The Role of Cultural Differences in Customer Retention: Evidence from the High-Contact Service Industry

Customer Loyalty

Based on previous reviews, Solnet and Kandampully (2008) emphasized a hotel’s customer base as an asset of the highest value that is imperative for success. Customers act as brand ambassadors in an industry that lacks unique differentiating attributes and have far-reaching influence within the marketplace. Comparing loyal and nonloyal customers, Bowen and Chen (2001) had found that loyal customers tend to take a shorter time to decide on intended purchases as they are driven by preference instead of extended searches for better deals. This finding is in line with Skogland and Siguaw’s (2004) proposition that customer loyalty signifies faithfulness toward a service or product. Following the above discussion, this study chose customer loyalty as the outcome of interest in view of its implications pertaining price insensitivity. Specifically, loyal tourists have an increased willingness to spend more or change dates for their intended travel due to their preference for a favored hospitality service provider or accommodation.

Although the title of this article carries the term customer retention, loyalty was tested as the dependent variable of its framework. Singh and Khan (2012) referred to loyalty as the disposition of a particular customer to acquire an organization’s tangible and intangible offerings against its competitors. On the other hand, retention constitutes the emotional aspects induced through actual experience and corporate image; the connections forged by these emotions then support loyalty (Magatef & Tomalieh, 2015). Keiningham et al. (2007) concurred that retention can occur through mere recommendations from acquaintances but is insufficient to cultivate loyalty on its own. Comprehensive definitions thus describe loyalty as relationship development and retention as relationship continuation. By outdoing competitors in term of products, pricing, and delivery, an organization can cultivate customer loyalty by creating positive customer experiences which lead to repeat purchases (Cheng, Gan, et al., 2019; X. J. Lim et al., 2018; Y. M. Lim et al., 2019; Magatef & Tomalieh, 2015). Moreover, as the main element of customer retention, recurring clientele to an organization is closely related to relationship marketing (Hennig-Thurau & Hansen, 2013).

Overall, retention is viewed as the process of achieving long-term loyalty. In this study on how relationship marketing, service quality, and corporate image promote customer satisfaction and ultimately loyalty in the high-contact service industry, the term customer retention therefore represents the proposed framework in its entirety.

Relationship Marketing

Relationship marketing was defined by Beetles and Harris (2010) as a marketing paradigm that emphasizes the development and retention of long-lasting customer relationships. It is further regarded by Wu and Lu (2012) as a continuous process aimed to transfer long-term value to customers. Palmatier (2008) then proposed that sustainable customer relationships are an influencing factor of service firms’ long-term profitability. As reciprocity is necessary for organizations and customers to maintain a relationship, the success of such a relationship is dependent on both parties’ engagement (W. G. Kim & Cha, 2002). It should be noted that customers’ perceptions are determined through their actual experience of a particular service; relationship marketing therefore calls for hotel operators to personalize their services to enhance customer experiences (Narteh et al., 2013). In this manner, relationship marketing transcends the marketing mix’s short-term transactional gains by adopting a customer-oriented approach that offers personalized experiences and greater accessibility to achieve continued engagement and retention (W. G. Kim & Cha, 2002).

Relationship marketing is an irreplaceable asset in today’s digital environment, as it engenders loyalty from customers of diverse national and societal backgrounds and motivates destination-based buzz through well-targeted information across integrated channels (Arminda & Sergio, 2017). Accordingly, positive relationships have been exhibited between trust and customer loyalty, social responsibility and relationship engagement, as well as specialized treatment and customer appreciation (Chou & Chen, 2018). While customer loyalty holds a dominant role in developing lasting buyer–seller relationships in service organizations, differences prevail between high-contact and moderate- or low-contact services, whereby in the latter, social responsibilities and personalized treatment are deemed secondary to patronage (Srihadi & Setiawan, 2015). Kinard and Capella (2006) further highlighted the greater benefits of customer loyalty in high-contact organizations with customizable services and encouraged programs capitalizing on customer involvement in these firms. This, however, was rebutted by E. L.-Y. Li et al. (2017), who found that the indirect relationship between customer trust and loyalty towards a service provider is more apparent in low-contact services.

Regardless of service type, every employee–customer interaction has substantial potential to affect customer satisfaction (Hennig-Thurau & Hansen, 2013). As demonstrated by Berezan et al. (2017), relationship-building programs that increase information exchange and customers’ knowledge directly enhance customers’ positive feelings toward service providers, which indirectly boost customer satisfaction. Additionally, the study by Al-Nazer (2010) has shown that relationship marketing directly affects customer loyalty by stimulating customers’ connectedness, confidence, networking, satisfaction, and engagement. As such, the importance of relationship marketing is evident in its impact on a satisfactory customer experience and customer loyalty.

Recent studies on the effect of relationship marketing on customer satisfaction (e.g., Panjaitan & Laely, 2017) and customer loyalty (e.g., Dedeoğlu et al., 2015) have mainly explored the financial and banking sector, in view of its speculative nature and its management of customers’ private information. Although today’s marketing environment has advanced in many ways, relationship marketing continues to be a critical part of current and future marketing efforts in various business sectors (Berezan et al., 2017). Therefore, given the inconsistent findings on customer loyalty across service firms’ contact levels and a lack of exploratory research in the tourism sector, the current study aspired to investigate relationship marketing as an independent variable affecting various customer-related outcomes in the tourism industry.

Perceived Service Quality

Perceived service quality represents a customer’s overall assessment or impression of the relative inferiority or superiority of a specific service (Parasuraman et al., 1988). As outlined by Tsoukatos and Rand (2006), one’s service evaluation is commonly dependent on elements of the service’s delivery process, production, and outcome. Service quality is perceived as superior when customers’ quality expectation is fulfilled, which can be detrimental if customers hold unrealistically high expectations despite excellent service (Grönroos, 1988). In other words, perceived service quality reflects the comparison of customers’ actual service experience to their prior expectations (Parasuraman et al, 1988). Due to close mutual interactions, employees’ service performance carries a significant impact on the fulfilment of customer expectations; likewise, customers’ involvement in a service affects their perceptions of the service’s quality (Tsoukatos & Rand, 2006).

Perceived service quality has been commonly researched in terms of its direct and indirect influence on customer satisfaction (e.g., Miranda et al., 2018) and customer loyalty (e.g., Marković & Janković, 2013). Dedeoğlu et al. (2015) stated that perceived service quality affects repurchase intention (i.e., loyalty) through the development of value perception. Additionally, a study by Hu et al. (2009) suggested a positive association between service quality and image, which ultimately influences customer satisfaction. In the context of high-contact services, studies in the banking and hotel sectors have demonstrated customers’ perception of service quality as a significant predictive factor of positive customer outcomes (Ali et al., 2017). Perceived quality established through customer experience is indeed shown to influence the corporate image of hotels and resorts (W. J. Chen & Chen, 2014). Therefore, enhanced perceptions of service quality via effective relationship marketing would directly promote customer retention and loyalty (Cham et al., 2016; Ryu & Lee, 2017).

From another perspective, Holmlund and Kock (2006) proposed a favorable service experience as an essential precondition of a constructive business relationship. Customers’ service perceptions act as the foundation of their trust and pleasure, which in turn encourage purchasing behaviors (Ryu & Lee, 2017). A direct association between perceived service quality and customer–organization relationship quality was also proposed by Su et al. (2016) within the context of tourism. Moreover, both service quality and relationship quality emerged as independent variables that connect service characteristics to revisits (Sadeghi et al., 2017). Notably, relationship marketing is shown to precede relationship quality and ensure positive word-of-mouth (Bennett & Barkensjo, 2005). However, recent studies have mainly appraised the influence of service quality on relationship-building, while neglecting the direct influence of relationship marketing on perceived service quality. Hence, direction of the causal influence between both variables remains ambiguous. Based on the above discussion, the current study proposed an inverse relationship in the high-contact tourism industry. The following hypotheses were thus postulated:

Corporate Image

According to Christou (2003), corporate image signifies an organization’s behavioral, psychological, and physical components, constituting elements such as services offered and perceptions of service providers. In the literature on service quality, corporate image is interchangeably termed corporate reputation (Chun, 2005). Within the context of tourism, a hotel’s image is ultimately formed through customers’ experiences; thus, it is crucial to identify the antecedents of hotels’ corporate image and their influence on other constructs (Dedeoğlu et al., 2015). For example, hotels and resorts in Thailand attract new customers and promote long-term relationships by creating a favorable corporate image using relationship marketing.

The intangible nature of services demands greater emphasis on creating a superior corporate image, as a positive image in combination with effective marketing entails more benefits than just marketing efforts for service firms (Onkvisit & Shaw, 1989). This may be challenging as the intangibility of services poses difficulties in evaluating the relationship of image with perceived service quality (Wilkins et al., 2009). Nevertheless, in the area of service improvement, corporate image has surfaced as an advanced basis for firms to map the effectiveness of public apologies and conflict resolutions in improving their service as well as to measure poststrategy customer satisfaction (Mostafa, 2017). Essentially, corporate image plays an intermediary role between organizational efforts and customers’ perceptions.

On one end of the spectrum, corporate image is viewed as an outcome of organizational capabilities in the areas of service quality and relationship marketing (W. J. Chen & Chen, 2014; Hu et al., 2009). Consistent with this view, W. J. Chen and Chen (2014) discovered the significant influence of relationship marketing and perceived service quality on hotels’ corporate image, where service experiences provide the foundations for customers’ perceptions. Likewise, relationship marketing, being an important determinant of customers’ trust and satisfaction, has proven beneficial in forming a favorable organizational image (Christian et al., 2017). An integrated model was also proposed by C. K. Li and Hung (2009), which posited marketing tactics as direct antecedents of firm image and firm image as a predictor of customer loyalty. Despite numerous studies placing relationship marketing, service quality, and image as independent factors affecting the service landscape (Anggraini et al., 2016; Christian et al., 2017), the general consensus has been that a positive corporate image is a potential outcome of service providers’ relationship marketing behaviors geared toward meeting customer expectations.

At the other end of the spectrum, corporate image has been found to positively influence customer loyalty (Cham & Easvaralingam, 2012). This is supported by Wang and Wu (2012), who revealed corporate image as a direct antecedent to customer loyalty, though the strength of the association relies on customers’ switching costs. Similarly, findings by Cheng and Rashid (2015) have indicated that corporate image is a direct predictor of customer satisfaction and loyalty. Cham and Easvaralingam (2012) further established that service quality has a stronger effect on customer loyalty than corporate image, though both are significant. Apart from direct effects, Engriani et al. (2019) proposed that organizational image, service quality, and perceived value indirectly influence loyalty and actual purchase intention through experience-based satisfaction. Kandampully et al. (2011) supported the indirect relationship between image and loyalty by way of excellent service. Based on these findings, the influence of image on loyalty appears to be dictated by various underlying mechanisms.

Undeniably, customers tend to favor firms with positive corporate images (Kandampully et al., 2011; X. J. Lim et al., 2020). Azhar et al. (2018) explored the indirect effects of organizations’ service quality, environmental layout, customer-oriented conflict resolutions, network communications, and trust on customer satisfaction through the mediation of corporate image. In contrast, the current study acknowledged the significance of image as an intermediary by investigating it as a possible outcome of service quality and relationship marketing, as well as a potential predictor of customer loyalty. In the context of the high-contact industry, the following hypotheses were proposed:

Customer Satisfaction

Cadotte et al. (1987) explained that customer satisfaction is based on the expectancy-disconfirmation theory, which puts forth that customers form feelings of dissatisfaction or satisfaction after experiencing a service. This concept was further expanded by Smith et al. (1999) to include three dimensions of service evaluations, that is, perceived justice, procedural justice, and interactional justice. These dimensions link expectations to disconfirmation of service quality in situations of poor service. Harris and Goode (2004) subsequently asserted the importance of evaluating satisfaction as an overall perception rather than a transaction-specific event due to its impact on businesses.

In the service industry, satisfaction has often been shown to be a direct result of service quality (Jiang & Zhang, 2016; Lacap et al., 2021; Meesala & Paul, 2018). Service practitioners have thus made radical attempts to improve service quality to stimulate higher customer satisfaction (Marković & Janković, 2013). According to Kasiri et al. (2017), (1) service quality can be enhanced through both service uniformity and personalization, (2) uniformity possesses a greater influence on service quality compared with personalization, (3) a service’s practical characteristics have stronger impacts on satisfaction than service mechanisms, and (4) satisfaction derives long-term customer loyalty. Thus, while service quality might not directly predict loyalty, satisfaction may be an effective bridge linking these constructs (Jiang & Zhang, 2016). Nonetheless, contextual differences have emerged between service industries (e.g., hospitals and hotels), as the attributes that lead to customer satisfaction vary across these industries (Cham, Cheng, Low, & Cheok, 2020; Meesala & Paul, 2018).

Rahimi and Kozak (2015) claimed that management efforts in strengthening customer relationships enhance customer satisfaction and experience. Studies in the hotel sector further verify relationship marketing activities as an important factor in guest satisfaction and customer relationship continuity, among other benefits (Gunjan, 2018; W. G. Kim & Cha, 2002). In reference to customer loyalty, relationship marketing has a direct influence on satisfaction (Martínez & Bosque, 2013), which subsequently motivates loyalty. Indeed, past studies have consistently reported that customer loyalty is formed based on customer satisfaction (e.g., Cheng, Cham, et al., 2019; Martínez & Bosque, 2013). However, Leveren and Liljander (2006) suggested otherwise, finding that service-based relationships do not improve customer loyalty regardless of their transactional level, and that customer satisfaction achieved through enriched relationships is not necessary to retain profitable customers. Rather, relationship quality is implied as a mediating variable between effective relationship marketing and customer satisfaction (Santouridis & Veraki, 2017).

In meeting customers’ expectations, transactional value and product offerings remain crucial contributors to tourists’ satisfaction (Rahimi & Kozak, 2015). Drawing from the expectancy-disconfirmation theory, service providers’ efforts toward meeting and surpassing customer expectations would yield customer satisfaction, which in turn has a long-term impact on customer loyalty in the high-contact service industry. Therefore, the following hypotheses were developed:

The Moderating Effect of Cultural Differences

Customers’ cultural upbringing, as proposed by Reisinger and Turner (1998), plays an essential role in determining their perceptions of service providers, as culture affects social interaction and individual satisfaction. Academic endeavors on culture in the hospitality industry have prioritized modernized societies such as Western and Asian communities (M. Li, 2014). For instance, Paris et al. (2014) noted differences in values upheld by Asian and Western tourists, though they share some common values. Cultural foundations thus affect tourists’ perceptions at the individual level, especially their perception of personnel and service quality (Cham et al., 2021; Deng et al., 2019).

It has been reported that cultural background and geographical location drive travel patterns and destination planning (Paris et al., 2014). The consonance between both factors is hence critical in understanding the influence of an individual’s culture on his or her destination image and consequent visit decision (Murphy et al., 2007). Loyalty to a tourist destination is thus attainable, though Western tourists display greater individual-destination congruency than Asian tourists (W. Kim & Malek, 2016). This echoes the findings of Choi and Chu (2000) that transactional benefits outrank accommodation quality among Asian tourists, but not among their Western counterparts. However, contradictory evidence was presented by Williams et al. (2017), who found that Asians are more emotionally oriented, while Westerners are more economically oriented.

Additionally, Reisinger and Turner (2002) suggested that Asian tourists’ perceptions go beyond service qualities to include cultural aspects, social interactions, and relationships. Seo (2012) then revealed that in their hotel experience, Western tourists prefer service-centric elements (i.e., service compatibility, reliability, customizability, and punctuality), whereas Asian tourists prefer elements of association and connection (i.e., staff politeness, communication, attentiveness). Therefore, cultural background has undeniable diverging effects on individuals’ motivation to visit specific destinations (Jin et al., 2016). Based on this discussion, the following hypotheses were developed:

Based on the thorough discussions presented in the earlier sections, the research model tested in the current study is presented in Figure 1.

OPEN IN VIEWER

Figure 1

Research Model

Measurements and Questionnaire Development

A paper-based questionnaire was used to assess the hypotheses of this study. The measurement instruments employed to operationalize the constructs were adapted from the existing literature as suggested by Luarn and Lin (2005), as the scales’ validity has already been established. Although relationship marketing is a multidimensional concept contingent on customers’ desires, it was treated as an independent construct in this study based on the extant literature. Indeed, scholars have investigated this variable among multiple accommodation service providers in the Asia-Pacific region, further highlighting the possibility of deviated findings among different cultural backgrounds. Additionally, relationship marketing, as a unidimensional construct, was examined in various later studies on relationship management in the hospitality sector (e.g., Chan et al., 2018; Sota et al., 2020). Since this study targeted the hotel industry, relationship marketing was measured using a 13-item scale by W. G. Kim et al. (2001), which assesses a hotel’s overall effort in maintaining a relationship with their customers through their marketing activities. The instrument had been proven reliable in this context, thus solidifying relationship marketing as a viable construct.

Perceived service quality was measured on a five-item scale adapted from Cham et al. (2021), Cheng et al. (2014), Hapsari et al. (2017), and Suhartanto (2011) in the high-contact service industry context. It was operationalized into the dimensions of assurance, responsiveness, reliability, tangibility, and empathy of the service experience. Next, corporate image was measured with a three-item scale developed by Ladhari et al. (2011) to gauge guests’ overall impression and perceived image of their respective hotels. For customer satisfaction, the four-item scale proposed by Olorunniwo et al. (2006) was adapted for this study to measure hotel guests’ satisfaction based on their affective responses toward their service experience. Finally, customer loyalty was measured using a seven-item scale adapted from Skogland and Siguaw (2004), which appraises the likelihood of hotel guests’ repeated purchase, patronage to the hotel, and positive recommendations to close acquaintances (e.g., friends and family members). All the measurement items used in this study are presented in the online Supplemental Appendix.

All items were rated on a 6-point Likert-type scale ranging from 6 = Strongly Agree to 1 = Strongly Disagree. The decision to use a 6-point Likert-type scale is supported by the works of Infosurv (2006), Johns (2010), and Kulas et al. (2008). These scholars posit that the midpoint on an odd number scale (e.g., “3” on a 5-point scale) is commonly used by respondents to avoid sensitive questions and to provide socially desirable answers. This suggests that the midpoint is a “dumping ground” though respondents possess clear answers to questions, which distorts the impartiality of collected data. Infosurv (2006) thus proposed that a research sample must include individuals with real opinions or experiences pertaining to the investigated issue to reflect the rarity of a neutral response in their answers. Given the high-credence nature of hotel selection, guests are expected to be able to fully express their views on their experienced service. This study, therefore, adopted a 6-point Likert-type scale to assess respondents’ perceptions, which is consistent with the recent tourism studies by Andrades-Caldito et al. (2014), Cham et al. (2021), and G. Chen et al. (2020).

The developed questionnaire was pretested with 15 experts from the industry and academia and subsequently pilot-tested among 60 hotel guests prior to actual data collection. The preliminary responses obtained from both tests evinced that the questionnaire was well designed with no ambiguities. Face validity was thus established.

Procedure and Sample

To test the proposed hypotheses, customers’ perceptions toward the hotel industry were determined as the most relevant context to exemplify the nature of high-contact services. Apart from tourism being fairly customer-/tourist-centered, multiple previous studies on high-contact services have focused on appraising hospitality dimensions such as service development (Ottenbacher et al., 2006), service quality (Briggs et al., 2007), and critical service components (Stefano et al., 2015). Based on the rationale that satisfaction arises from performance that is equal to or higher than expectations, hotels involve in-depth engagement between employees and customers to ensure personalized solutions to customers’ various requests without forgoing hotels’ consistent standards in and above accommodations (i.e., timely transportation, suitable meals, customizable amenities, and resourceful entertainment). This justifies the hotel industry as a high-contact service industry, as corroborated by the Association for Supply Chain Management (2020). As such, it is pertinent to explore hotel services in this study on high-contact services.

The target respondents of this study were international tourists who, at the time of data collection, were visiting Malaysia for a vacation and had engaged hotel services during their stay. Data were collected at the Kuala Lumpur International Airport and the Kota Kinabalu International Airport in Malaysia. These airports were selected as they are the busiest airports in Malaysia with a significant volume of passengers (Joseph, 2018). Both quota and purposive sampling approaches were used to select 400 international tourists. Quota sampling was undertaken by evenly dividing the 400 respondents between both research locations. Meanwhile, purposive sampling was used to identify appropriate respondents who fulfilled two key criteria: (1) must be an international tourist and (2) must have stayed at least one night at a hotel in Malaysia during the visit. The questionnaires were personally distributed to tourists who fulfilled these requirements. According to Malhotra (2004), the researcher-administered survey prevails as the most effective method to create a cooperative atmosphere among respondents, which improves response rates and reduces uncertainties.

On an average, it took 15 minutes for the respondents to complete the questionnaire. Out of the 400 distributed questionnaires, 391 were usable while nine were rejected due to incompleteness and suspicious response patterns (e.g., alternating extreme pole responses, diagonal lining, and straight-lining). As for sample size, Saunders et al. (2012) argued that a sample size of 300 is sufficient to represent a large population, while a sample size of 200 is needed for structural equation modeling (SEM) analysis (Hair et al. 2010). The use of the G*Power software also shows that the sample size of 391 exceeded the minimum sample size of 129 based on an effect size of 0.15 at a 95% power level (Faul et al., 2009). Hence, the final sample size of 391 in the present study was sufficient (Memon et al., 2020). The collected data were analyzed using the Statistical Package for the Social Sciences Version 25 and the Analysis of Moment Structures Version 24.

Table 1 reports the demographic profile of the study respondents. In terms of gender, the respondents were almost equally distributed between females (52.6%) and males (47.4%). A majority of the respondents were married (77%) and aged 25 years and younger (28.3%). The top four tourist nationalities were Singapore (13.6%), Indonesia (11.3%), the United Kingdom (10.7%), and China (9.7%). Most respondents had visited the researched country between one and three times (87.8%). In terms of cultural background, about half (50.6%) the respondents were Asians, while the remaining half were Westerners.

OPEN IN VIEWER

Table 1

Respondents’ Demographic Profile

Variable	Percentage
Gender
Female	52.6
Male	47.4
Age (years)
25 and younger	28.3
26-35	28.1
36-45	25.3
46-55	15.8
56-65	2.0
65 and older	0.5
Marital status
Married	77.0
Single	18.4
Divorced	2.3
Widowed	2.0
Others	0.3
Nationality
Singapore	13.6
Indonesia	11.3
United Kingdom	10.7
China	9.7
United States	9.5
Australia	5.9
France	5.4
Thailand	5.1
Germany	4.3
Brunei	3.3
Canada	3.1
Netherlands	2.8
India	2.6
Saudi Arabia	2.0
Italy	2.0
Korea	1.3
Denmark	1.3
Pakistan	1.0
Sweden	1.0
Philippines	0.8
New Zealand	0.8
Japan	0.5
Kazakhstan	0.5
Qatar	0.5
Spain	0.5
Iran	0.3
Russia	0.3
Number of visit(s) to Malaysia
1-3 times	87.8
4-6 times	9.2
7-9 times	2.6
More than 10 times	0.5
Cultural background
Asians	50.6
Westerners	49.4

Measurement Model Assessment

Confirmatory factor analysis (CFA) was employed to test the validity of the measurement items for all the variables. The CFA examined the fit of the measurement model and validated both convergent and discriminant validity of the study constructs. As per Hair et al. (2010), the model fit was determined based on various indices. The common indices used to assess model fit are normed chi-square (χ²/df), root mean square error of approximation (RMSEA), goodness of fit (GFI), Tucker–Lewis index (TLI), and parsimony normed fit index (PNFI). Hair et al. (2010) recommended that a model be considered fit if χ²/df is less than 3, RMSEA is less than .08, GFI is greater than .90, TLI is greater than .90, and PNFI is greater than .50. The results of this study’s measurement model assessment showed that χ²/df was 2.162, RMSEA was .055, GFI was .873, TLI was .920, and PNFI was .790. Therefore, the measurement model was deemed reasonably fit.

Convergent validity was assessed based on Hair et al.’s (2010) guideline, which states that convergent validity is established if the loadings for each construct item exceeds 0.60, the value of each construct’s average variance extracted (AVE) is greater than 0.50, and the composite reliability for each construct is equal to or greater than 0.70. Initial assessment of the measurement model resulted in two items under relationship marketing being removed from further analyses due to loadings less than the recommended threshold value of 0.60. After the removal of said items, the results in the online Supplemental Table 1 show that the measurement model satisfied the requirements of convergent validity. Specifically, factor loadings for all the items ranged from 0.626 to 0.887, the AVE value for all constructs were greater than the threshold value of 0.50, and the composite reliability for all constructs were more than 0.70.

Discriminant validity was evaluated based on the criterion by Fornell and Larcker (1981). It was determined based on two criteria: (1) comparison of the square roots of AVE with the variance of the constructs and (2) comparison of AVE values with the maximum shared variance (MSV) of the constructs. According to Fornell and Larcker (1981), discriminant validity is achieved if the values of the variance shared between any two constructs are smaller than the values of the square roots of AVE, and AVE values for all constructs are greater than the values of their respective MSVs. As depicted in the online Supplemental Table 2, all the square roots of AVE (diagonal entries in bold) were greater than the correlation values (off-diagonal entries in italic), and the AVE values for all constructs were greater than their MSV values, affirming that discriminant validity was established.

Common Method Variance

Common method variance (CMV) can be defined as the “variance that is attributable to the measurement method rather than to the constructs the measures represent” (Podsakoff et al., 2003, p. 879). CMV may produce a false internal consistency and intercorrelations that affect the validity of the results (Malhotra et al., 2006). This means that significant results from the relationships between constructs may be due to CMV instead of actual hypothesized effects. In this study, the issue of CMV was assessed as per the suggestion of Malhotra et al. (2006), who stated that CMV is not a problem if all the measurement items modeled as a single factor do not fit. Following this guideline, the single factor analysis revealed that the model did not fit; hence, it was ascertained that CMV was not an issue in this study. Having addressed and satisfied the criteria for validity, reliability, and CMV, the subsequent section presents the results of the structural model and hypotheses testing.

Structural Model Assessment and Hypotheses Testing

The SEM technique was used to assess the structural model of the present study. SEM is commonly used by researchers to examine the overall fit of a structural model and to test its individual causal paths (Hair et al., 2010). The structural model depicted in Figure 1 portrays that the model’s χ²/df was 2.192, GFI was .870, RMSEA was .055, TLI was .918, and PNFI was .793, indicating that the structural model had reasonable fit. Table 2 presents the hypothesized direct path results, which were found to be significant and in support of Hypotheses 1 to 7. Specifically, hotels’ relationship marketing was found to have a significant positive impact on perceived service quality (β = .351, p < .001), corporate image (β = .291, p < .001), and customer satisfaction (β = .239, p < .001). Similarly, hotel guests’ perceived service quality was found to have a significant positive impact on their satisfaction (β = .315, p < .001) and their image of the hotel (β = .215, p < .001). The results also indicate that both corporate image (β = .398, p < .001) and customer satisfaction (β = .239, p < .001) positively influence guests’ loyalty toward the hotel they stayed in during their trip.

OPEN IN VIEWER

Table 2

Result of Path Analysis

Hypothesized Path	Standardized Estimate (β)	Critical Ratio	Significant
Hypothesis 1: Relationship marketing → Perceived service quality	.351	5.814 **	Yes
Hypothesis 2: Relationship marketing → Corporate image	.291	4.687 **	Yes
Hypothesis 3: Perceived service quality → Corporate image	.215	3.444 **	Yes
Hypothesis 4: Corporate image → Customer loyalty	.398	6.427 **	Yes
Hypothesis 5: Relationship marketing → Customer satisfaction	.239	4.199 **	Yes
Hypothesis 6: Perceived service quality → Customer satisfaction	.315	5.194 **	Yes
Hypothesis 7: Customer satisfaction → Customer loyalty	.239	4.390 **	Yes

**

Significant at 99% confidence level.

Multiple Group Analysis

Multiple group analysis (MGA) was used to examine the differences between Asian and Western tourists in relation to the causal paths in the structural model. The MGA analysis was performed based on the guidelines proposed by Jöreskog and Sörbom (1993). It began by splitting the data into two cultural groups, that is, Asian and Western. Next, SEM analysis was performed on the structural model whereby all the causal paths were fixed with equal regression weights for both groups. Then, another SEM analysis was conducted without imposing any equality constraints for the causal paths, such that they were freely estimated. Last, the significance of the moderating effect of culture on all the causal paths was determined based on the chi-square difference test. The moderating effect of culture was deemed to exist when the path coefficient’s critical value was greater than 3.841 at the 95% confidence level.

Further analysis was conducted to determine the differences between the structural models of the Asian and Western groups. The results found that both the Asian group (χ²/df = 1.606, GFI = .833, RMSEA = .055, TLI = .900, and PNFI = .722) and the Western group (χ² = 243.334, df = 182, CFI = .978, IFI = .979, RMSEA = .052) had reasonable fit. As highlighted in the online Supplemental Table 3, the p value of .000 indicated that the chi-square difference between the nonconstraint model and the constraint model was significant, suggesting that the model differed across Asian and Western hotel guests.

To analyze the moderating effect of culture as postulated in Hypotheses 8 to 14, a chi-square difference test was used to evaluate the differences in each of the seven path coefficients. Table 3 exhibits that only two paths were found to be significant. In particular, the p value of .019 reflects the significant difference between Asian and Western tourists in the effect of relationship marketing on perceived service quality, whereby the effect is stronger for Western hotel guests (β = .508). Similarly, the influence of customer satisfaction on customer loyalty was also found to significantly vary across Asian and Western groups (p value = .018), such that this influence is more evident for Western hotel guests than their Asian counterparts.

OPEN IN VIEWER

Table 3

Results of Multiple Group Analysis

Hypothesis and Path Name	Asian (β)	Western (β)	Difference in β	p Value	Significant
Hypothesis 8: RM → PSQ	.178 *	.508 **	−.330	.019	Yes
Hypothesis 9: RM → CI	.268 *	.272 *	−.004	.929	No
Hypothesis 10: PSQ → CI	.266 *	.254 *	.012	.843	No
Hypothesis 11: CI → CL	.348 **	.322 **	.026	.796	No
Hypothesis 12: RM → CS	.128ns	.327 **	−.199	.218	No
Hypothesis 13: PSQ → CS	.279 **	.438 **	−.159	.378	No
Hypothesis 14: CS → CL	.163 *	.372 **	−.209	.018	Yes

Note: RM = Relationship marketing; PSQ = perceived service quality; CS = customer satisfaction; CI = corporate image; CL = customer loyalty; ns = not significant.

*

Significant at 99% confidence level.

**

Significant at 95% confidence level.