Parasocial Interaction, Destination Personality Fit and Consumers’ Behavioral Intentions: The Case of TV Shopping

This study examines whether a TV shopping host’s introduction and promotion on a tourism destination could develop viewers’ parasocial interaction (PSI) with that host and thus influence viewers’ behavioral intentions on that destination. In addition, moderating effects of two factors (self-destination personality fit and host-destination personality fit) on the relationship between the TV viewers’ parasocial experience with a TV host (PSI) and the viewers’ behavioral intentions on the advertised destination are tested. Data were collected from 310 TV shopping customers in Taiwan using a self-administrated questionnaire. Simple linear regression and hierarchical regression techniques were implemented to test the proposed hypotheses. Findings indicate that PSI between a TV host and TV viewers is strongly correlated with the significant positive behavioral intentions of viewers. In addition, self-destination and host destination personality fits were both reinforcements of the positive impact that PSI has on TV viewers’ behavioral intentions. Findings and data collected from this study would yield useful practical implications for tourism marketers in terms of wisely selecting a TV host and utilizing a TV host’s parasocial interaction with audiences to increase audiences’ behavioral intention on the promoted destination.


Introduction
Similar to online shopping, TV home shopping has become an important alternative shopping medium to brick-and-mortar stores [1,2]. This is particularly true in Taiwan, as more and more retailers have adopted such a business model to market and sell their products [3]. Many tourism businesses in Taiwan have also benefited from such a business model, as an increasing number of travel deals and packages are sold through this manner. Having seen this opportunity, two of Taiwan's biggest TV home shopping channels, ETMall and MOMO shop, have collaborated with many travel agents to promote various package deals for local and/or foreign destinations. It is estimated that there is an audience of around 10.7 million people engaging with these two TV shopping channels combined, constituting 50% of Taiwan's whole population [4]. With this tremendous number of TV home shopping audiences, the question of how to increase this audience's purchase rate on those travel package deals becomes a critical task for many Taiwan tourism businesses to increase their revenue.
The semiotic application approach could be a theoretical underpinning of why many destination markers use TV shopping as a channel to promote tourism destinations. As the study of signs and symbols, semiotics conveys meaning through language, gestures, or any other visible signs, which all serve as essential elements for information exchange and communication. Every individual can interpret a sign differently based on their social and cultural background [5]. According to Peirce's (1839-1941) semiotic model, the semiotic process consists of an object, a sign, and an interpretant. In the tourism context, the object is the destination itself, the signs are the advertisement's information message, and the message receivers or interpretants are the target tourists [5].
Consider city branding: the advertisement message is in a specific context, indicating its meaning to the city's target tourists [6], who are active recipients of information messages. Specifically, the advertisement's signs, symbols, and messages are important determinants of consumers' interpretation of the city. Therefore, making good use of important city signs and symbols and designing appealing messages in advertisements should create a favorable city image in consumers' minds, which in turn influences their visit intention to that city. Similarly, when consumers see a tourism destination promoted by a TV host in a TV shopping program, they interpret all of the signs, symbols, and messages presented in the show and delivered by the host, which forms their image of the promoted destination. Therefore, the content presented in the TV program and how the TV host promotes the destination and interacts with consumers are critical determinants of how consumers interpret those signs, symbols, and messages, and form their impression of the promoted destination. The preceding discussion explains why TV shopping becomes a popular channel for destination marketers to promote tourism destinations. It also emphasizes the important role of a TV host in promoting a tourism destination in a TV program.
In comparison with online shopping, TV home shopping is distinct in two aspects. First, while online shopping requires the necessary information technology skills of consumers', TV home shopping is relatively more convenient/accessible, and thus has the potential to attract more consumers. Second, having a host to introduce and promote products is a unique feature of TV home shopping, whereas online shopping rarely adopts such a business format [7]. Due to virtual face-to-face conversations between a TV host and TV viewers, TV home shopping easily promotes an interpersonal relationship between the host and the viewers, which could further increase viewers' intentions to purchase the proffered products [8,9]. The relevant literature suggests that the success of this shopping mode is due to so-called "parasocial interaction", a virtual interaction between media users and the media [10,11].
Many prior empirical studies have shown that parasocial interaction (hereafter PSI) is an influential factor of media users' cognitive, affective, and behavioral outcomes, such as enjoyment, selective exposure, attitude change, social learning, and satisfaction with the media (e.g., Lim and Kim, 2011;Madison and Porter, 2015) [1,12]. This literature provide an important implication for destination marketers who adopt TV shopping as one of the major distribution channels to promote their destinations, as developing and strengthening PSI might help to generate more audiences' positive attitudes with the TV program and thus lead to better sales volume and reputation for the program. In the tourism literature, however, research on how PSI might affect media users' attitudinal and behavioral outcomes is still lacking. In addition, while most research on consumers' shopping behaviors focuses on the online shopping context, the TV shopping context receives little attention. To address these gaps, the current study aims to investigate whether TV shopping hosts' introduction and promotion of a tourism destination could develop a viewers' parasocial experience with that host, which in turn would influence viewers' behavioral intentions concerning that destination.
Furthermore, this study addresses two possible moderators regarding the relationship between PSI and TV viewers' behavioral intentions. They are viewers' perceived self-destination personality fit and perceived host-destination personality fit. The term "destination personality" refers to using human personality traits when describing a destination [13,14]. As with brand personality, where consumers appreciate and select compatible brands that display their actual and desired self [15,16], destination personality is found to increase consumers' satisfaction and visit intention concerning a certain destination when that destination fits well with consumers' self-image (i.e., consumers' self-congruity) [17,18]. On this basis, the current study proposes that, in the TV shopping context, perceived self-destination personality moderately fits with the influence of PSI on TV viewers' behavioral intentions.
On the other hand, an advertisement is, for all intents and purposes, the description and promotion of a product. The audience understands an advertisement or the product itself through the techniques used in the advertisement. Therefore, advertising is a typical mode of information transmission. Advertising itself is used to focus on persuasion and communication, conveying information to consumers to influence their purchase behavior [19]. However, the audience could have different interpretations of the information conveyed by the advertisement, depending on the positioning of the information, and thus the advertising appeals. Semiotically, in addition to the explicit meaning, advertisements have meanings and ideologies not readily apparent in the text, where the concealed meanings require interpretation by the audience with the social knowledge they possess [5]. Therefore, it is necessary that human emotions be considered along with the use of techniques to persuade consumers to influence their mindset to become amenable and then form the desire to buy and generate purchase behavior.
Marketing literature has suggested that a celebrity endorsement of a certain product can generate favorable attitudinal and behavioral responses from consumers towards that specific product. Compatibility amidst the personality of the celebrity and the personality of the endorsed product is essential in producing such a positive effect [20,21]. This rationale also provides an important implication for TV shopping programs that a TV host, functioning in a similar role to that of a celebrity who promotes and endorses a certain product, might become more appealing and convincing on TV if he/she is perceived as having a similar personality or image related to the advertised product. In line with this logic, in the context of TV shopping, TV viewers' host-destination personality fit might also play an important moderating role for the effect of PSI on TV viewers' behavioral intentions in regard to the advertised destination.
In summary, the purpose of this study is two-fold. First, to empirically test and determine whether or not TV viewers' parasocial experience with a TV host (i.e., PSI) can significantly impact viewers' behavioral intentions on the advertised destination. Secondly, to identify whether the two factors (self-destination personality fit and host-destination personality fit) may have a significant moderating effect regarding the influence of PSI on TV viewers' behavioral intentions. The results and conclusions originating from this study should produce insightful connotations conducive to destination marketers and TV shopping programmers.

Parasocial Interaction and Behavioral Intention
Parasocial interaction was first introduced by Horton and Wohl [22] and was defined as a "simulacrum of conversational give-and-take" (p. 215) that audiences experience as a response to a media performer (the "persona") in a media exposure situation. They suggest that audiences regard parasocial interaction as an intimate reciprocal social interaction, even though they know it is only an illusion. Hartmann and Goldhoorn [10] incorporated Horton and Wohl's [22] original ideas with more recent socio-psychological insights and defined an audience's parasocial interaction experience as "a felt reciprocity with a TV performer that comprises a sense of mutual awareness, attention, and adjustment" (p. 1107). They also suggested that a media performer's cues such as gazing and bodily-addressing cues could effectively induce viewers' active mindreading activities. This in turn leads to a parasocial experience, and is understood as a viewer's intuitive feeling of participating in a normal social interaction.
As TV shopping program users are unable to physically experience the advertised merchandise, a TV host needs to provide detailed descriptions of the functions and characteristics of the promoted product in order to reduce audiences' perceived risks while also providing them with a feeling of trust and authority [8,9]. The role of a TV host becomes even more significant when the promoted product is a destination, as tourism services have unique features (i.e., intangibility, perishability, heterogeneity) that differ from that of manufactured products. Consequently, as the major middleman between the audiences and the advertised destination, the TV host's performances, behaviors, and professionalism would be highly critical factors in determining audiences' perceptions, attitudes, and behavioral intentions related to the advertised destination.
The preceding discussion outlines the importance of not only the professionalism of the TV shopping host, but also of the parasocial interaction effect that a TV shopping host could bring to the audiences. According to Grant et al. [23], parasocial interaction is one critical criterion that audiences consider when selecting TV programs, as audiences who engage in parasocial interactions with a TV host may see the host as more of a friend providing advice rather than a sales agent marketing a product. A number of empirical studies have also found that TV viewers who engage in pleasant parasocial interactions with TV hosts spend more time watching the program [24,25], make more purchases from that program [26], and are more satisfied with that program [1] altogether. Based upon the results and implications from these prior studies, this current study proposes that, in the context of TV shopping with a TV host promoting a tourism destination, TV viewers' parasocial experience with the host should positively influence their behavioral intentions concerning that destination. Hence, the following hypothesis is developed:

Hypothesis 1 (H1).
In the context of TV shopping, PSI positively influences TV viewers' behavioral intentions concerning the promoted destination.

Self-Destination Personality Fit as a Moderator
Aaker [27] suggested that consumers often regard product brands as having human characteristics; moreover, she defined brand personality as "the set of human characteristics associated with a brand" (p. 347). For example, someone may describe popular brands such as Coca Cola as "cool", Pepsi as "young" [27], Marlboro cigarettes as "masculine" [28], and BMWs as "sophisticated" [28]. In line with Aaker's [27] rationale of brand personality, Ekinci and Hosany [29] introduced the concept of destination personality, defined as "the set of personality traits associated with a destination" (p. 127). Since then, numerous studies conducted on destination personality have applied Aaker's [27] brand personality measure to characterize different destinations, and have confirmed the effects that destination personality would have on tourists' behavioral intentions (e.g., Pan [13,17,18,30,31]. According to self-concept theory (or self-image theory), consumers usually use brands as part of their self-concept and for communicating their identities to others and to themselves [32]. Therefore, self-congruity theory, a natural extension of self-concept theory, suggests that consumers are largely attracted to the idea of evaluating and utilizing products or brands that share similar characteristics with themselves [33]. The self-congruity theory indicates that an individual's perceived match between his/her self-concept and a product's/brand's attributes is likely to affect his/her attitudes and behaviors related to that product/brand [34]. A number of previous studies have drawn the conclusion that the relationship between consumers' self-concept and a brand's image (i.e., self-brand image congruency) is positively correlated with the brand attitude, intention, and behavior [18,35,36]. Likewise, many tourism studies have also applied the self-congruity theory to the destination context and discovered that a good congruity between a tourist's and a destination's image (i.e., self-destination image congruency) would positively affect his/her attitudes and behaviors toward that destination [13,[37][38][39][40].
Hosany et al. [41] demonstrated in their study that destination image and destination personality are two associated but rather distinct concepts; destination personality pertains to destination image but is more akin to the affective aspects. Their study produced important empirical evidence suggesting that self-destination personality congruence, similar to self-destination image congruity, might also have a positive effect on consumers' attitu-dinal and behavioral outcomes pertaining to a destination. Based on this rationale, some tourism researchers have indeed found a positive influence of self-destination personality in relation to tourists' attitudes and behaviors toward a destination [30]. The literatures and discussions mentioned above prompt the current study to propose that, in the context of TV shopping with a TV host promoting a destination, the effect of PSI on TV viewers' behavioral intentions might be strengthened if the audiences perceive themselves as having similar personality traits to those of the promoted destination. Hence, the following hypothesis is formulated: Hypothesis 2 (H2). In the context of TV shopping, perceived self-destination personality congruity strengthens the effect of PSI on TV viewers' behavioral intentions concerning the promoted destination.

Host-Destination Personality Fit as a Moderator
Many modern marketers have adopted celebrity endorsement in advertising to increase sales [21], as such a method helps to transfer the symbolic meaning from the celebrity to the advertised product/service [42]. According to the meaning transfer model [42], consumers usually transfer various symbolic meanings about a celebrity to the promoted product through endorsement, and then from the product to the consumer through purchase. McCracken's study also suggests the idea that celebrity endorsements are more potent and successful if the celebrity and product are well-suited with one another in terms of cultural meanings and images. The notion of fit between endorser and product has been researched previously under the product match-up hypothesis. In the context of physical attractiveness, earlier studies have revealed consumers to be more persuaded by physically attractive celebrities when they promote products associated with physical beauty [43]. Other prior studies indicate that favorable compatibility between celebrity image and product image is a key factor in generating consumers' positive responses to advertising [44][45][46].
A TV host, who mainly describes and explains the functions and advantages of the promoted product, is a critical endorser of the advertised product. Based upon the rationale of the meaning transfer model [42] and the discussion above, the current study postulates that, in the context of TV shopping, consumers might also associate various symbolic meanings of the TV host with the advertised product and find the host more favorable and persuasive if he or she has a good match-up with the advertised product in terms of personalities. Such match-up might thus help strengthen the effect of PSI on the audiences' behavioral intentions. The following hypothesis is thus developed:

Hypothesis 3 (H3).
In the TV shopping context, perceived host-destination personality congruity strengthens the effect of PSI on TV viewers' behavioral intentions on the promoted destination.

Sampling and Data Collection
This study devised a self-administrated questionnaire to collect data. The questionnaire contained two sections. The first section addressed questions of the variables in the proposed model while the second section dealt with participants' demographic information. This study used a purposive sampling technique. The purpose was to understand TV viewers' parasocial experiences with a TV host; hence, only those who had ever watched one or more TV shopping channels with a TV host introducing and promoting a certain tourism destination in the past three months were included in the sample. Additionally, this study only included those who purchased the travel deal using the TV shopping channel for the first time, as those who are repetitive purchasers might have already formed preferences on certain TV shopping channels or TV hosts, which could bias the results of this study. To meet the above criteria, initially, researchers contacted and asked the public relation manager of the TV shopping channel to provide a list of 1000 members who met these criteria. Since Mandarin was the first language of the participants in this study, a back translation method was conducted to translate all English questions into Mandarin. Afterwards, the finalized survey was sent out to these 1000 members with a retuned envelope and a flyer about a sales activity promoted by the TV shopping channel. The flyer also specified that anyone who fills out the survey and mails the survey back to the TV shopping channel would be included in a lottery draw activity to have a chance of winning the grand prize. Three hundred and ten valid responses were ultimately collected.

Measurements
All question items were well-established scales with minor revision to fit the context of this study. For parasocial interaction between the TV host and the subjects, 6 items from Levy [46] were employed (sample item: "The TV host's introduction and promotion of that destination gave me a visual experience of traveling to that destination"). For selfdestination personality congruity, 4 items from Sirgy and Su [39] were adopted (sample item: "I considered that the personality of the destination is consistent with my personality characteristics"). These 4 items were also used to measure the host-destination personality congruity by changing the context into comparing the personality congruity between the TV host and the destination (sample item: "I considered that the personality of the destination is consistent with the TV host's personality characteristics"). As for behavioral

Sampling and Data Collection
This study devised a self-administrated questionnaire to collect data. The questionnaire contained two sections. The first section addressed questions of the variables in the proposed model while the second section dealt with participants' demographic information. This study used a purposive sampling technique. The purpose was to understand TV viewers' parasocial experiences with a TV host; hence, only those who had ever watched one or more TV shopping channels with a TV host introducing and promoting a certain tourism destination in the past three months were included in the sample. Additionally, this study only included those who purchased the travel deal using the TV shopping channel for the first time, as those who are repetitive purchasers might have already formed preferences on certain TV shopping channels or TV hosts, which could bias the results of this study. To meet the above criteria, initially, researchers contacted and asked the public relation manager of the TV shopping channel to provide a list of 1000 members who met these criteria. Since Mandarin was the first language of the participants in this study, a back translation method was conducted to translate all English questions into Mandarin. Afterwards, the finalized survey was sent out to these 1000 members with a retuned envelope and a flyer about a sales activity promoted by the TV shopping channel. The flyer also specified that anyone who fills out the survey and mails the survey back to the TV shopping channel would be included in a lottery draw activity to have a chance of winning the grand prize. Three hundred and ten valid responses were ultimately collected.

Measurements
All question items were well-established scales with minor revision to fit the context of this study. For parasocial interaction between the TV host and the subjects, 6 items from Levy [46] were employed (sample item: "The TV host's introduction and promotion of that destination gave me a visual experience of traveling to that destination"). For self-destination personality congruity, 4 items from Sirgy and Su [39] were adopted (sample item: "I considered that the personality of the destination is consistent with my personality characteristics"). These 4 items were also used to measure the host-destination personality congruity by changing the context into comparing the personality congruity between the TV host and the destination (sample item: "I considered that the personality of the destination is consistent with the TV host's personality characteristics"). As for behavioral intentions concerning the promoted destination, this study adopted 3 items from Lee et al. [47] (sample item: "After watching the host's introduction and promotion concerning the destination, I would consider purchasing the deal for traveling to that destination"). All question items were anchored with a 5 point-Likert scale to record and evaluate the participants' responses.

Analysis
Descriptive and correlation analysis were conducted to showcase participants' demographic profiles and correlations between the constructs. Cronbach's α was calculated to test internal consistency of each construct. The first hypothesis was tested by means of one simple linear regression analysis, whereas the second and third hypotheses were tested by means of two individual hierarchical regression analyses. Since this study is crosssectionally designed, Harman's single factor test was also utilized to determine whether common method variance would be a serious problem.

Results
This research investigates whether or not TV viewers' parasocial experience with a TV host (i.e., PSI) can significantly impact viewers' behavioral intentions on the advertised destination and identifies whether the two factors (self-destination personality fit and host-destination personality fit) may have a significant moderating effect regarding the influence of PSI on TV viewers' behavioral intentions. A total of 310 valid responses were collected and analyzed using the SPSS. Table 1 shows that the subjects were comprised of 132 males (42.6%) and 178 females (57.4%), with the majority of subjects (87%) being in the 18 to 40 year-old age range; 88.5% of the subjects have an annual income of TWD 20,000 or lower while 11.5% of the subjects have an annual income higher than that of TWD 20,000. Moreover, 70% of the subjects have a college degree or lower, whereas 30% have a master's degree or higher.  Table 2 shows that all constructs are positively correlated (p < 0.01), with their respective Cronbach's Alpha values all above 0.70, suggesting acceptable internal consistency. In addition, Harman's single factor test revealed 41% of the total variance, indi-cating that common method variance was not a critical threat to the magnitudes of the hypothesized relationships.  Table 3 displays the CFA results. The results were an indication of all standardized factor loadings being significant and above 0.50 (ranging from 0.50 to 0.90), which suggests a satisfactory convergence validity [48]. Moreover, all constructs' average variance extracted (AVE) were greater than 0.5 (ranging from 0.5 to 0.73), which also represents a good convergence validity. Discriminant validity was tested by means of comparing AVE with the coefficient of determination (square of the correlation coefficient) for any two constructs [49].

Descriptive Statistics and CFA Results
The AVE values of all pairs of constructs were larger than their corresponding coefficients of determination (squared correlation coefficient), indicating a satisfactory discriminant validity. The overall fit of this final measurement model was X 2 (113, N = 310) = 438.99 (p < 0.01); X 2 /d.f. = 3.88; CFI and TLI were both 0.90. SRMR and RMSEA were 0.07 and 0.08, respectively. The numbers from all indexes suggested an acceptable fit for the measurement model. Lastly, the data collected from Harman's single factor test resulted in a single-factor model accounting for just 41% of the total variance, implying that common method variance was not a critical threat Note: X 2 = 438.99 (df = 113, p < 0.00001); X 2 /df = 3.88; CFI = 0.90; TLI = 0.90; SRMR = 0.07; RMSEA = 0.08. All factor loadings are significant at p < 0.01.

Hypotheses Test Results
The regression analysis results coincided with our H1 in that parasocial interaction has a positive influence on behavioral intentions (b = 0.74, p < 0.01). As Table 4 presents, parasocial interaction × self-destination personality fit interaction term (∆R 2 = 0.01, p < 0.05) adds significant incremental variance, indicating that self-destination personality fit does indeed play a significant moderating role with regards to the relationship between PSI and behavioral intentions. The interaction plot was then created to show a clear moderating effect of the self-destination personality fit. As Figure 2 indicates, subjects who perceive high or low self-destination personality fit do not have a significantly different level of behavioral intentions when the parasocial interaction effect is low. However, subjects who perceive high self-destination personality fit report significantly higher behavioral intentions than those who do perceive a low self-destination personality fit when the parasocial interaction effect is high. This provides support for H2, namely that perceived self-destination personality congruity strengthens the effect of PSI on TV viewers' behavioral intentions. Table 4. Moderating effect of self-destination personality fit on the relationship between parasocial interaction and behavioral intentions (N = 310).

(Step 1)
Parasocial interaction 0.69 ** Self-destination personality fit 0.11 * R 2 0.56 ** ( Step 2) Parasocial interaction × Self-destination personality fit 0.08 *  Table 5 displays the results of another hierarchical regression analysis concernin moderating effect of the host-destination personality fit on the relationship betwee and behavioral intentions. The interaction term of parasocial interaction × host-destin personality fit also adds significant incremental variance to the equation (∆R 2 = 0.0  Table 5 displays the results of another hierarchical regression analysis concerning the moderating effect of the host-destination personality fit on the relationship between PSI and behavioral intentions. The interaction term of parasocial interaction × host-destination personality fit also adds significant incremental variance to the equation (∆R 2 = 0.01, p < 0.01), suggesting that the host-destination personality fit is also a significant moderator for the relationship between PSI and behavioral intentions. Another interaction plot was graphed to show the exact moderating effect of the host-destination personality fit. As Figure 3 presents, subjects who perceive a high or low host-destination personality fit do not report a significantly different level of behavioral intentions when parasocial interaction effect is low. However, subjects who perceive a high host-destination personality fit report a significantly higher level of behavioral intention than those who perceive a low hostdestination personality fit when the parasocial interaction effect is high. This lends support to H3, namely that the perceived host-destination personality congruity strengthens the effect of PSI on TV viewers' behavioral intentions.

Discussions
The findings reveal that, in terms of TV shopping, PSI between a TV host and TV viewers has a positive influence on viewers' behavioral intentions. This finding corre sponds with a number of prior studies exploring the effects of PSI in the context of TV

Discussions
The findings reveal that, in terms of TV shopping, PSI between a TV host and TV viewers has a positive influence on viewers' behavioral intentions. This finding corresponds with a number of prior studies exploring the effects of PSI in the context of TV programming, i.e., that PSI significantly influences viewers' attitudes and behaviors (e.g., Hartmann and Goldhoorn, 2011;Lim and Kim, 2011;Madison and Porter, 2015;Su et al., 2011) [1,10,12,50]. In terms of contribution to the tourism literature, such a finding not only addresses the research gap mentioned earlier (that there are scant studies addressing the possible beneficial effect of PSI on consumers' attitudes and behaviors), but also provides an important implication for tourism researchers in marketing that PSI might also exist in other kinds of media environments in addition to TV shopping, such as social media (e.g., Facebook Instagram, Twitter, YouTube, Blogs, etc.) which involve virtual one-to-one or one-to-many conversations and interactions.
The results also support the idea that self-destination personality fit strengthens the positive influence of PSI on TV viewers' behavioral intentions. Such a finding suggests that the personality congruity between one's self-concept and a certain destination is not only a significant predictor of one's attitudinal and behavioral outcomes toward a certain destination [17,18,30], but also an important moderator that strengthens the influence of PSI on one's attitudes and behavioral intentions toward a certain destination. This discovery enriches the understanding of extant tourism literature in that the rationale of self-congruity theory can also be applicable in the context of tourism destination (i.e., the compatibility between destination personality and consumers' self-concept) and sheds light on the fact that this theory has an important interaction effect with PSI in affecting consumers' behavioral intentions in the context of TV shopping.
Furthermore, the findings of this study reveal that the host-destination personality fit also strengthens the positive influence of PSI on TV viewers' behavioral intentions. This result suggests that the significant effect of personality congruity on one's attitude and behavioral intention toward a certain destination can be generated not only from the congruity between oneself and the destination, but also from the congruity between the destination and the third party who promotes that destination (i.e., the TV host). Such a finding also contributes to the existing tourism literature in that the rationales of the meaning transfer model [42] and match-up hypothesis (the compatibility between an endorser and an endorsed product) can be applied in the context of tourism destinations as well.

Implications
The findings of this study supply destination marketers with several managerial implications with respect to developing more effective TV shopping strategies to promote their destinations. First, the results suggest that consumers' parasocial experience with the TV host increase their intention to purchase the deal of traveling to the advertised destination. Therefore, destination marketers opt to choose a host who is an expert in utilizing vivid conversations and interactive approaches when advertising/promoting a vocational destination in a TV shopping program, as PSI takes place when a media user perceives the media performer as an intimate conversational partner [22]. It should also be noted for the TV host to address the audiences on both a bodily and verbal level during the show. Bodily addressing implies acknowledging and making eye contact with the audience on a consistent basis, whereas verbal addressing denotes directly referring to the audience during the show, for example, in the opening line, such as "good evening ladies and gentlemen", or by making remarks during the show such as "our audience is probably wondering why we are doing this" [8]. The literature suggests that both addressing styles of TV performers could intensify the viewers' parasocial experience [10].
Additionally, the findings of this study indicate that a good personality fit between the advertised destination and the audiences would help enhance audiences' parasocial experience with the TV host, thereby raising audiences' behavioral intentions to make a purchase. Hence, the TV host should prioritize the creation of a relation between destination personality and audiences' self-concept during the show. For example, destination marketers could first conduct a pretest to identify what the personality traits (e.g., sophistication, competence, sincerity, excitement, and so on) that the target audiences perceive concerning the destination to be promoted. Having such information, the TV shopping host could consistently use and highlight these personality characteristics and terms in the show when describing and promoting the destination, which should help draw more attention and arouse more interest from audiences that perceive that they have similar personality characteristics to that of the advertised destination.
Lastly, this study also reports that a good personality fit between the advertised destination and the TV host helps to enhance audiences' parasocial experience with the TV host, which boosts audiences' behavioral intentions to purchase a deal. Consequently, when selecting a TV host for a promoting a tourism destination, destination marketers should note whether the host and the promoted destination are well fitted in terms of personality characteristics. For example, it might be better to choose a younger and more energetic TV host to advertise a destination which is perceived as young, energetic, and exciting, and choose a more sophisticated and competent host to advertise a destination regarded as more sophisticated.

Limitations
The limitations in this study that need to be addressed in future research are as follows. The first is the lack of diversity among participants and the requirement of additional examinations to increase the generalizability of its results. Because this study was solely conducted with Taiwanese participants, the findings and implications may not be generalizable to other regions and countries. Therefore, future research can consider including participants from different countries or cultural backgrounds to validate the generalizability of the model in this study. Second, the study only included respondents who had purchased a travel deal through a TV shopping channel. However, whether such a parasocial interaction effect can occur in other media contexts is unknown. Therefore, future research could examine the parasocial interaction effect when consumers purchase products/services on social media platforms such as Facebook and YouTube.
The third limitation could be the uncertain causal relationship between independent and dependent variables, as this study adopted a one-wave, cross-sectional design to collect data. Future research should thus try to collect the data of predictor and outcome variables separately at different points in time to identify more accurate causal relationships between predictor and outcome variables.