessing electronic service quality. J. Serv. Res. 7 (3), 213–233. Porter, Michael E., 1980. Competitive Strategy: Technologies for Analyzing Industries and Competitors. Free Press, New York. Preacher, K.J., Kelley, K., 2011. Effect size measures for mediation models: quantitative strategies for communicating indirect effects. Psychol. Methods 16 (2), 93–115. Prentice, C., 2014. Who stays, who walks, and why in high-intensity service contexts. J. Bus. Res. 67 (4), 608–614. http://dx.doi.org/10.1016/j. jbusres.2013.02.044. Ratnasingham, P., 1998. The importance of trust in electronic commerce. Internet Res. 8 (4), 313–321. Roca, J., García, J., Vega, J., 2009. The importance of perceived trust, security and privacy in online trading systems. Inf. Manag. Comput. Secur. 17 (2), 96–113. Rod, M., Ashill, N.J., 2013. The impact of call centre stressors on inbound and outbound call-centre agent burnout. Manag. Serv. Qual. 23 (3), 245–264. Rossi, P.H., Wright, J.D., Anderson, A.B. (Eds.), 2013. Handbook of Survey Research. Academic Press, New York. Santouridis, I., Trivellas, P., 2010. Investigating the impact of service quality and customer satisfaction on customer loyalty in mobile telephony in Greece. Total Qual. Manag. J. 22 (3), 330–343. Seo, D., Ranganathan, C., Babad, Y., 2008. Two-level model of customer retention in the US mobile telecommunications service market. Telecommun. Policy 32, 182–196. Shirin, A., Puth, G., 2011. Customer satisfaction, brand trust and variety seeking as determinants of brand loyalty. Afr. J. Bus. Manag. 5 (30), 11899–11915. Singh, J., Goolsby, J.R., Rhoads, G.K., 1994. Behavioral and psychological consequences of boundary spanning burnout for customer service representatives. J. Mark. Res. 31 (4), 558–569. Tabachinick, B.G., Fidell, L.S., 2001. Using multivariate statistics. Allyn and Bacon, Boston. Teddlie, C., Yu, F., 2007. Mixed methods sampling a typology with examples. J. Mix Methods Res. 1 (1), 77–100. Telenor, 2015. Towards a Connected world Telenor Group. Available at: ⟨http:// www.telenor.com/wp-content/uploads/2012/03/Towards-a-Connected-World1MB.pdf⟩ (accessed 01.04.15). Thaichon, P., Lobo, A., Prentice, C., Quach, T.N., 2014. The development of service quality dimensions for internet service providers: Retaining customers of different usage patterns. J. Retail. Consum. Serv. 21 (6), 1047–1058. Thaichon, P., Quach, T.N., 2013. Integrated marketing communications and their influences on brand loyalty: a Thai perspective. International Journal of Studies in Thai Business. Soc. Cult. 2 (1), 55–79. Thaichon, P., Lobo, A., Mitsis, A., 2012. Investigating the antecedents to loyalty of Internet service providers in Thailand: Developing a conceptual model. Paper presented at the Australian and New Zealand Marketing Academy Conference, Adelaide, Australia. Toufaily, E., Ricard, L., Perrien, J., 2013. Customer loyalty to a commercial website: Descriptive meta-analysis of the empirical literature and proposal of an integrative model. J. Bus. Res. 66, 1436–1447. True, 2013. True Corporation Plc At a Glance. Available at:: ⟨http://true.lis 113 tedcompany.com/misc/PRESN/20130831-TRUE-glanceAugust2013-02.pdf⟩ (accessed 01.04.14). Vlachos, A., Vrechopoulos, P., 2008. Determinants of behavioral intentions in the mobile internet services market. J. Serv. Mark. 22 (4), 280–291. Wang, Y., Lo, H., Yang, Y., 2004. An integrated framework for service quality, customer value, satisfaction: evidence from China’s telecommunication industry. Inf. Syst. Front. 6 (4), 325–340. Wedel, M., Kamakura, W., 2003. Market Segmentation. Conceptual and Methodological Foundations. Kluwer Academic Publishers, Norwell, MA. Weinstein, A., 1987. Market Segmentation. Probus Publishing, Chicago, IL. Wolfinbarge, M., Gilly, C., 2003. eTailQ: dimensionalizing, measuring and predicting etail quality. J. Retail. 79 (3), 183–198. Zeithaml, V.A., Berry, L.L., Parasuraman, A., 1996. The behavioral consequences of service quality. J. Mark. 60 (2), 31–46. Zhang, R., Li, X., Zhang, Y., 2010. Service quality, customer satisfaction and customer loyalty of mobile communication industry in China. J. Glob. Acad. Mark. Sci. 20 (3), 269–277. Zhao, Y.L., Benedetto, C.A.D., 2013. Designing service quality to survive: empirical evidence from Chinese new ventures. J. Bus. Res. 66, 1098–1107. Thu Nguyen Quach holds a Master's Degree in Marketing and is currently working towards a PhD programme in Marketing in the Faculty of Business & EnterpriseSwinburne University of Technology, Melbourne, Australia. Her research interests are in the area of services marketing, marketing research, consumer behaviour and relationship marketing. Thu's research has been published in the Journal of Retailing and Consumer Services, Journal of Business and Industrial Marketing, and Services Marketing Quarterly, among others. Paramaporn Thaichon is an Assistant Professor of Marketing at the S P Jain School of Global Management, Sydney, Australia. His research has been published in leading marketing journals throughout Europe, North America and Australasia including but not limited to the Journal of Retailing and Consumer Services, Journal of Business and Industrial Marketing, Asia Pacific Journal of Marketing and Logistics, and Journal of Relationship Marketing. Dr Charles Jebarajakirthy is based in the Faculty of Business and Law at the Swinburne University of Technology, Melbourne, Australia. His research interests are in the areas of market orientation, consumer behaviour, social marketing and corporate social responsibility. Charles's research has been published in the Journal of Retailing and Consumer Services, Asia Pacific Journal of Marketing and Logistics, International Journal of Consumer Studies and Journal of Young Consumers, among others.
