Traveling consumption brings an indulgent experience and represents extraordinary life quality. Tourists are induced to treasure new shopping goods and psychologically ready to buy luxury goods as souvenirs for rewarding travel experiences. Prestigious tourist consumption goods are unusual, authentic with psychological, symbolic and emotional values (Kapferer & Bastien, 2009; Tynan, McKechnie, & Chhuon, 2010). Tourists consume travel service and high-end brands to manifest their own social status (Strizhakova, Coulter, & Price, 2008). From the consumption perspective, luxury goods not only carry tangible meanings, but more importantly, also represent intangible meanings by which tourists satisfy both symbolic and material needs. Vigneron and Johnson (2004) stress that behavior toward luxury consumption is much related to self-expression and sociability. Luxury consumption during traveling carries some personal values, apart from the basic functional values in consumption. The luxury brand value consists of hedonic elements associated with experiential, emotional attachment and relationship.

In this study, tourist-consumption based motivation is grounded on self-determination theory (SDT) which focuses on individual's intrinsic psychological disposition toward personality integration as well as self-motivation (Ryan & Deci, 2000). As stated in motivation theories, SDT addresses that humans are active organisms striving for psychological development and growth (Sheldon, Ryan, Deci, & Kasser, 2004). Regarding luxury goods consumption behavior, the goals of humans include self-acceptance (growth), materialistic success (money and luxury), and personal appearance and image (Kasser & Ryan, 1993, 1996). This segment of tourists tends to make use of exclusive luxury goods to satisfy psychological needs and convey private meaning. The following sections will examine in detail the related motivation-based consumption constructs and affective response within tourism innovation contexts and discuss how they often create significant managerial implications for traveling consumption.

The first construct is social esteem. This is a prominent theorized element in the perspectives of self-concept and motivation (Hogg, Cox, & Keeling, 2000). It is generally known that nature of mankind strives to boost social esteem (Crocker & Nuer, 2004). The second construct is desire for experiential travel. From the perspective of tourism innovation, desire for possessions is the synonym of desire for experiential travel. Tourist consumption outside home country creates favorable and enjoyable travel experience. As originated from interpersonal insecurity, desire for material possessions during travel enhances self-appeal (Jiang, Zhang, Ke, Hawk, & Qiu, 2015). Materialism is related to social esteem (Jiang et al., 2015) and desire for self-enhancement (Kilbourne & LaForge, 2010). The third one is need for uniqueness. The purchase of special tourist product or service is a way to develop distinctiveness of an individual as owning unique items is an important means of defining one's sense of self (Snyder, 1992). Exclusive possessions differentiate tourists and express social image without creating an extreme social reaction which could result in deviation (Ruvio, 2008).

Truong and McColl (2011) consider the pursuit of social esteem as one of the important motivational dimensions that will reward tourists emotionally. Social esteem is also a basic reference point for tourists to orient their behavior, as long discussed in the marketing theories of self-other psychological meshing (Ferraro, Shiv, & Bettman, 2005). It can well be understood as a consumer's whole being of liking, respect and acceptance feelings to individual own “self” (Rosenberg, 1965). By nature, individuals strive to enhance and reinforce their own social esteem owing to human ego. Acquiring goods and service improve tourists’ social esteem by inflating their ego (Pyszczynski, Greenberg, Solomon, Arndt, & Schimel, 2004). Expensive tourist product or treasurable travel experience as self-gifts can satisfy the need for social esteem. The need to improve social esteem offers psychological rewards to tourists who engage in status consumption. These ego-enhancing benefits render the consumption of status-oriented goods as an attractive affirmational option. Eastman, Iyer, and Thomas (2013) conclude that status consumption is a motivational process that enhances an individual's standing in a social hierarchy by consuming status goods with symbolic meanings. This study also emphasizes status consumption as an independent difference variable that concerns a social prestige value conferring onto corresponding consumers. Status consumption is a behavior-related need for status at whatever levels of income or social class.

Desire for experiential travel is a synonym of materialism in tourism innovation context. As a personality trait, desire for possessions combines acquisitiveness, non-generosity and envy (Belk, 1985; Richins & Dawson, 1992). These studies conceptualize materialistic possession as a centrally held personal belief in human life, and it consists of three aspects – acquisition centrality, success and happiness. Acquisition centrality refers to a psychological “extant” that a consumer succeeds to hold with respect to materialistic ownership, thereby functioning as a life goal. Such success is based on how far a person can measure his/her own material possessions. Happiness refers to the respective psychological reward that a person can gain through the ownership or acquisition of the desired possessions. In such course, happiness and better well-being exist. Possessions of higher involvement products, such as clothing and jewellery, are found to be the source of happiness because publicly displayed products express higher levels of symbolic meanings and provide deeper impressions to significant others (Lertwannawit & Mandhachitara, 2012). Highly materialistic tourists have a preference to spend more time and energy on consumption during travel and spending to impress others (Fitzmaurice & Comegys, 2006). Prior studies have reported that the desire for possessions has direct effect on status consumption (Kim, Ko, Xu, & Han, 2012). Tourists seeking for treasurable traveling goods and substantial tourist experience do not only demonstrate material ownership and superiority, but also express social status (Amaldoss & Jain, 2005).

Snyder and Fromkin (1977) set forth its seminal discussions on desire for uniqueness, which manifests itself in the courses of material pursuits and differentiated self-image (Simonson & Nowlis, 2000). The need for uniqueness is where a person shows the need to pursue products and service to express differentness and achievement. It can also be embodied by acquisition, utilization, and disposition of goods during travel (Tian, Bearden, & Hunter, 2001). The study comes up with three related dimensions: (1) avoidance of similarity, (2) creative choice for counter-conformity, and (3) unpopular choice for counter-conformity. The latter refers to a situation in which tourists use deviant products to differentiate themselves. Yet, some studies show that unpopular choice for counter-conformity is a weak measurement for the need for uniqueness owing to its low factor loadings (less than 0.50) (Bian & Forsythe, 2012). Therefore, only the first two dimensions are considered in the present study.

Similarity avoidance is a synonym of avoidance of rituality within the context of tourism innovation. It describes a tourist preference that they try to avoid tourist products and services that are seemingly popular and common. Tourists tend to discontinue these popular travel consumption and activities to avoid replication or “being the same” to others. Gierl and Huettl (2010) also discuss the effect of scarcity consumption leading to the suitability of product for conspicuous consumption. When it involves exclusive luxury goods that enhance status, the value of product scarcity increases, because of limited supply.

For creative choice-oriented tourists, even if they want to express self-identity by consuming new and exclusive tourist products, they might still want to follow social norms (Knight & Kim, 2007; Tian et al., 2001). In order to present a distinctive image and at the same time, be accepted by others, traveling with a unique style and in a limited opportunity is the best for offering distinguishing attributes. Innovative traveling offering unusual attributes, such as novel experience and exclusivity, is appealing to the tourists who demand this kind of traveling preference.

In considering tourist consumption behavior, the antecedents of status consumption are proposed to be social esteem, desire for experiential travel, avoidance of rituality and creative choice counter-conformity. The value of luxury goods involves personal preferences that take into consideration personal characteristics, such as hedonism and identity (Wiedmann, Hennigs, & Siebels, 2009). These personal consumption preferences are more evident in tourists who evaluate tourist products and travel experience with individual-based standards, and derive self-directed pleasure. The consumption of luxury goods facilitates the affective response of tourists as personal reward and fulfillment (Shukla, 2008; Tsai, 2005). Status-conscious tourists are affected by symbolic meanings and feelings evoked by a brand and make affective judgments (Agarwala & Malhotra, 2005; Shukla, 2008). These psychological benefits indicate one's self-gratification and satisfaction obtained from a certain product and the related consumption experience during travel. Particularly for tourist travel experience outside home country, the satisfaction of new place and activity treasuring is significantly related to affective behavioral responses, especially in terms of expenses.

Semi-random sampling was applied in the recruitment of every tenth tourist consumer who shopped in retail stores or carried shopping bags in famous buying attractions in Hong Kong. All of the respondents were screened for only those who indicated tourism innovation as the main reason in their trip and were more willing to shop, no more than three days before the date of the survey. The respondents were invited for a self-administered survey questionnaire. The respondents were so located in different places and times in order to maximize external validity against geographic and temporal bias. A total of 316 respondents completed the questionnaires. After data cleaning and screening, 21 cases (respondents) were deleted because of missing data or inconsistent responses, which left 295 samples for study. The data were examined and modeled by confirmatory factor analysis, and structural equation modeling using SPSS AMOS v.21.

The survey measurement accessed ratings in a 7-point Likert-scale, which ranked a “strongly agree” in scale (7) and “strongly disagree” in (1). Considering the time taken to complete the questionnaire, we select reliable items with higher factor loadings. Social esteem was based on the Rosenberg Self-esteem Scale (Rosenberg, 1965). This study borrowed 5 items with loadings exceeding 0.70. Desire for experiential travel was measured by using Materialism scale from Richins and Dawson (1992) with some modifications to meet the research purpose. For tourist consumption, two items from each dimension (acquisition centrality, success and happiness) with high factor loadings over 0.70 were adopted (Richins & Dawson, 1992). Avoidance of rituality and creative choice for counter-conformity were measured by adjusted constructs initially in Tian et al. (2001). Two items with factor loadings higher than 0.70 were adopted (Tian et al., 2001). Status consumption measure was based on the scales in Eastman, Goldsmith and Flynn (1999) with some amendment. This study adopts five items from their scale showing strong reliability (Lertwannawit & Mandhachitara, 2012). Affective response was measured by the 5-item consumer perceived emotional value, as referred to Sweeney and Soutar (2001). Altogether, there are 25 question items developed in the measurement instrument (Table 1).

Cronbach's alpha reliability tests were conducted to ensure measurement scale reliability, and eliminated those items with scores lower than 0.50 corrected item-to-total correlation. The overall alpha score of item set was 0.863 when no additional item, if deleted from the scale set, could significantly improve the scale reliability. The Cronbach's alpha coefficients for each dimension were greater than 0.73, above the suggested threshold of 0.70 as acceptable score of construct reliability (Nunnally & Bernstein, 1994). A Harman's one-factor test was assessed for the results of the common method effect on our variables (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). The analysis cross-checked if common method variance existed; it might be either because of a single factor emerging in a factor analysis, or because of one general factor that accounted for the majority of covariance among all the variables. At the beginning, exploratory factor analysis results showed the largest factor not accounting for the variance in majority. Empirically, no general factor was apparent and significant. Then, all the variables were loaded onto a factor to examine the regression fitness. If the one-factor model fit the data well, the resultant factor model had a poor fit. The common method variance was not an issue against the result interpretation.

The study proceeds the analysis with the maximum likelihood method to finalize the measurement, and assesses a number of fit indices: GFI, goodness-of fit index, AGFI, adjusted goodness-of-fit index, NFI, normed fit index, CFI, comparative fit index and TLI, Tucker–Lewis index, which have to be greater than 0.9. RMSEA, root mean square error of approximation, has to be less than 0.08. The χ2 value to the degree of freedom (χ2/df) should not exceed 3. Table 2 shows a final measurement model of six dimensional constructs from 17 scale items. All the factor loadings are above 0.5, and significant (p<0.001). The model fit indices demonstrate an acceptable fit with: χ2 (df)=250.968 (102); χ2/df=2.46; p=0.000; GFI=0.9; NFI=0.90; AGFI=0.87; CFI=0.94; TLI=0.92; and RMSEA=0.07.

Convergent validity as an internal consistency is the degree of interrelatedness among the observed items, i.e. the unidimensionality and AVE, average variance extracted. In this study, convergent validity is an AVE ranging from 0.576 to 0.701 (Table 2). All indicators of AVE for each item exceed the suggested cut-off value of 0.50 for the convergent validity. The study evaluates construct validity using standardized factor loadings within each construct, AVE, and correlation between constructs (Anderson & Gerbing, 1988; Bagozzi & Yi, 1988). The standardized factor loadings on the latent construct are greater than 0.5 for meaningful and interpretable model results. All the loadings on all the latent constructs are satisfactory, thus indicating satisfactory item convergence on the intended dimensional constructs. The CR, composite reliabilities in the measurement scale were greater than 0.70, thus exceeding the suggested cut-off of 0.70. Table 3 shows the descriptive statistics, correlations and AVE for the measurement scale. Social esteem has the highest mean score and the lowest standard deviation among the six factors. The squared correlation between constructs is less than the individual AVEs, that confirms discriminant validity. In summary, the scale items of measurement fit the data reasonably well, thus demonstrating good reliability and validity.

The theoretical model consists of four exogenous variables, namely social esteem, desire for experiential travel, avoidance of rituality, and creative choice counter-conformity, and two endogenous variables, status consumption and affective response. MA shows a fair fit (Table 4): χ2/df ratio=2.71; p<0.001; GFI=0.90; CFI=0.92; TLI=0.90; RMR=0.075; and RMSEA=0.076. The model fits the data well with χ2/df ratio less than 3, CFI above 0.90 and RMSEA less than 0.08. The estimate of standardized path coefficient (P) indicates that the relationship between social esteem and status consumption (H1) is positively significant (P=0.24, p<0.001). Desire for experiential travel has the strongest positive impact on status consumption (P=0.50, p<0.001), which supports H2. The relationship between avoidance of rituality and status consumption is also highly significant and positive (P=0.35, p<0.001), thus supporting H3. However, the path between creative choice counter-conformity and status consumption (H4) is not supported. The confirmation of H5 indicates the relationship between status consumption and affective response positive and highly significant (P=0.54, p<0.001). The R2, squared multiple correlation coefficient measures how well the observed outcomes can be replicated by the proposed model, pertaining to the percentage of total variance explained by latent variables. The R2 for status consumption is 0.66, indicating that 66% of the total variance in status consumption can be explained by the four antecedent variables. About 29% in affective response can be explained by the other five variables.

This study analyzes and compares several competing models, similar to the approach adopted by Chi and Qu (2008). The assessment of competing models is to compare MA (Fig. 1) with other models which serve as alterative explanations. The aim is to decide the best fitting model among a set of competing models. The present study proposes two alternative models for comparison purposes (Fig. 2): MB and MC.

Fig. 2.

Competing SEM models.

(0.2MB).

As shown in MA, the flow path between ‘status consumption’ and ‘creative choice for counter-conformity’ is insignificant. As a subscale of ‘need for uniqueness’, ‘creative choice for counter-conformity’ positively affects attitude and emotional value (Kumar, Lee, & Kim, 2009). Therefore, MB excluded the insignificant path (H4) and added a direct path between ‘creative choice counter-conformity’ and ‘affective response’ (H6). In building personality profile, social esteem is related to self-liking in cognitive-affective regulatory process (Tracy, Cheng, Robins, & Trzesniewski, 2009). Based on this, MC further included a path between ‘social esteem’ and ‘affective response’ (H7). Table 4 shows the SEM results for all three models.

This study employs CFI difference tests and sequential chi-square (χ2) difference tests (SCDTs) to compare the nested structural models. The χ2 difference test between MA and MB (Δχ2=12.29; Δdf=1) suggests that MB performs significantly better than the theoretical model MA. The χ2 difference test between MB and MC (Δχ2=6.25; Δdf=1) suggests that MC performs better than MB. This analysis achieves a significant improvement in fit. CFI differences exceeding 0.01 between models would indicate loss of fit in a nested model. The study also assesses Akaike information criterion (AIC) and expected cross-validation index (ECVI) to confirm the choice for the best fitting model. The preferred model is the one with the lowest AIC value. The model with the smallest ECVI exhibits the greatest potential for replication. MC has the lowest AIC and ECVI among the three models and is preferred. The above assessment indicates that MC performs better than MA and MB.

The path linking creative choice counter-conformity and status consumption is not significant in MA. This implies that there is no direct relationship between creative choice for counter-conformity and status consumption. Therefore, this path is excluded and the extra path is included in MB. This model confirms that the relationship between creative choice counter-conformity and affective response is highly significant and strong (H6: P=0.40, p<0.001). The fit indices of MB are better than those of MA. As shown in Table 4, the model fits the data well: χ2/df ratio=2.62; p<0.001; CFI=0.93; TLI=0.91; GFI=0.91; RMR=0.064; and RMSEA=0.074. The χ2/df ratio is less than 3. Both CFI and GFI are larger than 0.90. The RMSEA is less than 0.08, thus showing a closer data fit. With the exception of creative choice counter-conformity, the magnitude and direction of other antecedent variables on status consumption are very similar. Again, desire for experiential travel has the strongest positive impact on status consumption (P=0.47, p<0.001), followed by avoidance of rituality (P=0.28, p<0.001) and social esteem (P=0.23, p<0.001). For the value of R2, about 48% of the total variance in status consumption in MB can be explained by social esteem, desire for experiential travel and avoidance of rituality. Approximately 34% in affective response can be explained by the other five variables.

For MC, an additional path between social esteem and affective response was included. MC fitted the data well (Table 4): χ2/df ratio=2.58; p<0.001; CFI=0.93; TLI=0.91; GFI=0.91; RMR=0.065; and RMSEA=0.073. This set of indices meets the model fit criteria. As compared with MB, the strength and direction of other causal paths in MC are similar with the exception of a few disparities. The relationship between social esteem and affective response is positively significant (H7: P=0.19, p<0.001). This suggests that there should be a direct path between social esteem and affective response. The R2 values for status consumption (0.48) and affective response (0.34) in MC are the same as MB. With significant difference in χ2, the lowest AIC and ECVI, MC provides the best fit among all the tested models. This study achieves a significant improvement in model fit. Fig. 3 illustrates the final SEM model in the study.

Fig. 3.

Finalized SEM model with parameter estimates. Note: ***p<0.001; **p<0.01.

(0.12MB).