How Instagram Influencers Contribute to Consumer Travel Decision: Insights from SEM and fsQCA

In recent years, social media influencers (SMIs) have become independent entities capable of influencing their audiences. Businesses look towards SMIs on Instagram as a marketing communication tool due to their popularity and effectiveness. In tourism, few studies investigate the role of SMI in influencing travel behaviors. The study examines how SMI can influence Instagram followers' travel behaviors. This study determined SMI based on attractiveness, similarity, and expertise. SMI can generate parasocial interactions with followers and build trust by promoting destinations based on these three dimensions. This study tested the research hypothesis on 364 respondents using a dual-approach analysis of structural equation modeling and fuzzy set qualitative comparative analysis. The results of the SEM analysis confirm that consumers were more likely to trust SMI based on their expertise and similarity. Also, this study demonstrated that highly attractive SMI and similarity with followers can lead to parasocial interactions. When consumers trust and feel parasocial interaction with SMI, they are more likely to consider traveling. FsQCA results confirm the presence of two configurations with high travel intention. The causal conditions configuration presented in this study demonstrated the interdimensional relationship between SMIs (attractiveness, similarity, and expertise), trust, and parasocial interaction in terms of travel intention. This study also achieved theoretical contributions and managerial implications on how scholars and tourism managers leverage SMIs to create high travel intentions. Intention; Fuzzy Set Qualitative Comparative Analysis.

Instagram has grown in popularity as a new and effective medium for businesses and brands to communicate with customers. Furthermore, SMI, through Instagram, can supplement marketing communications and influence consumer purchasing decisions [8,9]. The role of SMI in tourism is to promote tourist destinations, which is an essential component of increasing visits [10]. SMI focuses on developing content about tourism, culture, and travel experiences that is the best option for tourism marketers [11]. SMI is called a "travel influencer" in the tourism industry because they can drive followers to vacations with photos or videos on Instagram [12]. The effectiveness and increasing popularity of SMIs has forced many tourism marketers to use SMIs as a marketing tool to promote tourist destinations [9,13]. This will allow tourists to obtain a great deal of information about tourist destinations they share and are interested in visiting [14]. Zhang & Huang (2021) [15] also asserted that the presence of SMI can significantly impact people's travel decisions and the number of people who visit certain places.
Tourism marketers are increasingly interested in hiring SMIs to influence tourists' purchasing decisions. (1) SMI can inspire followers' trust in the services offered by the tourism marketers to influence their purchasing decisions. At this point, SMI is an "opinion leader", influencing followers to travel to destinations posted on their Instagram [16]. Marketing strategies using SMI are also widely used since they have successfully gained the trust of their followers [17]. The interaction between the content created by SMI and their followers on Instagram creates direct engagement, increasing trust and influencing tourists' purchasing behavior [1,18]. (2) SMI helps destinations reach a larger audience and increase visits through social media. The easiest way to reach tourists and strengthen communication is through social media, especially with the support of the SMI in interacting and sharing their experiences with tourists [9]. Furthermore, tourists' visits will increase, and the decision to visit the destination will be promoted [19]. (3) SMI can be a marketing tool with significant cost benefits and ubiquitous access. Using SMI to promote specific products or services is possible without incurring high costs. Different social media platforms can help marketers and SMIs figure out how to reach the right target audience.
In addition to its increasing popularity, SMI significantly impacts online marketing. Travel influencers can communicate with their followers and establish trust in the promoted destination [20]. It is critical to promote tourist destinations through SMI to influence tourists to visit [12]. For this reason, it is necessary to know several essential criteria considered by SMI when promoting tourist destinations through posts shared by influencers on Instagram. This study aimed to identify how the SMI dimensions, such as attractiveness, similarity, and expertise, can be used to create parasocial interactions with followers and build trust [21,22]. The formed trust influences the visitors' intention to travel to the promoted destination [17]. Through SMI, tourism marketers can make thoughtful decisions about marketing communications and utilize SMI in marketing tourism.
Scholars have recently paid close attention to the phenomenon of SMI influencing consumer behavior. In particular, SMI are increasingly crucial in influencing consumer decisions in the tourism sector [15,16,20]. Scholars have incorporated several perspectives in their investigations of SMI in the tourism industry. Pop et al. (2021) [10] examined how trust in SMI influences visiting decisions via Facebook posts. Choi et al. (2019) [23] investigate SMI, especially in online travel community platforms. They found that the formation of parasocial interaction and trust can influence tourist satisfaction by influencing tourists' intentions to visit. Magno & Cassia (2018) [24] explain that tourist interest in travel is influenced by the quality of information and the reliability of SMI, especially on online travel blogs. Meanwhile, SMI have identical and attractive characteristics that allow them to influence consumer behavior. Therefore, this study highlights SMIs' characteristics such as attractiveness, similarity, and expertise in influencing their audiences. Further, it is imperative to investigate the mechanisms that facilitate the formation of emotional bonds between the audience and the influencers. Due to the emotional state consumers feel towards influencers, they may make purchases as a result. Accordingly, this study identified a research gap in the platform used by SMI, specifically on Instagram, which warrants further investigation. This study aimed to examine SMI on Instagram to fill the research gap, emphasizing attractiveness, expertise, and similarity. This study also investigated how trust influences travel intention behavior and how Instagram fits this realm.
This study introduces a consumer decision journey model based on SMI, parasocial interaction, and trust in travel intentions. In addition, this study examined Instagram influencers, which are currently becoming very popular and can significantly impact consumer behavior. The role of SMI in influencing the pursuits of tourists' travel intentions by sharing photos and videos of tourist destinations is investigated. This study has also applied the parasocial interaction theory through specific Instagram influencers. Horton & Wohl (1956) [25] defined parasocial interaction as a face-toface illusion relationship between the audience and their favorite media figures. Parasocial interaction describes how social media influencers create relationships with their followers on Instagram based on their attractiveness, expertise, and similarity [26]. The study also found that these three things, together with the formed parasocial interaction, will create tourist confidence in the posts shared by SMI and ultimately affect the tourists' travel intention [19]. This study focuses on SMI on Instagram for the tourist destinations in North Sumatra, Indonesia. Glimpsing the pleasures of tourism in North Sumatra, such as natural scenery, culinary, and historical culture, can create a good impression on tourists. Therefore, it requires further development and promotion. The services of SMI can help increase tourist visits. That is why the Indonesian tourism industry is improving. Using Structural Equation Modeling (SEM) and fuzzy set qualitative comparative analysis (fsQCA), researchers and tourism businesses can make better decisions about using SMI as a marketing tool to get more people to visit North Sumatra, Indonesia's tourist destination.
The paper remains structured as follows: Section 2 presents the theoretical background and literature review. Section 3 is hypothesis formation and developing a conceptual framework. Section 4 is the methodology that explains the research design, operationalization, and development of questionnaire items, sampling techniques, data collection methods, and statistical analysis techniques (SEM and fsQCA). In Section 5, this study presents the results of our research and discussion. Lastly, Section 6 concludes with conclusions, implications, and directions for future research based on the research conducted.

2-1-Instagram and Tourism Marketing
The use of social media in the tourism industry plays an essential role in finding information, assisting tourists in decision making to visit a specific destination, as a marketing tool for destinations, and communicating with tourists [7]. Instagram is one of the most popular platforms used today. Tourists widely use a platform to find travel stimulation, build interest, and engage in social commerce [27]. Instagram can convey essential and valuable information using photos, videos, or text [28]. Moreover, consistent visual representations on Instagram can mean that the shared tourist destinations are appealing and suitable for a visit. As a result, Instagram can effectively attract tourists' attention and convince them to visit [29]. Consequently, Instagram has become a popular marketing tool used by tourism marketers.
Instagram can also enhance the destination image through direct tourist involvement [28]. The use of social media influencers (SMI) can enable effective and long-term communication with tourists [30]. Various content related to tourist destinations shared on Instagram by SMI will create a sense of involvement, social presence, trust, and enjoyment and impact travel intentions [6]. In addition, the content shared by SMI is considered more credible than standard advertisements, which is confirmed by previous research to influence tourists' travel intentions [2]. Moreover, the information from Instagram is valuable to tourists, but marketers may also use it to develop promotional strategies and improve destination image [31]. Marketers can use Instagram and SMI to better design and deliver their destinations and services to prospective tourists [32]. Tourism promotion can successfully drive higher tourist travel decisions by precisely utilizing Instagram and SMI attractions [15].

2-2-Social Media Influencer (SMI)
The concept of social media influencers (SMI) is a marketing practice where specific individuals promote destinations through posting a photo or videos on social media [33,34]. Every shared post has the potential to provide benefits to its followers in the form of information, motivation, role models, and attachment relationships [35]. As a result of its effectiveness and advertising power, SMI is used as a critical model for influencing consumer behavior in the tourism industry [8]. Therefore, marketing strategies involving SMI are widely used and have been found to gain the trust of their followers [17]. SMI's credibility is also essential to its success in using influencers in marketing [36]. There are several critical credibility factors that SMI must possess, namely: attractiveness, expertise, and similarity that can be perceived by their followers [20].
Attractiveness is the first dimension of SMI. It refers to the visual appeal and physical appearance [20]. SMI can induce curiosity in its followers and influence their decision-making through its attractiveness [37]. Furthermore, if a specific character is owned (i.e., well-known and unique), it can increase the sense of liking and interest in influencers' content [38]. The second dimension is expertise, which refers to SMI's ability to respond to specific inquiries or provide detailed instructions [39]. Once followers feel that the message conveyed is correct and not confusing, an influencer is recognized for his/her expertise [13]. SMI must ensure its message is compelling and persuasive by paying attention to the content shared through his/her social media channels [40]. The last dimension to consider is followers' similarity with SMI on social media. The similarity results from the emotional attachment that SMI has successfully established with its followers. The behavior of followers tends to accept the presence of influencers and respond favorably to the posts shared [41]. In this case, the similarity between SMI and followers can objectively be evaluated based on the facts, circumstances, appearance, and lifestyle they possess; and subjectively based on their similar perceptions [42]. It is believed that substantial similarities between influencers and followers will foster trust between them [43,44]. As a result, followers will think SMI is a suitable partner and feel closer to SMI [45].
As part of the tourism industry, the role of SMI is to ensure that each post looks attractive and raises the intention of tourists to plan a trip [12]. As a result of SMI's recommendations and activities in interacting and re-sharing their travel experiences on social media, tourists can select tourist destinations they are interested in visiting. In addition, the credibility it enjoys encourages tourists to trust the posts shared by SMI [13]. The information provided by an influencer is usually rational and objective to stimulate consumer interest and persuade them to make a tourist trip [11]. SMI will ultimately significantly impact tourists' decisions to travel to specific tourist destinations [10].

2-3-Parasocial Interaction with Social Media Influencer
The parasocial interaction concept was coined by Horton and Richard (1956) [25] as a face-to-face illusionary relationship between the audience and media figures such as celebrities, artists, and presenters. Parasocial interaction is defined as creating a friendly relationship between consumers and personas (characters or media figures) by interacting with them through certain media. According to Horton and Richard's study, the media figure in question is SMI on Instagram. Instagram can lead to parasocial interactions between influencers and followers, which is one of the critical factors in successful influencer marketing [46]. Furthermore, the support of messages and information provided by SMI can encourage the parasocial phenomenon with their fans and causes a positive response to the posts they share [47].
By engaging in parasocial interaction, followers will view SMI as friends, desire to meet them personally, and imagine them as favorite figures [8]. Further, if an SMI establishes a positive relationship with his followers, this will provide a sense of affordability for their presence [48]. Additionally, besides social and physical attractiveness, the credibility of the information supplied by SMI to their followers on Instagram can also result in solid parasocial interactions [45,49]. Ultimately, parasocial interactions will be persuasive and influence followers' behavior, such as purchasing and attitudes toward promoting the product or service [8]. In this sense, the purpose of SMI in tourism marketing is to introduce tourist destinations to tourists through interaction with them and to develop positive relationships. SMI must produce parasocial interactions that contribute to emotional attachments and affect tourists' trust in the promoted destinations [50]. This emotional attachment may enhance tourists' curiosity about shared tourist destinations [51]. Consequently, tourists' interest in tourist destinations shared by SMI is formed, influencing their decision to visit [52,53].

2-4-Trust in Social Media Influencer
For SMI, trust is a key element for establishing and maintaining a successful long-term relationship with followers [10]. SMI can establish trust through interaction with their followers through posts shared on social media [54]. Through the interactions, followers are encouraged to seek additional information, ultimately influencing their buying behavior [19]. Conversely, influencer followers tend to view SMI information with greater confidence since they are perceived as credible and experienced [36]. The experience and credibility of SMI were found to impact trust in SMI [17]. SMI is an active social media user whose followers listen to and regard them as a trusted source [50]. As a result of trust in SMI regarding shared tourism information, tourists are interested in and use it when planning trips [55]. Tourist confidence in a destination is influenced by the informational value of posts by social media influencers, including perceived attractiveness and similarity [21]. With the quality and expertise of SMI posts, the interest of tourists in following the travel advice provided by SMI also increases [24]. In the end, the influenced tourists by SMI are perceived as a reliable source, affecting the decisions of other prospective tourists [50].

3-1-1-SMI Dimensions and Parasocial Interaction
In tourism marketing, the role of SMIs is considered significant in introducing tourists to specific destinations and their influence on the decision to travel [10]. This is because the source of messages and information provided by SMI can stimulate parasocial phenomena among followers and result in positive reactions to posted content [47]. Thus, an influencer must possess attractiveness, expertise, and similarity that followers can identify with [45]. Their physical appearance determines their perceived attractiveness [20]. Influencer expertise provides information about promoted tourist destinations [39]. The similarities between followers and influencers can be seen in the facts, circumstances, appearance, and style evaluated similarly [42]. Based on these three factors, SMI is capable of creating parasocial interactions with followers [46]. By utilizing attraction and the perception of similarity, influencers can establish a positive communication relationship with their followers, forming a parasocial relationship [33]. The expertise possessed by influencers may also lead to positive parasocial interactions and influence tourists' behavior and travel decisions [45]. According to   [46], SMI's credibility can create parasocial interactions, which are critical to the success of influencer marketing programs. Accordingly, this study hypothesize:

3-1-2-SMI Dimensions and Trust
SMI's posts can influence followers' trust in them [56]. An influencer with many followers will likely be liked and regarded as popular [57]. Lou and Yuan (2019) [21] explain that SMIs are crucial for building trust with followers because of their popularity, credibility, and expertise [21]. This study investigated that SMIs must meet specific requirements to be trusted by tourists, including attractiveness, expertise, and similarity [20,44]. By providing information about tourist destinations, SMI will generate tourist travel intention through credibility, foster trust, and develop positive responses [22,55]. An influencer who can attract tourists based on his credibility may be able to construct an intention for their followers in what they share on social media [37]. Additionally, trust is established when followers believe that the message conveyed is accurate and not confusing [13]. Likewise, the similarity between influencers and their followers will also result in a close trust relationship [43,44]. With the three factors of credibility that have successfully influenced tourists, the information provided is considered reliable while planning tourist excursions. The following hypothesis is derived from the three dimensions of the credibility of social media influencers as it relates to consumer trust: H2ac. SMIs' attractiveness, expertise, and similarity significantly influence trust

3-1-3-Parasocial Interaction and Trust
SMI posts can create a parasocial interaction that contributes to tourist confidence to travel to particular destinations advertised through social media [50]. Parasocial interactions between influencers and followers can enhance trust, increasing interest in shared tourism information [37]. Moreover, consumer perceptions of influencers are likely to form an attachment relationship that will lead followers to believe the messages conveyed by SMI are more reliable when making travel plans [48,55]. SMI posts are generally regarded as authoritative and dependable sources of information. Therefore, they are effective sources of recommendations [22]. The creation of parasocial interactions will be a significant factor in promoting tourists' trust in the destination. The following is the hypothesis developed:

3-1-4-Trust and Travel Intention
SMI engages tourists through recommendations and activities to share their travel experiences and influence their decisions regarding tourist destinations [58]. Given this, SMI has a significant role in establishing trust and maintaining successful relationships with tourists through social media posts [10]. The travellers influenced by SMI posts are more likely to trust the source and seek additional information [22]. Informative and trustworthy posts with attractiveness, expertise, and similarity with followers will result in greater consumer trust, awareness, and intentions to visit [21,51]. The trustworthiness of information shared by influencers is crucial in generating interest in visiting tourist destinations and making the information useful for travel planning [12,37]. Therefore, the hypothesis is as follows:

3-2-1-Structural Model
This study investigated various theories by employing social media influencers (SMI) on Instagram for marketing tourism. This study investigated the influence of SMI, parasocial interaction, and trust on travel intention. The parasocial interaction theory describes the relationship constructed between SMIs and their followers. The SMI dimension emphasizes three aspects of influencers' credibility: the attractiveness, expertise, and similarity that followers perceive. The SMI dimension contributes to the development of parasocial interaction, which contributes to trust. The formation of trust ultimately affects the followers' intention to travel. Figure 1 illustrates the conceptual framework and research hypotheses in detail.

3-2-2-Embracing the Complexity of Theory with Configurational Models
This study investigated the relationship between the dimensions of SMI (attractiveness, expertise, and similarity), parasocial interactions, and trust toward high travel intentions. The key factors leading to the creation of travel interest have been identified as SMI, parasocial interaction, and trust in tourism marketing on Instagram [45,59]. Therefore, these casual conditions were combined to explain the results supporting the formation of tourist interest. The general H4 assumption is that tourists are likely to travel based on the recommendations provided by SMI. These recommendations feature charm, expertise, and similarities to posts on Instagram. However, it is also possible that tourists will travel based on the attractiveness of SMI posts. Furthermore, tourists may be more likely to travel if they perceive a close relationship between themselves and a social media influencer. The results of the relationship between the audience and the influencers will increase trust. The reasons for developing an interest in travel can vary greatly depending on the factors that motivate the development of such an interest. In this study, factors that promote a high level of interest in travel were examined. Accordingly, this study proposed that each configuration of social media influencers, parasocial interactions, and trust in travel intention is distinct and individual. As a result, the following conceptual framework has been developed in Figure 2.

4-1-Construct Operationalization and Item Development
The variables used in this study include SMI dimensions (attractiveness, expertise, and similarity), trust, and parasocial interaction. Table 1 illustrates the operational definition of each variable.

Construct
Operational Definition Adapted Sources

Attractiveness
A social media influencer's attractiveness consists of the ability to attract followers through their physical appearance, and the followers will judge whether the influencer is attractive.
Wiedmann & Mettenheim (2020) [20] Expertise Social media influencers' abilities to share information, provide answers and explain a topic accurately. Goldman (2018) [39] Similarity Describe the value perceived by the followers under the influence of social media influencers, objectively (state, appearance, and lifestyle) and subjectively (thoughts  [61] In this study, the measurement items were modified from previous research. Measurement items were selected and matched according to the research context ( Figure 2). This study modified the attractiveness scale from Wiedmann and von Mettenheim (2020) [20] and Weismueller et al. (2020) [37]; and five items were created. The expertise was obtained from Xiao et al. (2018) [36] and Ki et al. (2020) [41], and five were developed. The similarity adapted from Lou and Kim (2019) [45] were then developed into three items. Parasocial interaction was adapted from research by Hwang and Zhang (2018) [59] and consisted of three items. Each item is evaluated using a seven-point Likert scale (1 = strongly disagree, 7 = strongly agree).

Parasocial Interaction Trust
Travel Intention I see SMI as a competent person.

Similarity
In general, SMI has a lot in common with me.
SMI and I have an identical closeness.
SMI can be easily identified.

Parasocial Interaction
I am looking forward to SMI's post on Instagram.
If my favourite SMI shows up on other Instagram accounts, I will take a look.
I followed SMI on Instagram and became a fan of it.
I want to meet face-to-face with my favourite SMI.
I will read the content of magazines and blogs related to my favourite SMI.
Through their posts on Instagram, I feel like I know SMI.
I can rely on the information obtained from SMI.
I feel sorry if SMI made a mistake.
Trust SMI is trusted.
SMI is honest.
I trust the travel information provided by SMI.
In my opinion, SMI is more trusted than travel agents.
I believe SMI conveys information that does not harm its followers.
Travel Intention I intend to visit tourist destinations posted by SMI in the future.
I prefer to visit destinations posted by SMI over others.
If I have a chance, I will visit the destinations posted by SMI.

4-2-Sampling Technique and Data Collection Procedure
The study used non-probability sampling with purposive sampling. To be eligible to complete the questionnaire, respondents must be active users of Instagram, follow several travel influencers' accounts, and have visited destinations that SMI recommends. This is to achieve research objectives: examining visitors' travel intentions along the dimensions of SMI (attractiveness, expertise, and similarity), parasocial interaction, and trust. This study used Google Forms for a questionnaire and to compile the previously developed items. The questionnaires were randomly shared using Instagram, WhatsApp, and Facebook and received 364 responses.

4-3-Analysis Technique
This study used structural equation modeling (SEM) to analyze data with Smart-PLS 3.0. This study determined the measurement model's convergent and discriminant validity. Three metrics were used to assess convergence validity: Average Variance Extracted (AVE), Composite Reliability (CR), and Cronbach's Alpha. Using the Fornell-Larcker criterion, discriminant validity was determined by the square root of the AVE as it was more significant than the correlation between the other constructs. The heterotrait-monotrait ratio (HTMT) was also used to assess discriminant validity. The structural model was tested using two methods. First, measure the R-square value of the structural model based on independent variable factors. Secondly, compute the fit model indices: the SRMR, the d_ULS, the d_G, and the NFI. In addition, hypotheses were tested after the model fit criteria have been satisfied in structural equation modelling.
A fuzzy set qualitative comparative analysis (fsQCA) was used to assess the complexity of the theory and the interdependence between the variables. The fsQCA test has contributed to developing theory and practical guidance derived from the asymmetric analysis results. The fsQCA test began with the calibration of the seven-point Likert scale into fuzzy data, with "2" for full non-membership, "6" for full membership, and "4" for intersection [62]. The results were compiled into a truth table for further analysis. Based on the causal conditions of the independent variable, this study conducted data analysis to identify a solution to the dependent variable. A predictive validity test was performed on the XY axis to determine the validity of the results from the fsQCA.

5-1-Demographics Characteristics
This study received a total of 364 responses. In terms of gender, 64.6% of respondents were male. The age range of 45.9% was between 20 and 29 years old. The majority of respondents (56.3%) were single. The majority of respondents (66.2%) had undergraduate educational backgrounds. Most respondents were students (bachelor's and master's degrees), whereas 27.2% were government employees. Most respondents (54.7%) spent less than two hours a day using social media platforms (Instagram). In addition, 32.7% of respondents followed more than eight influencers on Instagram. Table 3 provides a detailed overview of the demographic characteristics of the respondents.

5-2-1-Validity and Reliability Test
Validity and reliability have been assessed at several stages throughout this study. To determine construct validity, this study compared each loading value obtained with a value of 0.7, as suggested in Hair et al. (2017) [63]. This indicates that construct validity has been achieved. Furthermore, convergent validity was tested using the Average Variants Extracted (AVE) value of 0.5 [63], and the obtained results were satisfactory. The internal consistency of a construct was evaluated using Cronbach's Alpha of 0.7 [63] and composite reliability of 0.7 [63]. These results indicate that the construct has strong internal consistency for each item, suggesting that internal consistency has been achieved. The results of construct validity, convergence, and internal consistency are presented in Table 4. The discriminant validity test was evaluated using three approaches. Firstly, this study compared the value of AVE square roots to the inter-construct correlation. It is known as the Fornell-Larcker criterion. The value of the square roots of AVE is greater than the value of the inter-construct correlation. Therefore, the Fornell-Larcker criterion approach demonstrated discriminant validity as suggested [64][65][66][67]. The discriminant validity results with the Fornell-Larcker criterion are shown in Table 5. Furthermore, discriminant validity was evaluated using the heterotrait-monotrait (HTMT) approach. According to Henseler et al. (2015) [67], HTMT is a new comprehensive approach for assessing discriminant validity measured by the HTMT value of 0.85. Results in Table 6 indicate that all HTMT values are less than 0.85. This suggests that this study had a strong discriminant validity category, as recommended [65].  Table 7 presents a matrix of cross-loadings. For a study to be considered a good discriminant valid from cross-loading criteria, the correlation coefficient of one construct must be greater than the factor loading item value of the other construct. The test results indicate that all constructs have significant factor loading than the other constructs' correlation coefficient. Therefore, each construct has good discriminant validity.

5-2-2-Hypothesis Results
This study used Smart PLS 3.0 software to perform a structural analysis. There are several steps in the evaluation of the structural model. The objective is to establish the strength of the research model used to test hypotheses. The path coefficient between constructs was first used to assess the model strength of endogenous variables. The feasibility of this approach was investigated by comparing the value of R 2 to 0.1. Falk & Miller (1992) [66] indicated that a structural model is considered viable when its R 2 value is greater than 0.1 or close to 1. According to the findings, the trust construct had an R 2 value of 0.634, which was explained by path coefficients of attractiveness, expertise, similarity, and parasocial interaction. Parasocial interaction had an R 2 value of 0.572 based on the path coefficients of attractiveness, expertise, and similarity. According to the trust path coefficient, the travel intention construct had an R 2 value of 0.465. The endogenous construct in this research model had an R 2 value greater than 0.1 [68], indicating that it is viable. Additionally, model fit was measured concerning the structural model. These results indicate that each of the model fit criteria, such as SRMR = 0.047; d_ULS = 0.942; d_G = 0.730; and NFI = 0.873, met the suggested requirements [63]. The results of structural modeling can be seen in Figure 3.

Figure 3. Structural Model Results
As illustrated in Figure 3 and Table 8, attractiveness and similarity significantly influence parasocial interaction. Therefore, H1a and H1c were supported (β = 0.258 and 0.499, t = 4.106 and 9.615, respectively). While expertise did not significantly affect parasocial interaction, H1b was not supported with β = 0.100, t = 1.822. The effect of expertise and similarity on trust was significant, supporting H2b and H2c (β = 0.296 and 0.211, t = 3.637 and 3.635, respectively). In contrast, the attractiveness effect on trust was insignificant, so reject H2a with -0.032, t = 0.462. Moreover, parasocial interaction had a significant effect on trust, which supported H3 (β = 0.437, t = 7.472). This study found that users who trusted SMI were likely to increase travel intention, indicating that H4 was supported (β = 0.682, t = 20.017).

5-3-1-Calibration Selection and Truth Table Construction
The present study used fsQCA analysis to develop configurations of high travel intention based on attractiveness, expertise, similarity, parasocial interaction, and trust. Additionally, fsQCA could assist this research in solving the complexity of the theory derived from the inter-dimensional relationships within the constructs associated with high travel intentions. Therefore, it is essential for researchers in consumer behaviors, management, or marketing to use fsQCA-based methods for formulating strategies in complex environments. To perform the fsQCA analysis, this study first calibrated the seven-point Likert scale in three groups, namely "6" as a full membership, "4" as the intersection, and "2" as a full non-membership based on calibration selection as suggested [67,68]. Afterward, the calibrated data were converted into fuzzy scores of "0", "Low," and "1", "High." Thus, a truth table was derived and shown in Table 9, revealing all-composite conditions that reflect high travel intention. The truth table analysis indicated fifteen composite indicators could predict travel intention, including eleven for those with "high" travel intentions and four for "low" travel intentions. In particular, the first composite that forms high travel intention was composed of 118 cases obtained from five independent constructs with the "High" condition. Currently, four composites with 30 instances were created based on the "High" condition on four different constructs that formed high travel intentions. For the state of high travel intention with three constructs with the state "High", there were 23 cases. Further, two composites with 7 cases formed high travel intentions from the two constructs with the state "High". The fifteenth composite indicator represented low travel intention, with the five constructs reflecting the "Low" condition for 45 cases. Furthermore, the fourteenth composite indicator represented low travel intention with three "Low" condition constructs containing eight instances. The outcome of low travel intention was formed from four constructs with a "Low" condition, which consisted of seven cases.

5-3-2-Analysis of Necessary Conditions
This study also examined the necessary conditions to determine whether certain constructs are always present or not. Table 10 describes the necessary conditions for high and low travel intentions. The construct is considered necessary if the consistency value is between 0.8 and 0.9 [69]. The results indicate that (1) the presence of all constructs (attractiveness, expertise, similarity, parasocial interaction, and trust) was necessary to shape the outcome of high travel intention and (2) the absence of similarity, parasocial interaction, and trust led to low travel intention.

5-3-3-fsQCA Findings
This study used fsQCA 3.0 software to perform data analysis. Table 11 presents the results of a fsQCA analysis obtained from intermediate solutions. The solution identified core and peripheral conditions that correlated with high travel intentions. There were six configurations for high travel intention solutions with an overall solution consistency and coverage values of 0.89 and 0.91, respectively. The results were acceptable and relevant based on the recommendation of Ragin (2006) [69], which suggested consistency and coverage values greater than 0.75. Furthermore, each causal condition from the first to the sixth configuration had a consistency value greater than 0.75, as recommended by Ragin (2006) [69]. Therefore, each configuration had a causal condition that ensured consistent high travel intentions. According to Ragin (2009) [70], the first (C.1) and sixth (C.6) configurations were the most relevant causal conditions that indicate high travel intention. Hence, each configuration was unique and different, supporting the research proposition. Each configuration for the high travel intention solution in Table 11 is unique. The following three configurations exist presence (*), absence (~), and don't care (blank space). The first configuration (C.  Figures 4 and 5 illustrate how fsQCA results with high consistency and relevance can be visualized to create high travel intentions. Based on the results of the fsQCA analysis, this study evaluated two configurations as the best solutions for forming high travel intentions based on consistency and raw coverage values. Firstly, this study found that attractiveness, expertise (social media influencers), and trust contribute to forming high travel intentions. Second, high expectations of travel can be created through social media influencers' attractiveness, expertise, and parasocial interaction.

5-3-4-Predictive Validity
A predictive validity approach was used in this study to test the ability of a research model and to predict the dependent construct in additional samples [67,71]. Predictive validity is fundamental for a model to make a good prediction [67]. It began by dividing the data into two subgroups of subsamples and hold samples. If the consistency between the two subgroups is high (> 0.75), a model can be considered fit and is of good predictive value. Table 12 provides the results of subsample analyses (*ATT, *EXP, ~SML, ~PSI, *TS). Furthermore, Figure 6 presents the analysis results of hold samples from prepositions (*ATT, *SML, *TS, ~EXP, ~PSI). Based on the results, the model calculated from the subsamples shows good predictive validity with the value of each causal condition greater than 0.75. The model calculated from hold samples also shows good predictive validity with a consistency of 0.95 and coverage of 0.337, indicating that the present study had good predictive validity.

6-1-Main Findings
This study aimed to understand the behavior of tourists who visited specific destinations via Instagram using the dimensions of social media influencer (attractiveness, expertise, and similarity), parasocial interaction, and trust. This dimension of social media influencer explains how attractiveness, expertise, and similarity impact trust and parasocial interactions. Parasocial interaction explains how trust may be impacted by perceived face-to-face illusion relationships with social media influencers. Embedding trust in the research model illustrates how consumers are in the affective stage of converting to purchasing behavior. Data were collected from the followers of SMI who have visited tourist destinations based on SMI's Instagram posts (videos, photos). The following are the conclusions of this study.
This study examined the effects of attractiveness, expertise, and similarity on parasocial interaction. Attractiveness reflects SMI's ability to influence consumers based on their physical appearance. Attractiveness (H1a) significantly impacts parasocial interactions. Accordingly, consumers tend to interact with highly attractive SMIs in a parasocial manner. Users form an emotional connection with SMI when they find SMI's Instagram posts (videos and photos) attractive, stylish, and charismatic. The results also confirm those of previous findings which suggested that SMI's attractiveness significantly increased parasocial interaction [72]. Expertise is the second dimension of SMI. Expertise was assessed by observing how SMI presented and explained information about destinations on their Instagram. Therefore, the expertise (H1b) did not significantly influence the parasocial interaction. This indicates that SMI's expertise in explaining information about destinations was not a significant factor in shaping parasocial interaction. Consumers who view SMI's posts on Instagram were not focused on the skill, expertise, or knowledge of explaining information about a specific destination. In addition, this finding was also confirmed by previous research, which suggested that the expertise of SMI did not significantly affect parasocial interactions [73]. The third dimension similarity refers to the similarity between consumers and SMI concerning their appearance, lifestyle, and thinking. The study results indicate that similarity (H1c) significantly affected parasocial interaction. The customers who perceive similarities with SMI were likely to form parasocial interactions. The higher the similarity, the greater the likelihood of developing a parasocial interaction. Previous research has also supported this outcome [74].
Second, all SMI dimensions were also evaluated against the concept of trust. It is expected to understand how these dimensions influence trust in this context. The results demonstrate that attractiveness (H2a) did not significantly affect consumer confidence. Thus, the image of an SMI who is charismatic, stylish, and attractive did not significantly impact consumer confidence. The results are also consistent with a previous study [75]. Meanwhile, expertise (H2b) and similarity (H2c) played a significant role in trust. This indicates that to increase trust, consumers were more likely to view SMI as a function of their ability, knowledge, and skill regarding a particular destination. In contrast, the similarity of lifestyle and thinking promoted trust in SMI. This result is consistent with previous studies [17,76].
Third, this research revealed that parasocial interaction (H3) significantly affected trust. This demonstrates that consumers perceived that parasocial interactions with SMI strongly influenced trust. Parasocial interaction leads to a greater sense of trust among consumers. In this case, parasocial interaction promoted consumers' entry into the affective phase, the trust stage. The parasocial interaction items are more likely to indicate an emotional connection between social media influencers and consumers. This may lead to the trust phase of the affective stage. Consumers who look forward to SMI posts feel friendly with them and rely on information provided by SMI. These items contribute to the building of trust. Moreover, this result is also supported by previous research [77]. Fourth, this study showed that trust (H4) was a significant predictor of travel intention. When SMI wishes to learn whether consumers intend to travel, they rely on it. The results also align with previous research [78].
Furthermore, the fsQCA explained how the theory represents the output of high travel intention using the predictive dimensions of SMI (attractiveness, expertise, and similarity), parasocial interaction, and trust. Based on the results of the fsQCA analysis, six configurations of causal conditions were associated with high travel intention. This study demonstrated how the causal condition is an antecedent to high travel intent derived from SMIs on Instagram. Overall, the consistency and raw coverage values were high, indicating that all solutions can be applied to form high travel intentions. However, this study suggested that two (C.1 and C.6) of the six configurations of the most relevant high travel intention solution were based on the highest raw coverage value. The first configuration mentioned the presence of the dimensions of SMI attractiveness, expertise, and trust to achieve high travel intention. While similarity and parasocial interaction are "don't care" conditions. On the other hand, the sixth configuration indicated that attractiveness, expertise, and parasocial interaction were conditions of SMI, while similarity and trust were characterized by "don't care" conditions. The fsQCA analysis of this study indicated that the formation of high travel intentions from Instagram can guide the configuration obtained, as each configuration is unique and different.

6-2-Implication for Research
This study provided significant academic contributions regarding the importance of social media in tourism marketing. In this case, social media influencers were considered an important aspect that should be investigated to develop tourism marketing knowledge through modern media, especially Instagram. This study has identified three dimensions of SMI on Instagram: attractiveness, expertise, and similarity. These were considered as adequate antecedents in tourism marketing and could increase travel intention. In this study, attractiveness is defined as the capability of SMI to attract followers through a stylish, charismatic, and attractive physical appearance [20]. Expertise has been described as the ability of SMI, measured according to his/her knowledge, experience, and skill in creating Instagram content [39]. Meanwhile, similarity refers to the feeling of similarity that followers experience with SMI, influencing their behaviors [45]. There are three dimensions to SMI, including the value of consumers who can develop a strong relationship with SMI through sharing tourism content on Instagram. This has been demonstrated by results that show the SMI dimension can influence the relationship between parasocial interaction and trust.
This study provided an in-depth analysis of parasocial interaction as a mechanism to increase tourists' trust in SMIs. The results confirm that attractiveness and similarity were the most critical factors shaping parasocial interactions with SMI. Nonetheless, expertise had no significant impact. SMI could increase consumers' trust by first building parasocial interactions through similarity and attraction. These two dimensions of SMI represented consumer perceptions of the emotional value associated with Instagram SMIs. This study has demonstrated that increased emotional value influenced the perception of parasocial interactions to create trust in SMI. Additionally, trust is a mechanism linking-social media influencers and travel intentions. The analysis results confirm that similarity and expertise were significant dimensions influencing trust. In particular, the dimension of expertise was more important in establishing trust in SMI than similarity. Therefore, functional value (expertise) was the most critical factor in forming trust. When the consumer is in the affective stage, similarity is a factor that can increase emotional purchase value. In this case, the functional value is more significant in trust, which is one of the consumer's effective conditions in the decision-making process. Travel intention results from the trust mechanism created. The trust mechanism is more predominant than the consumer's functional value.
Parasocial interaction refers to the ongoing experience of consumers connecting with media figures, such as social media influencers on Instagram. This conceptual model outlines how Instagram users look forward to attending events, following Instagram accounts, relying on information provided, and feeling emotionally attached to their favorite SMIs. SMI plays a significant role in forming parasocial interactions in the context of travel intentions. When consumers have significantly felt parasocial interaction, they will trust SMI. This is where trust plays an essential role in maintaining the parasocial interaction influencing travel intentions. For example, consumers who have followed, liked, wished to meet, or felt like friends with SMI are more likely to trust the destinations. The situation is similar to everyday consumers when building relationships with people, and trust is a vital component. Therefore, this study confirmed that trust was essential to realize the illusory relationship consumers feel about SMI into a new episode that results in behavioral investment that influences travel intentions in the future.
The theoretical contribution of the QCA configuration analysis can guide researchers in investigating travel behaviors through emerging technologies. The theoretical sophistication achieved in this study successfully determined the causal conditions that lead to high travel intentions. As a result of the configuration of the causal condition antecedent high travel intention used in this study, this study will help scholars comprehend consumer behaviors and emerging technologies. Accordingly, two main configurations explained the interdimensional interactions among social media influencers (attractiveness, expertise, and similarity) and parasocial interaction and trust toward high travel intentions. In the first configuration (C.1), conditions of attractiveness, expertise, and trust were combined to form a firm travel intention. As a result of this configuration, it can be seen that emotional value (attractiveness) and functional value (expertise) come together to create a high level of travel intention. Moreover, trust explains the affective stage of consumers in the decision-making process, which is a very significant factor. The combination of the two is the interaction of consumer values (emotional and functional) driven by the affective stage (trust) of decision-making is most likely converted to behavioral intention (travel intention). According to the second configuration (C.6), travel intention was achieved by combining attractiveness, expertise, and parasocial interaction. This configuration explains the interactions between value (emotional and functional) and the illusion of a direct consumer relationship with a high intention to travel formed by SMI. The value perceived by consumers concerning SMI on Instagram is attributable to parasocial interaction, which will significantly influence behaviors (travel intention).

6-3-Implication for Managerial
This study contributed to the management implications of using Instagram as an effective marketing tool for tourism destinations. Destination marketing using Instagram is a suitable alternative for tourism managers. It can reach target consumers effectively. Instagram is also becoming increasingly popular with its users. It is used as a medium that helps facilitate purchasing decisions. This study summarized a discussion of how Instagram was used to assist consumers in making travel decisions. Tourism managers will better understand consumer behavior and purchase decisions by investigating how social media influencers on Instagram can establish trust and parasocial interactions with followers. Tourism managers can employ the findings of this study as a marketing strategy and better understand consumer behaviors on Instagram.
First, social media influencers can enhance parasocial interactions. In social media, two dimensions can increase the parasocial interaction between followers and influencers: attractiveness and similarity. When consumers are watching, viewing, and observing the content posted by SMI on Instagram, they are experiencing parasocial interactions. Parasocial interactions increase due to the lifestyle, way of thinking, and physical attractiveness of SMI followers. The parasocial interaction that follows feelings towards SMI represents a virtual emotional connection that can influence followers and build a sense of closeness. To hire SMI for marketing, it is crucial to consider how to use SMI to create bonds with consumers. Tourism managers use Instagram to promote destinations, and they expect consumers to have a sense of proximity to the destinations. The SMI who share their travel activities on Instagram may influence their followers' travel intentions to the same destination.
Second, in establishing trust in SMIs, expertise is an essential factor. In the context of social media, trust is a critical component that can encourage consumer decisions. Therefore, the efforts of tourism managers to build consumer trust toward destinations on Instagram need to be identified. In this study, tourism managers were advised to use SMIs who were experienced, knowledgeable, skilled, and skilled at managing interesting content to influence their followers. Hence, when SMI shares its travel experiences on Instagram, it will tend to gain the trust of its followers. However, content concerning travel to specific destinations is also posted in photographs and videos, through which SMI can provide detailed information to its followers. This is confirmed in the research results as an increase in trust.
Thirdly, this study suggested that parasocial interaction with SMI could improve trust. When consumers feel close to SMI, their trust in SMI is increased. The tourism industry benefits from the parasocial interaction that followers have with SMI. The close relationship between SMI and its followers is one of the emotional factors that can influence behavior. As a result, when SMI shares their travel experiences on Instagram, their followers trust them. Tourism managers can benefit from SMI as a marketing tool to develop a sense of proximity, increase consumer awareness of the destination, and strengthen consumer trust. The study also shows that consumers tend to increase their travel intentions when they trust SMI. Therefore, trust is a significant factor in influencing travel intentions to promote destinations communicated to consumers through Instagram.
Lastly, the results obtained through fsQCA broaden the perspectives of tourism managers on how to increase visits to destinations with social media influencers on Instagram. A fsQCA configuration comprises different and unique causal conditions, predictive dimensions of SMI (attractiveness, expertise, and similarity), parasocial interaction, and trust. Therefore, each fsQCA configuration of this study can be used by tourism managers to determine high travel intentions. The causal conditions (C.1) that can result in a high travel intention are the presence of conditions, attractiveness, expertise, and trust. Tourism managers can benefit from this insight because it shows that attractive and expert SMI is needed to create high travel intentions to promote destinations. Hence, when tourism managers decide to use SMI as a marketing communication tool, they should consider configuration (C.1) to consider how consumers' interaction of emotional and functional aspects can create trust. Meanwhile, other configurations (C.6) that can result in high travel intention output include causal conditions, present conditions of attractiveness, expertise, and parasocial interaction. This provided insight for tourism managers that consumers must interact with SMI in a parasocial way determined by the dimensions of attractiveness and expertise to form strong travel intentions. The overall solution configuration generated by fsQCA is validated through predictive validity. The model derived from each configuration of the increased travel intention solution can be implemented.

7-Conclusion
This study successfully synthesized how Instagram influencers can influence consumer travel behavior. The data in this study were analyzed using two approaches, namely SEM and fsQCA. As a result of the SEM analysis, consumers were more likely to trust SMI when promoting destinations because of their expertise and similarity. In addition, this study demonstrated that highly attractive SMI and similarity with followers could lead to parasocial interactions. Travel decision by consumers was influenced if they had parasocial interaction with SMI. The FsQCA results indicate two configurations with a high level of travel intention. This study's causal conditions configuration illustrated the interdimensional relationship between SMIs (attractiveness, similarity, and expertise), trust, and parasocial interaction.
There are some limitations to this research. First, this study examined three dimensions of SMI, namely attractiveness, expertise, and similarity, but did not identify a particular SMI. For instance, Shen et al. (2019) [79] classified attractiveness into task, social, and physical. Therefore, further research could identify the dimensions of attractiveness, expertise, and similarity to investigate more specific aspects of SMI. This is because social media platforms have different characteristics. Hence, SMI follows a different approach to reporting on its followers' experiences. Secondly, this study examined the implications of SMI and parasocial interaction on trust without examining consumer value. Additionally, one factor contributing to trust is consumer value [80]. Further research can identify consumer values (hedonic, utilitarian) when using Instagram and SMI to make decisions. This will allow the continuous development of research models for shaping travel intention and destination loyalty according to consumer value, trust, and other factors.

8-2-Data Availability Statement
The data presented in this study are available on request from the corresponding author.

8-3-Funding
The authors received no financial support for the research, authorship, and/or publication of this article.

8-6-Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this manuscript. In addition, the ethical issues, including plagiarism, informed consent, misconduct, data fabrication and/or falsification, double publication and/or submission, and redundancies have been completely observed by the authors.