Testing moderating effects on the relationships among on-board cruise environment, satisfaction, perceived value and behavioral intentions

Abstract

Purpose

This study aims to investigate both the relationships among the on-board environment, overall satisfaction, perceived value and behavioral intentions and the moderating effects of gender, employment status, group composition and the propensity to stay on board in the context of cruise experience.

Design/methodology/approach

Relationships among constructs were tested on the basis of 417 surveys collected and analyzed with the structural equation modeling approach of partial least squares path modeling. A multi-group analysis was used to test the moderating effects.

Findings

The research findings suggest that on-board environment is a good predictor of behavioral intentions, but that the relationship is strongly mediated by satisfaction and perceived value. Moreover, the multi-group analysis of moderating effects indicated various differences that offer interesting insights for segmenting passengers; these insights have substantial implications for future studies and cruise line companies alike.

Practical implications

This study offers useful insights for managers who want to differentiate their value proposition with ship-centered elements.

Originality/value

This study contributes to the literature by providing a theoretical framework and empirical evidence for analyzing the role of the perceived on-board environment in passenger experience. From a managerial perspective, the moderating effects offer new insights for targeting and customizing the cruise experience value proposition.

Citation

Calza, F., Pagliuca, M., Risitano, M. and Sorrentino, A. (2020), "Testing moderating effects on the relationships among on-board cruise environment, satisfaction, perceived value and behavioral intentions", International Journal of Contemporary Hospitality Management, Vol. 32 No. 2, pp. 934-952. https://doi.org/10.1108/IJCHM-09-2019-0773

1. Introduction

In recent years, the cruise industry has grown rapidly, from a market of 20.9 million passengers in 2012 to a market of 27.2 million passengers in 2018, with more than 30 million expected in 2020 (Cruise Lines International Association, 2018). From 2016 to 2018 alone, 27 new ships were introduced to the cruise market (MedCruise, 2018). One current trend is to invest in giant ships that can accommodate over 6,000 passengers to improve the attractiveness and profitability of each cruise brand (Cruise Lines International Association, 2018). Another trend is to invest in ships customized for specific target markets. For example, recently two new cruise ships have been designed specifically for the Chinese market, with wider internal common spaces and a reduction of external spaces, and a “no frills” ship design has been introduced specifically for veteran cruisers. Common to both trends is greater investment in intangible, atmospheric elements of the on-board cruise experience (e.g. lights, music and odors) in recognition of consumer experience as a top commercial priority. As the ship is a container of passenger experience and provides a series of services (e.g. food and entertainment) (Walls et al., 2011; Yim et al., 2019), managerial decisions regarding investments in particular aspects of a vessel internal design and on-board services become crucial toward offering tourists an optimal cruise experience.

Empirically speaking, elements of the “servicescape” have been found to affect consumers’ response and behavioral intentions (Lockwood and Pyun, 2019). Indeed, elements of the physical environment are particularly relevant in service settings where interaction between consumers and the hosting organization lasts longer and the consumption experience is predominantly hedonic in nature, as happens in the tourism sector and particularly in cruise tourism. The cruise vacation’s mix of benefits, ranging from the tangible (e.g. on-board entertainment and cuisine) to the emotional (e.g. escapism and recharging), make it an exemplary hedonic service experience. Although some academics have highlighted the important role of the physical environment in determining cruisers’ satisfaction and behavioral intentions (Kwortnik, 2008; Lyu et al., 2017; Risitano et al., 2017), this area requires further research. Some authors are still investigating aspects of the cruise ship servicescape (Lyu et al., 2017), while others have concentrated their analysis on specific national targets (e.g. American cruisers) (Li and Kwortnik, 2017), and still others have analyzed elements of the environment in terms of perceived on-board service quality (Chua et al., 2015; Li and Kwortnik, 2017; Petrick, 2004).

The scarcity of studies and the differences in both results and approaches mean that further empirical work needs to be done. To fill this gap, this study examines the relationships among on-board environment, passenger satisfaction, perceived value and behavioral intentions. In particular, it tests the moderating effects of gender, employment status, attitude to staying on board and group composition, thus offering an original contribution to theoretical studies and insights for practitioners. To achieve these aims, the work uses partial least squares path modeling (PLS-PM) on a sample of 417 international cruise passengers disembarking in Italy during their vacation.

This study contributes to the body of knowledge in several areas. By incorporating the moderating effects of gender, employment status, group composition and propensity to stay on board, it generates the opportunity for cruise managers to segment the market and provide a customized ambiance, thereby enhancing the level of cruise satisfaction and stimulating word-of-mouth recommendations and return intentions. Due to the subjectivity of environmental perceptions (Mehrabian and Russell, 1974), analysis of these moderating factors offers the opportunity to seek out similar behaviors within a specific consumer cluster (Akbar, 2013).

2. Literature review and research hypotheses

2.1 The on-board cruise environment

The term “on-board cruise environment” refers to all the ambient, design-related and social factors (e.g. music, lights, temperature and colors) that comprise a ship’s overall atmosphere during a cruise vacation, known as “shipscape” in the cruise tourism literature (Kwortnik, 2008), it is adapted from the original model of the servicescape (Bitner, 1992), which encompasses three latent dimensions: ambiance, aesthetics and social atmosphere. Recent cruise tourism literature has focused on the effects of environmental stimuli on consumer perceptions, but results are still nascent. Specifically analyzing the perceptions of Chinese passengers, Lyu et al. (2017) studied ship environment to validate a novel measurement scale with 26 items. Other studies have analyzed elements affecting the construct. For example, using a choice experiment, Mahadevan and Chang (2017) studied willingness to pay for on-board cruise services, and verified that cruisers will pay extra for perks such as upgraded cabins and leisure facilities. Similarly, Hung (2018) analyzed Chinese cruisers’ behaviors in depth by adopting an exploratory research method (photo-interviewing 20 passengers) to hierarchically gauge their on-board experience. The findings showed that passengers are influenced by both a ship’s tangible and intangible attributes (e.g. convenience, pampering service, level of cruisers’ sentiments, fun and comfort).

As noted, the effects of the on-board cruise environment are still under investigation and empirical research testing the effects of this construct on behavioral consequences is underdeveloped. A first attempt to fill this gap was made by Petrick (2004), who tested the impact of perceived ship quality on perceived value and repurchase intention, and found interesting results that segmented first-time cruisers and repeaters. Afterwards, Chua et al. (2015) evaluated the effects of specific aspects of on-board service quality on passengers’ satisfaction. Moreover, studying an exemplary population of Chinese cruisers, Wu et al. (2018) analyzed the influence of on-board experiential quality on experiential value, satisfaction and behavioral intentions in the context of a local cruise trip around Hong Kong.

While empirical research explores the concept of on-board environment with various measurement scales, most uses the construct in a rational manner by asking passengers to judge the perceived quality of various elements (Chua et al., 2015; Petrick, 2004). The only study that tried to expand this approach to a more experience-centric analysis is a study of the link between cruise brand experience and perceived value (Ahn and Back, 2019), where sensory, affective and behavioral experience were found to be associated with functional and hedonic value. Moreover, most studies were conducted using online surveys, and so did not capture actual perceptions while travelers were still experiencing the cruise trip. This combined with the inconsistency of theoretical models and methods implemented, motivated the present research, which aims to explain the effects of perceived on-board environment on passenger response. In doing so, the present study considers on-board environment to be an important predictor of passenger response (e.g. perceived value and satisfaction) and behavioral intentions (e.g. repeat custom, word-of-mouth recommendation, electronic recommendation). We assume that the on-board environment might be an antecedent of the cognitive responses of satisfaction and perceived value, and by extension behavioral intentions. This is expressed in the following hypotheses:

H1.

The on-board environment has a significant effect on perceived value.

H2.

The on-board environment has a significant effect on overall satisfaction.

2.2 Perceived value and satisfaction

Traditionally, value for customers has been considered from a utilitarian perspective, as the utility derived from a product or service (Tellis and Gaeth, 1990; Yi et al., 2014; Zeithaml, 1988). Some researchers have started to analyze differences and similarities among the concepts of value, quality and satisfaction. Cronin et al. (2000) argued value as a consequence of perceived quality, while Sweeney and Soutar (2001) contended that quality is part of overall value. In contrast, Holbrook and Hirschman (1982) presented a more complex model of perceived value that comprises utilitarian and hedonic dimensions. Their framework shows that value is perceived not only as a trade-off between benefits and sacrifices (rational perspective) but also as a consumption experience that positively or negatively impacts feelings and emotions (experiential perspective). They defined consumer value as an “interactive, relativistic and preferential” experience and proposed eight categories of consumer value (efficiency, excellence, play, aesthetics, esteem, status, ethics and spirituality) related to three main dimensions, namely, extrinsic vs intrinsic (utilitarian vs hedonist); active vs reactive (as the consumer exerts active or passive control over the object); and self-oriented vs other-oriented (when a social dimension of the act of consuming is adopted) (Gallarza and Saura, 2006). Similarly, Sheth et al. (1991) theorized a multidimensional model that frames perceived value as four sub-dimensions, namely, functional, social, emotional and epistemic.

Consequently, consumer value could be considered to reside not in the product/service purchased – or in the brand chosen and/or the object possessed – but in the consumption experience itself (Vargo and Lusch, 2004). This approach implies a more personal evaluation of perceived value, which might depend on situational and contextual factors (Sánchez-Fernández and Iniesta-Bonillo, 2007; Zeithaml, 1988). In fact, the concept of perceived value from an experiential perspective has been broadly used in the service context (e.g. retail, hotel, events and cruise ships), in which intangible factors are especially pertinent. Over the past 10 years, scholars in the cruise tourism field have approached the study of perceived value by analyzing its antecedents and their consequences. Duman and Mattila (2005) verified the effects of novelty, control and hedonics on satisfaction, perceived value, and behavioral intentions among American cruisers and found positive relationships among the latent constructs. In the present work, the construct is considered to be the result of utilitarian and experiential perceived value derived from the perception of the on-board environmental stimuli. This view reflects a concept of value as a one-dimensional construct that depends on certain personal evaluations.

In the general marketing literature, satisfaction has been well-distinguished from the concept of value. While value is perceived during each stage of a consumer journey, satisfaction is measured at the final stage, after a certain consumption experience, as the gap between personal expectations and actual experience (Parasuraman, 1997; Sweeney and Soutar, 2001). Oliver (1996) analyzed the constructs of satisfaction and value in depth and proposed a wider perspective that considers the two concepts differently. He posed a series of compelling questions about the link between satisfaction and value, and considered that satisfaction may be linked to, but conceptually different from, value. In their study of these relationships, Overby and Lee (2006) found that both utilitarian and hedonic value positively affect consumer satisfaction. According to this approach and contrary to some authors who have considered overall satisfaction to be an antecedent of perceived value (Parola et al., 2014; Sanchez et al., 2006), the present study considers satisfaction a consequence of the value of the perceived on-board environment to emphasize the link to value but with a different meaning. Thus, the following hypothesis was developed:

H3.

Perceived value has a significant effect on overall satisfaction.

2.3 Behavioral intentions

Usually, the construct of behavioral intentions has been theorized as loyalty behaviors (i.e. repurchase intentions, word-of-mouth recommendation, willingness to pay a premium) (Wu et al., 2018; Zeithaml et al., 1996). The term “behavioral intention” signifies the intention to enact plans to perform or not perform future behaviors (Ahn and Back, 2019; Warshaw and Davis, 1985). In the tourism literature, behavioral intentions are widely measured by multiple items to capture repurchase intentions, word-of-mouth recommendation, electronic recommendation and loyalty (Wu et al., 2018). Managers wanting to strengthen the efficacy of their marketing must understand the analysis of these behavioral intentions; in the present work, they represent effects of the perceived on-board experience. Indeed, from a marketing perspective, on-board experience generates value if it is related to passengers’ behavioral intentions (Bitner, 1992; Jang and Namkung, 2009). However, the existing literature does not explain the simultaneous effect of the on-board environment on behavioral intentions, mediated by the effects of satisfaction and perceived value. Therefore, the following additional hypotheses are proposed:

H4a.

The on-board environment has a significant effect on behavioral intentions.

H4b.

Perceived value has a significant effect on behavioral intentions.

H4c.

Overall satisfaction has a significant effect on behavioral intentions.

2.4 Moderating effects on the relationships among constructs

When data are derived from a single population, the assumption of homogeneity is often unrealistic and too general, so some moderating effects need to be tested. Although in the marketing literature the effects of consumer profiles are considered important, in the tourism field – particularly in studies of cruiser behaviors – investigation into the effect of moderating variables on passengers’ behavior is limited (Parola et al., 2014; Sanz-Blas et al., 2017c) and mostly focused on external marketing stimuli (e.g. excursion package, information sources and service expertise).

In the present research, some socio-demographic and behavioral variables are taken into account: gender, employment status, group composition, and the propensity “to stay on board” (rather than “to get off” to visit the destination when the ship is in a port of call). The selected moderator variables were chosen for the following reasons. First, in the tourism literature, analyses of socio-demographic variables are limited and have yielded contradictory findings (Calza et al., 2019). Second, socio-demographic variables are generally used to segment the market. Although perceptions of environmental stimuli have the potential to drive behavioral response, they are subjective (Mehrabian and Russell, 1974), so there is a need to test for their moderating effects to find similar behaviors in homogeneous groups of consumers (Akbar, 2013). In doing so, the selected socio-demographic and behavioral variables represent a first attempt in the cruise literature to offer an opportunity for cruise managers to profile the market and implement marketing differentiation strategies. Third, the majority of moderating effects in the cruise literature have been little explored and have generally related to cruiser behaviors at the port of call. In this research, we test whether the effects of these variables (gender, employment status and group composition) are sensitive to ship environment.

2.4.1 The moderating effect of gender.

Gender may be analyzed from two perspectives: gender identity, related to the personality traits, which distinguish male from female (Gould and Weil, 1991) or simply biological sex (male vs female) (Kolyesnikova et al., 2009). Because it is more practical, the majority of studies investigate gender as biological sex (Khan and Rahman, 2016). Identification of gender differences confirms the validity of market segmentation (Kim et al., 2007). Gender has been fully recognized as a variable that influences consumer decision-making (Hansen and Møller Jensen, 2009); however, its impact on perceptions of the physical environment and behaviors has rarely been explored. Still, the few existing studies have shown that women are more sensitive to the physical environment (Chiu et al., 2005) and pay more attention to the appearance of a servicescape than men when engaging in shopping (Wu et al., 2017). On this basis, we suggest that gender influences the relationships among the latent constructs, and formulate the following hypothesis:

H5a.

Gender has a moderating role in the relationships among the constructs.

2.4.2 The moderating effect of employment status.

Employment status refers to the engagement of an individual in paid work. No prior study has directly tested the moderating effect of employment status and yet employment could affect decision-making and consumption experience because it can be closely related to income. In fact, few previous studies have analyzed the role of income in consumer choice (Akbar, 2013; Zeithaml, 1988). In the context of service experiences, some studies have analyzed the impact of income on consumer experience, but with inconsistent results (Grace and O'cass, 2005). Nevertheless, this variable is considered in the market segmentation when cruise managers design their vessels. In the context of cruises, we suppose that the ship environment is perceived differently by employed and unemployed people and that this could affect their satisfaction and behavioral intentions. Thus, we formulate the following hypothesis:

H5b.

Employment status has a moderating role in the relationships among the constructs.

2.4.3 The moderating effect of group composition.

Group composition is not normally considered in academic work (San Martìn, 2005) and yet it can have an important influence on individual perceptions. In the tourism field, Campo-Martínez et al. (2010) observed differences between tourists who travel alone and those who travel with friends in the formation of their behavioral intentions to revisit a destination. We, therefore, propose that group composition (small group vs. large group) influences the individual’s perceptions of on-board environment and its consequences in terms of perceived value, satisfaction and behavioral intentions. Hence, the following hypothesis:

H5c.

Group composition has a moderating role in the relationships among the constructs.

2.4.4 The moderating effect of the propensity to stay on board.

Visiting the port of call may be motivated both by push factors, determined by cruisers’ internal psychological wish to do so, and pull factors, consisting of port-of-call sightseeing, which might attract visitors’ interest (Sanz-Blas et al., 2017a, 2017b). The propensity to stay on board refers to the option to remain on board rather than to go ashore to visit the destination. This variable is particularly relevant for cruise companies because they should consider people who stay on board as an opportunity for encouraging shopping or for increasing passenger exposure to (and satisfaction with) environmental stimuli (e.g. music, colors). We presume that people who prefer staying on board during a port of call stop appreciate the on-board environment and prefer to enjoy the ship, especially as there are fewer other passengers. Consequently, cruisers who feel engaged in the on-board environment will be less willing to get off to visit the destination. The following hypothesis is suggested:

H5d.

Propensity “to stay on board” rather than “to get off” has a moderating role in the relationships among the constructs.

Figure 1 illustrates the hypothesized model, the theorized causal relationships among the constructs, and the effects of the moderating variables.

3. Research method

3.1 Sample and data collection

Data for this research were collected through a survey conducted with passengers disembarking from 12 cruise ships (Costa Fascinosa, Freedom OTS, Jewel OTS, MSC Armonia, MSC Poesia, MSC Symphony, Majestic Princess, Norwegian Epic, Norwegian Spirit, Royal Princess, Sovereign and TUI Discovery) between April and June 2017 at the Port of Naples, which is the third largest cruise port in Italy and the seventh largest in the Mediterranean area (Naples had 927,458 passengers in 2017 according to Italian Cruise Watch Annual Report, 2018). The 12 cruise ships are representative of the fleets of four worldwide shipping companies (Carnival, Royal Caribbean, Norwegian Cruise Line and MSC Cruise). Moreover, to keep the sample uniform, respondents had been on the same category of cruise in terms of itinerary (west Mediterranean), length of trip (seven days), ship size (mega ships) and target market (mainstream), as well as ship environment (casual atmosphere). To be sure of capturing impressions from those who had actually experienced the on-board environment, only passengers who were in the middle of or finishing their cruise experience were interviewed.

A trained group of university students distributed a self-administered structured questionnaire with multiple-item scales. The questionnaire consisted of five sections, namely, on-board environment (17 items), perceived value (4 items), overall satisfaction (3 items), behavioral intentions (4 items) and demographic background characteristics. The items on the questionnaire were rated on a seven-point Likert scale, ranging from 1 (“strongly disagree”) to 7 (“strongly agree”).

A pilot survey was carried out for the purposes of weeding out errors and ensuring that the questions were easy to understand. After the questionnaire was finalized in English, it was translated into French, Spanish and Italian. Non-probability judgmental sampling (due to the unavailability of a sampling frame) was used and 450 surveys were collected. In total, 33 questionnaires were rejected for being incomplete, leaving a total of 417. The sample was almost equally split between males (44 per cent) and females (56 per cent). Most of the respondents (73 per cent) were over 34 years old and lived in Europe (56 per cent). Additionally, 69 per cent of the respondents had a high level of education. Two-thirds (67 per cent) were married and 62 per cent were employed. In terms of annual household income, 51 per cent earned more than 50,000 euros. A third of the respondents were on their first cruise (32 per cent), while the majority were repeaters (67 per cent). Most of the respondents were traveling alone or with a partner, and most had decided to take this cruise trip to experience the on-board environment (66 per cent).

3.2 The measures

Once the literature was reviewed and subjected to an elicitation study, relevant constructs were mainly operationalized by adapting a multi-item scale. The measurement items for on-board environment were identified from studies in the literature (Kwortnik, 2008; Kim and Moon, 2009; Siu et al., 2012). The scales developed by Duman and Mattila (2005) were adapted to measure perceived value, while the constructs of satisfaction and behavioral intentions were measured using items adopted from the literature (Chua et al., 2015; Kwortnik, 2008). The indicators, constructs and Cronbach’s alpha values (Chin, 1998; Guilford, 1954) are displayed in Table I. The Cronbach’s alpha values provide a reliable indicator of internal consistency. The reliability statistics for all constructs range from 0.794 to 0.943. This exceeds the 0.7 rule-of-thumb (Nunnally and Bernstein, 1994) and confirms the scales’ reliability adequate internal consistency.

3.3 Data analysis

Constructs such as on-board environment, perceived value, overall satisfaction and behavioral intentions are not easy to measure directly. Moreover, such latent constructs are linked to other latent constructs. To test the research hypotheses, the proposed research model was estimated by PLS-path modeling (Lohmöller, 1989; Wold, 1982), a soft modeling approach to the analysis of the cause-effect relationships among latent constructs. Because it is a distribution-free data analysis approach, it does not require distributional assumptions and it is a more flexible technique for handling a small sample sizes, and different measurement scales.

PLS-PM is a two-step iterative algorithm that separately estimates each latent construct of the measurement model and then, in a second step, estimates the path model coefficients through ordinary least squares (OLS) regression (Chin, 1998). As PLS-PM has no distributional assumptions, the significance of the parameters is estimated through resampling and cross-validation methods. A bootstrapping re-sampling method with 5,000 re-samples was used to evaluate the model (Chatelin et al., 2002; Chin, 2010) and a Q² index was used to assess the cross-validation redundancy (Hair et al., 2017).

Once the main effects were shown to be supported, the moderating effect of background factors (gender, employment status, group composition and propensity to stay on board) was tested[1]. In the interest of confirming whether these socio-demographic factors have a moderating effect on model relationships, the authors performed a multi-group structural equation modeling analysis (Sörbom, 1974). Using a permutation procedure (Chin and Dibbern, 2010a; Edgington, 1987), the authors further yielded a subset of possible permutations in the data among the groups sampled. As a distribution-free test, this method eschews any need for parametric assumptions and is considered a good non-parametric alternative to the standard t-test.

4. Results

The global criterion of goodness of fit – i.e. the goodness of fit index (GFI) (Tenenhaus et al., 2005) – was equal to 0.551. As a rule of thumb, a GFI value equal to or higher than 0.36 (Wetzels et al., 2009) indicates the PLS model to be valid. The measurement model involved four constructs: on-board environment, perceived value, overall satisfaction and behavioral intentions. To assess the model, convergent validity and discriminant validity needed to be established. Convergent validity was assessed in two ways: first, by examining the indicator loadings; second, by examining the average variance extracted (AVE).

As shown in Table II, all factor loadings were higher than 0.70 (except 2_16 for on-board environment), and the AVE values were higher than 0.50 for each construct (Bagozzi and Yi, 1988; Chin, 2010). Thus, convergent validity was confirmed (Fornell and Larcker, 1981). Discriminant validity (Table III) was tested by comparing the AVE values with the squared correlations between pairs of constructs. As shown in Table III, all the squared correlations between pairs of constructs were less than the AVE for each construct, suggesting that discriminant validity was fulfilled (Fornell and Larcker, 1981). The overall fit of the structural model shows a good predictability, as R² of the dependent variables ranges from 0.358 to 0.649, and a predictive relevance (Chin, 1998), as Q² ranges from 0.236 to 0.506. The findings indicate that the model adequately predicts behavioral intention to repeat the cruise experience (Table IV).

In relation to the structural model, H1-H4 were tested (Table V). Analysis of H1 showed that the direct effect of on-board environment on perceived value was positive (path 0.598) and significant (p < 0.001), thus supporting H1. H2 determined the causal effects of on-board environment on overall satisfaction. There were positive (path coefficient of 0.336) and significant (p < 0.001) direct effects, accounting for 35 per cent of the variance (R² = 0.358) in support of H2. Perceived value showed an indirect positive effect (equal to 0.188) of on-board environment on behavioral intentions. H3 was tested to investigate the causal effect of perceived value on overall satisfaction. The findings revealed positive (path coefficient of 0.558) and significant (p < 0.001) direct effects, accounting for 64 per cent of the variance (R² = 0.649) in support of H3. Furthermore, on-board environment had an indirect effect on overall satisfaction (path = 0.334) via perceived value.

Finally, H4a-c proposed that on-board environment, perceived value, and overall satisfaction were predictors of behavioral intentions. The estimated structural paths were found to be positive and significant (0.088, 0.339 and 0.337, respectively); hence, H4a, H4b and H4c were supported. The R² for behavioral intentions was 0.484, which suggests that 48 per cent of the variance in behavioral intentions could be explained by its predictors (i.e. on-board environment, perceived value and overall satisfaction). The contributions of on-board environment, perceived value and overall satisfaction to R² were found to be 9 per cent, 45 per cent and 46 per cent, respectively. In addition, on-board environment and perceived value affected behavioral intentions through overall satisfaction (their indirect effects were equal to 0.428 and 0.188, respectively). The total effect (path = 0.516) of on-board environment on behavioral intentions was almost all via its indirect effects (83 per cent). We can, therefore, conclude that overall satisfaction had a mediating role in the relationships between on-board environment and behavioral intentions, and between perceived value and behavioral intentions.

4.1 The moderating effect of the socio-demographic factors

For multi-group comparisons with PLS, we used a distribution-free permutation procedure to construct a subset of all possible data permutations between the sample groups. This procedure is also useful for testing multi-group invariance measurement, an important prerequisite for group comparisons (Chin and Dibbern, 2010b). A permutation test (randomization test) is a good non-parametric alternative to the well-known t-test that fits well with PLS path modeling. It uses the S statistic (the absolute value of the difference on a parameter between two groups of observations) to compare all parameters, both standardized loadings and path coefficients. It also allows for an evaluation of measurement invariance; the loadings and weights of the constructs’ measurement models must not differ significantly within the model (Chin, 2000). The measurement invariance is necessary to ensure that the paths compared in the test are comparable in terms of the causal relationships that they represent. The S statistic is computed on the original samples and samples are permuted. The original S is compared with each permuted S to obtain a p-value:

p-value=1Nperm+1∑k=1NpermlSobs<Sperm

where S_obs is the observed difference in the generic i-th parameter between a priori formed groups for the j-th moderating variable, V^j and S_perm is the difference in the generic i-th parameter between randomly assembled groups, regardless of a priori distinctions. The hypotheses to be tested are:

• parameters (loadings and path coefficients) are not significantly different; and

• parameters are significantly different.

In our analysis, at the 5 per cent level, no difference in loadings between any subsample was significant. Measurement invariance is, therefore, established and we can proceed to interpretation of the results of the group comparisons.

4.2 Results of the structural difference analyses

The findings show that gender has a moderating effect in two of the six causal relationships: between on-board environment and overall satisfaction, and between perceived value and overall satisfaction. In particular, the difference between the path between on-board environment and overall satisfaction favored females, while the difference between the path between perceived value and overall satisfaction favored males.

A comparison of path coefficients for employed and unemployed cruisers allowed the authors to test the moderating variable of employment status. Employment status had a significant moderating effect on the relationships between on-board environment and overall satisfaction, between perceived value and overall satisfaction and between on-board environment and behavioral intentions. The influence of on-board environment on overall satisfaction was stronger for unemployed cruisers, while the impact of on-board environment on behavioral intentions was stronger for employed cruisers. Furthermore, overall satisfaction was found to be more strongly influenced by perceived value among employed cruisers than among unemployed cruisers. Regarding the moderating effect exerted by the variable “group composition,” we found that on-board environment had a stronger impact on perceived value for cruisers who traveled alone or with a partner than for cruisers who traveled in groups. Finally, the moderating variable of on-board environment was tested via comparison of path coefficients for two groups: cruisers who wanted “to get off” and cruisers who wanted “to stay on board.” The findings revealed that for cruisers who wanted to get off, overall satisfaction had a significant impact on behavioral intentions.

5. Discussion and conclusions

The framework proposed novel constructs that influence behavioral intentions in the cruise industry and the study analyzed the impact of the on-board environment on passengers’ behavioral intentions via perceived value and satisfaction. Moreover, important segmentation variables were tested in the multi-group analysis, namely, gender, employment status, group composition and propensity to stay on board. The findings have practical implications for cruise line companies.

First, this study reveals the important influence of the on-board environment on perceived value, satisfaction and willingness to engage in word-of-mouth recommendation and repeat a vacation. Second, the results show that the on-board physical environment is globally, one-dimensionally perceived as a single holistic experience of the ship’s overall atmosphere. This is surprising and interesting, because in earlier studies the so-called “shipscape” has been divided into three latent dimensions. However, the unidimensionality identified in this study needs to be confirmed. This result may derive from the fact that people who are still in the “cruise vacation mood” do not separate out the elements of the on-board cruise environment; instead, they tend to express their evaluations as an overall perception of a unique pleasant/unpleasant atmosphere.

Third, on-board environmental stimuli are predictors of behavioral intentions, and this relationship is strongly mediated by satisfaction and perceived value. Multi-group analysis indicated differences between female and male tourists, unemployed and employed tourists, group travelers and those traveling alone or as a couple, and those who prefer “to stay on board” rather than “to get off” at a port of call. That is, women are more sensitive to the on-board environment than men and unemployed people perceive a higher value from the on-board environment than those who are employed. Furthermore, on-board environment is also important to perceived value for cruisers traveling alone or with a partner rather than as part of a larger group. Finally, people who prefer to stay on the ship (who are presumably more sensitive to the on-board environment) rather than look around a port are more willing to recommend and/or repeat the cruise experience. The findings on these moderating effects are a significant contribution to previous studies and offer insights for cruise managers seeking to offer an optimized and customized on-board environment.

5.1 Theoretical implications

The present study enriches knowledge of tourism and hospitality by examining the link among on-board cruise environment, satisfaction, perceived value and behavioral intentions. The significance of this study is also demonstrated by the moderating effects of gender, employment status, group composition and the propensity to stay on board, which represent small but smart data for managers to customize services, enhance cruise satisfaction, and improve word-of-mouth and customer-return intentions. The findings advance academic knowledge in three ways. First, they support several studies reporting that on-board environmental stimuli affect passengers’ cognitive responses and behavioral intentions (Chua et al., 2015; Hosany and Witham, 2010; Kwortnik, 2008) but disconfirm that the ship environment is a multi-dimensional construct (Kwortnik, 2008). Although this result needs to be confirmed, it may derive from the experience-centric approach used to construct the measurement scale, where each element of the on-board environment is related to the pleasure/displeasure of experiencing it (previous studies used an impersonal list of elements to describe the quality of a ship’s environment). Perhaps this study’s approach rendered people unable to distinguish the latent components but instead simply perceive an overall atmosphere (e.g. “delightful”).

Second, while most previous studies used online surveys, the present research used a face-to-face survey to capture people’s emotions when they had disembarked, and this overcomes any possible methodological bias from surveying people potentially long after they had experienced the cruise. In fact, as pointed out in the discussion section, people who are still immersed in the mood of the vacation might be less inclined to separate out the various elements of the on-board environment. Third, this study investigates for the first time some interesting moderating effects that have important managerial implications. As the cruise is a collective trip experience, it is important to find a link between economic efficiency for cruise companies and a memorable experience for customers. In this research, women, unemployed persons, single people and couples were found to be more sensitive to the on-board environment, as were those who prefer to stay on board rather than visit the destination. Despite needing confirmation, these results offer new insights for marketing researchers and advance knowledge on-board environmental impact on perceived value, satisfaction and behavioral intentions.

5.2 Practical implications

In a competitive scenario oriented toward the “mega ship,” able to accommodate over 6,000 passengers, the need to understand and monitor customers’ needs and wants becomes crucial for cruise companies. This study emphasizes the importance of a strong orientation to the customer experience in the delivery of environmental stimuli. In particular, to engage people during their consumption experience, the cruise company should give more attention to personal engagement and go beyond the provision of mainstream and standardized services (i.e. entertainment, food, excursions and shopping) to offer some unexpected and tailored experiences. The positive relationship between on-board environment and behavioral intentions suggests how this element of the offering might become an important point of differentiation between cruise brands. Cruise companies are already heavily investing in new value propositions that are increasingly customer-oriented and fit with new travel behaviors (e.g. short cruise vacations, four-season availability), but this might not be enough. Cruise managers should attend to the more environmental elements of the cruise experience to demonstrate that passengers are not numbers but people (e.g. capture insights from socio-demographic data to surprise passengers on special dates or to greet each passenger at the end of the trip).

Being a mass service, one possible way to combine economic efficiency and customer value could be the implementation of a mass customization strategy. The trend is for cruise managers to deliver customized services to their wide but hyper-segmented customer base. A good example is the Mediterranean Shipping Cruise Group (MSC), the fourth largest cruise group in the world, which is investing in this direction. In fact, while the company continues to build new mega ships, the marketing managers are implementing some tailored services along the customer journey (before, during and after the cruise). With the use of 4.0 technologies, for example, they have implemented a virtual assistant in the premium cabins, and are encouraging the use of virtual reality for customer immersion in their ships before purchase. The challenge is to surprise each passenger (or at least groups of them) among the thousands of customers.

This study offers some important market insights regarding possible segmentation groups (gender, employment status, group composition and propensity to stay on board). This is because women, single people and couples are more sensitive to the on-board environment, cruise managers should offer customized services for these target groups, as well as for those who decide to stay on board rather than visit the destination. As women are more sensitive to their environment, new vessels could provide promenades with lights, music, colors, shops, bars and common spaces related to female preferences; for single passengers, ships could provide new single “studio” cabins, as well as family cabins for larger groups; for people who prefer to enjoy the ship during a stop at a port, managers could offer free access to certain exclusive ship areas (e.g. an exclusive solarium) or a free massage. Thus, by manipulating these small elements of the on-board environment, they may achieve important results in terms of cognitive, emotional and behavioral reactions, which are the main driving forces behind repeat experiences and recommendations (Ahn and Back, 2019; Rosenbaum and Massiah, 2011; Zomerdijk and Voss, 2010).

5.3 Limitations and future research

Although the results provide important contributions for marketing studies and cruise management, there are some limitations. First, convenience sampling was used, so the response/refusal rate could not be determined. Neither could response bias be measured. The results, therefore, cannot necessarily be generalized. Nonetheless, despite being tested with a convenience sample, the model is essentially an exploratory one and provides a good basis for further research. Second, because the shipscape construct has hitherto only been analyzed as a composite of three second-order factors, its unidimensionality could be considered a limitation, so the authors chose to call the construct “on-board environment.” Future studies could try to identify those elements (e.g. design, functionality and on-board social life) that have the largest influence on cognitive and behavioral responses and also the elements that “influence” the responses could be used to segment passengers on the basis of the environmental factors and type of ship (e.g. mega, medium or small ship). Last but not least, future studies could incorporate neuromarketing techniques, which are able to capture real emotions during a trip, and therefore, confirm or disconfirm what survey respondents say. Moreover, it could be interesting to analyze the on-board experience by comparing different samples of ships, with particular reference to those that implement the use of smart technologies to improve the passenger experience. As cruising remains a popular part of the travel industry, and consumers are looking for ever-more customized experiences, this is an important area for future study.

Figures

Figure 1.

The theoretical model

Table I.

Theoretical constructs, indicators and composite reliability indices

Theoretical constructs	Variables	Indicators	Cronbach’s alpha	D.G. rho (PCA)
On-board environment	2_1	The overall lighting in this ship environment is appropriate	0.943	0.950
	2_2	The temperature on this ship is comfortable
	2_3	The scents on this ship are pleasant
	2_4	The music makes this ship a more enjoyable place
	2_5	Overall, the environment is kept clean
	2_6	Overall, the atmosphere is elegant and cheerful
	2_7	The cruise motion is pretty calm during navigation
	2_8	This ship is decorated in an attractive fashion
	2_9	The interior décor of the cabins and of the common spaces is attractive
	2_10	The use of color in the décor scheme adds excitement to this ship environment
	2_11	The furnishings on this ship are modern and functional
	2_12	It is easy to walk around this ship and find what you are looking for
	2_13	The seats in the cabins and the common spaces are comfortable
	2_16	The crew is available and smiling anytime
	2_17	The crew is skilled and have a good orientation to problem solving
Perceived value	7_2	The value obtained from social interaction is high	0.818	0.884
	7_3	The value obtained from the aesthetic environment is high
	7_4	The value obtained from the on-board experience is higher than the onshore visits
	7_5	Going on the cruise trip was worth the time and money
Overall Satisfaction	7_1	Overall, the value obtained from this cruise experience is high	0.859	0.914
	7_6	Overall, I’m happy with the price paid for this cruise experience
	7_7	Overall, I am satisfied with the cruise trip
Behavioral intentions	8_3	I would like to repeat a cruise experience	0.794	0.867
	8_4	I would like to repeat a cruise experience with the same ship
	8_5	I’ll recommend the cruise line to family and friends
	8_6	I will write and upload a positive comment about this cruise trip on social networking sites (Facebook, TripAdvisor, etc.)

Source: Authors’ elaborations

Table II.

Factor loadings and convergent validity measures

Latent variable	Manifest variables	Standardized loadings	Loadings	Critical ratio (CR)	Lower bound (95%)	Upper bound (95%)
On-board environment	2_1	0.719	0.831	17.596	0.629	0.792
	2_2	0.720	0.901	20.331	0.637	0.783
	2_3	0.746	0.918	26.353	0.682	0.796
	2_4	0.722	0.907	23.376	0.659	0.776
	2_5	0.788	0.841	26.282	0.720	0.842
	2_6	0.822	0.895	36.091	0.771	0.862
	2_7	0.737	0.857	21.553	0.657	0.797
	2_8	0.826	0.973	39.544	0.786	0.868
	2_9	0.808	0.971	32.793	0.753	0.850
	2_10	0.775	0.962	29.023	0.722	0.823
	2_11	0.801	0.967	29.358	0.742	0.851
	2_12	0.746	0.901	25.151	0.676	0.796
	2_13	0.750	0.852	23.977	0.680	0.808
	2_16	0.613	0.660	10.143	0.486	0.715
	2_17	0.598	0.665	9.757	0.469	0.705
Perceived value	7_2	0.860	1.137	49.768	0.826	0.892
	7_3	0.863	1.070	48.869	0.821	0.896
	7_4	0.768	1.216	20.437	0.674	0.830
	7_5	0.751	0.864	24.386	0.685	0.804
Overall satisfaction	7_1	0.859	1.007	37.785	0.810	0.898
	7_6	0.880	1.067	45.284	0.839	0.916
	7_7	0.910	1.070	60.880	0.876	0.936
Behavioral intentions	8_3	0.785	1.161	23.407	0.710	0.844
	8_4	0.822	1.312	33.889	0.767	0.867
	8_5	0.855	1.281	39.631	0.813	0.897
	8_6	0.706	1.292	17.698	0.607	0.771

Table III.

Discriminant validity

Theoretical constructs	On-board environment	Perceived value	Overall satisfaction	Behavioral intentions
Fornell–Larcker criterion
On-board environment	0.559
Perceived value	0.358	0.660
Overall satisfaction	0.449	0.576	0.780
Behavioural intentions	0.266	0.418	0.426	0.631
Heterotrait–Monotrait ratio of correlations
On-board environment
Perceived value	0.678
Overall satisfaction	0.745	0.913
Behavioral intentions	0.583	0.797	0.785

Notes:

The italic values along the diagonal are AVE values. Squared correlations are below the diagonal

Source: Authors’ elaborations

Table IV.

Structural model assessment

Endogenous latent variable	R²	Adjusted R²	Redundancy Q2
Perceived value	0.358	0.358	0.236
Overall satisfaction	0.649	0.648	0.506
Behavioral intentions	0.484	0.482	0.305

Source: Authors’ elaborations

Table V.

Path coefficients and hypotheses testing

Hypothesis	From	To	Standardized path coefficients	Status
H1	On-board environment	Perceived value	0.598***	Supported
H2	On-board environment	Overall satisfaction	0.336***	Supported
H3	Perceived value	Overall satisfaction	0.558***	Supported
H4a	On-board environment	Behavioral intentions	0.088*	Supported
H4b	Perceived value	Behavioral intentions	0.339***	Supported
H4c	Overall satisfaction	Behavioral intentions	0.337***	Supported

Notes:

***p-value < 0.01;

**p-value < 0.05;

*p-value < 0.10

Source: Authors’ elaboration

Table VI.

Results of the multi-group analysis

	Gender			Employment status			Stay on board			Group composition
	G1	G2	|Diff|	G1	G2	|Diff|	G1	G2	|Diff|	G1	G2	|Diff|
OE → OS	0.205	0.449	0.244**	0.263	0.471	0.209**	0.378	0.272	0.106	0.297	0.390	0.092
PV → OS	0.659	0.467	0.192*	0.624	0.435	0.189*	0.559	0.582	0.023	0.618	0.473	0.145
OE → BI	0.131	0.044	0.087	0.172	−0.110	0.283**	0.051	0.087	0.036	0.128	0.067	0.062
PV → BI	0.288	0.355	0.067	0.241	0.499	0.258	0.551	0.136	0.416**	0.275	0.391	0.115
OS → BI	0.336	0.374	0.038	0.346	0.394	0.048	0.206	0.447	0.241	0.353	0.340	0.014
OE → PV	0.546	0.637	0.091	0.587	0.632	0.045	0.561	0.563	0.002	0.657	0.501	0.156**

Notes:

***p-value < 0.01;

**p-value < 0.05;

*p-value < 0.10. Gender: G1 = Male, G2 = Female; Professional status: G1 = Employed, G2 = Unemployed; Decision to stay on board: G1 = Get off, G2 = Stay on board; Group composition: G1 = Small group, G2 = Large group

Source: Authors’ elaboration

Note

1. Note that additional socio-demographic characteristics were tested as control variables (i.e. education, age, first-time and second-time passengers), but because they had no significant effects, for the purposes of brevity and clarity, the results are not reported here.