A Source Effect Theory Perspective on How Opinion Leadership, Parasocial Relationship, and Credibility Influencers Affect Purchase Intention

Objective: Influencers are known to have a solid influence on generating purchase intentions in consumers. Therefore, this study aims to examine the effect of opinion leadership, parasocial relationships, and credibility identified through the variables of attractiveness, trust, and expertise on purchase intentions. Design/Methods/Approach: 206 individuals who follow food vlogger influencers were surveyed for this study. The data analysis was carried out using GSCA SEM with GSCA pro software, and the mediation test was conducted through the Sobel test. Findings: The results indicate that opinion leadership and parasocial relationships have a positive impact on purchase intention. In addition, attractiveness and trustworthiness can also influence purchase intention through parasocial relationships. However, it was observed that the expertise variable did not directly or indirectly affect purchase intention. Originality: In this study, variables are utilized based on source effect theory to facilitate the contribution of the results to the advancement of marketing theory and science, specifically in the field of influencer marketing. Practical/Policy implication : For marketers, selecting influencers who wield substantial influence over their audience and have forged robust relationships with them is crucial. These influencers are renowned for their opinion leadership and parasocial connections, making them ideal for promoting products or brands.


Introduction
Social media is widely recognized for its immense potential in the marketing field. This is due to the staggering number of users worldwide, which currently stands at 4.74 billion, or 59.3% of the global population (datareportal, 2023). In Indonesia, for instance, YouTube boasts 139 billion users, or 50% of the country's population in 2022 (datareportal, 2022). The growing user base translates to increased interaction among users, which allows them to impact one another's purchasing decisions (Ennaji et al., 2018). As a result, marketers have come to see the relationships between users as increasingly vital (Yuan et al., 2016). These interactions also create influential users who are trusted by others, leading to the rise of the term influencer.
The emergence of social media has altered the role of influencers. Consumers have come to trust the content produced by influencers (Aw et al., 2023), making them significant in influencing consumer behavior (Sokolova & Kefi, 2020). Marketers have responded by emphasizing the importance of influencers in their strategies (Aw et al., 2023), including using them to promote products and services (Aw & Chuah, 2021). This has led to the development of influencer marketing, which involves creating a connection between influencers and consumers in a social media community (Farivar & Wang, 2022).
Individuals known as influencers possess valuable knowledge in their respective fields. Companies often collaborate with influencers to review products before releasing them to the market. This partnership allows influencers to express their opinions, contribute to product improvements, and share their perspectives on social media (Romero-Rodríguez et al., 2020). In addition to their expertise, influencers' credibility has a significant impact on consumers. Consequently, many brands use influencers to promote their products (Belanche et al., 2021). Influencers are highly effective in boosting brand awareness and encouraging purchasing decisions. The influence of YouTube influencers has been a topic of interest in the marketing literature (Sokolova & Perez, 2021).
Many studies have explored the various factors that enable influencers to sway consumers into making purchases or generating purchase intentions. One of these factors is opinion leadership. Influencers can be opinion drivers for consumers, thus making them more likely to follow the influencer's recommendations (Casaló et al., 2020). Furthermore, Thanh Ha & Thu (2022) explain that on social media, influencers can provide a positive social word of mouth (sWOM) if they have a pleasant experience with consumers. Consumers are drawn to original and unique accounts, leading to increased interaction, recommendations, and a willingness to follow the influencers' advice (Casaló et al., 2020). Opinions from influencers can ultimately lead to an intention to purchase products they recommend (Farivar et al., 2021).
One important aspect to consider is the parasocial relationship between consumers and influencers. This attachment can influence the consumer's intention to make a purchase based on the influencer's recommendations (Farivar et al., 2021). If the bond between the two is strong, consumers are more likely to buy the products suggested by the influencer (Sokolova & Kefi, 2020). Opinion leadership and parasocial relationships play different roles in shaping consumer purchase intentions. Opinion leadership refers to the influencer's expertise and knowledge about the product being reviewed. If an influencer recommends a product they like, it does not necessarily impact the consumers' opinions.
Developing a parasocial relationship can create a personal connection between influencers and consumers, ultimately leading to increased preference for the influencer (Farivar et al., 2021). With continued improvement in this relationship, consumers may become more engaged with and loyal to a brand (Zhong et al., 2021). Youtube is a particularly effective platform for building parasocial relationships (de Bérail et al., 2019), leading to increased trust from consumers (Fazli-Salehi et al., 2022) and improved credibility for the influencer. To establish this credibility, influencers should thoroughly understand the products they are reviewing, which can encourage consumers to follow their accounts, seek their advice, and recommend them to others (Belanche et al., 2021).
Despite previous research, there are still discrepancies in the results. Akdevelioglu & Kara (2020) found no correlation between opinion leadership on social media and the desire to use new products. Al-Harbi & Badawi (2022) discovered that opinion leadership does not affect the purchase intention of organic food. In contrast, Farivar et al. (2021) proved that opinion leadership is a significant factor in influencing purchase intention. Aw et al. (2023) demonstrated that there is no relationship between attractiveness and parasocial relationships. Additionally, Rungruangjit (2022) found that neither physical attractiveness nor trustworthiness affects purchase intention. However, Sokolova & Perez (2021) provided empirical evidence indicating that attractiveness can indirectly influence high parasocial relationships, thereby directing consumers' purchase intentions. Furthermore, Masuda et al. (2022) suggest that trustworthiness does influence purchase intention. These varying findings highlight the need for further research to explain this phenomenon.
While previous research has examined the impact of individual factors on consumer behavior, there is a need for a deeper understanding of the topic. Specifically, there has been limited research on the simultaneous influence of multiple factors, such as opinion leadership and parasocial relationships, on consumer purchase intentions (Farivar et al., 2021). This study uses the source effect theory to explore the role of influencers in increasing consumer purchase intentions. Additionally, this study considers the impact of opinion leadership and parasocial relationships, a variable that has not been widely tested in previous research. Other variables, including attractiveness, trustworthiness, and expertise, are also examined based on the source effect theory. These variables are expected to provide a more comprehensive understanding of the impact of influencers on consumers.
This study aims to uncover what attributes enable influencers to influence consumer purchase intentions. This will be determined by examining features such as opinion leadership, parasocial relationships, attractiveness, trustworthiness, and expertise. This research has made a valuable contribution to the marketing literature. First, this research aims to expand the application of the source effect theory by examining the variables of opinion leadership, parasocial relationships, and credibility through attractiveness, trustworthiness, and expertise. Second, this study adds to existing research on the influence of opinion leadership and parasocial relationships on purchase intention, which has yet to be discussed in previous studies. Third, this research strengthens the attributes influencing consumers' purchase intentions. Fourth, this study was conducted on a platform that has rarely been researched before -mega influencers in the food vlogger category on YouTube. Previous research only discussed influencers using the Instagram platform in the fashion category. Mega influencers were chosen for this study as little research has been conducted on them. Mega influencers are influencers with more than two million subscribers (Conde & Casais, 2023). The differences in social media from previous research and different types of influencers can provide a more general picture of the influence of influencer attributes that can affect consumer purchase intentions.
This research is divided into multiple sections. Section 1 provides the background information, section 2 includes a literature review and hypothesis preparation, section 3 outlines the data collection methods, section 4 covers the results and discussions, and Section 5 presents the conclusions with academic implications and practical applications.

Source Effect Theory
Studies have examined how the characteristics of information sources affect communication effectiveness and message reception (Farivar et al., 2021). These sources may include expertise, level of trust, attractiveness, and more (Kang & Herr, 2006). Of all the sources, expertise is believed to have the most significant impact. One's expertise can be determined by their level of education and experience (Wilson & Sherrell, 1993). Furthermore, Chen et al. (2021) discovered that expertise can enhance parasocial relationships and closeness between consumers and influencers.
An influencer should promote a product based on their expertise to establish trust. Trust is more likely to be earned if there is a fit between the product and the influencer's capabilities (Belanche et al., 2021). Mutual trust among social media users is the foundation of this influence (Ennaji et al., 2018). Jin et al. (2021) suggest that uploading images of influencers and products on feeds effectively increases consumer trust. These images create a parasocial interaction and social presence that fosters trust. Consumers who experience this interaction are more likely to trust the influencer, which ultimately leads to brand trust (Fazli-Salehi et al., 2022). Continued improvement of this parasocial relationship can lead to high brand engagement and consumer loyalty (Zhong et al., 2021).
According to Kang & Herr (2006), attractiveness significantly influences consumers' purchasing decisions. Attractiveness can be divided into two categories: physical attractiveness and social attractiveness, each with its impact. Physical attractiveness affects influencer credibility, while social attractiveness affects parasocial interaction (Sokolova & Kefi, 2020). As a result, Kok Wei & Li (2013) recommend that brands use physically attractive celebrities to increase consumer motivation.

Opinion Leadership on Purchase Intention
According to a study by Song et al. (2017), opinion leadership holds a significant influence over society. This is because opinion leaders serve as a valuable source of advice for consumers, as noted by Casaló et al. (2020). When influencers possess strong opinion leadership, consumers perceive them as having trustworthy information, as observed by López et al. (2022). These influencers are known for their expertise in their respective fields and willingness to share their knowledge on social media platforms, as confirmed by Romero-Rodríguez et al. (2020). To attain the status of an opinion leader, one must create unique and original content, as emphasized by Casaló et al. (2020). Opinion leaders are also mindful of their reputation and take steps to preserve it, as they do not want to tarnish their image with consumers.
In general, influencers tend to be less interested in promoting products in exchange for compensation (López et al., 2022). Influencers who possess strong opinion leadership skills can motivate consumers to engage with them and recommend their content (Casaló et al., 2020). Opinion leadership is particularly impactful in small groups with close interactions (Moldovan et al., 2017). On social media platforms like Instagram, influencers with strong opinion leadership are highly regarded by consumers (Casaló et al., 2020). Nowadays, consumers are increasingly turning to social media influencers who are perceived as opinion leaders and trust their content (Aw et al., 2023).
Individuals who possess opinion leadership as influencers on Instagram have numerous advantages, such as the ability to boost consumer engagement and recommend accounts to others. Consumers become active participants in the value-creation process, sharing their knowledge by interacting on accounts and bringing in new customers through word-of-mouth referrals. This can significantly benefit influencers by increasing their reach and influence. Additionally, opinion leadership can lead to a greater likelihood that consumers will consider the suggestions posted on the account, ultimately impacting sales and the company's success (Casaló et al., 2020). Yuan et al. (2022) further noted that strong opinion leadership can also increase social media traffic. Previous research supports the notion that opinion leadership can also influence purchase intention (Farivar et al., 2021;Farivar & Wang, 2022;Romero-Rodríguez et al., 2020). Therefore, it can be concluded as follows: H1: Opinion leadership has a positive effect on purchase intention

Parasocial Relationship on Purchase Intention
A parasocial relationship is a connection between an actor or influencer and their audience . When it comes to social media it refers to a perceived relationship between consumers and influencers that may not actually exist (Farivar et al., 2021). Social media has become a powerful tool for fostering these relationships in today's society (Aw & Labrecque, 2020). However, these parasocial relationships can sometimes lead to addiction rather than genuine friendship (Sokolova & Kefi, 2020). Consumers often desire a direct, two-way interaction with their favorite influencers through social media, hoping to feel affiliated, heard, and responded to by those they admire (Aw et al., 2023).
Social media has made it easier for influencers to create strong parasocial connections. The strength of these connections is determined by four dimensions: affective, behavioral, cognitive, and decision. The affective dimension refers to people's feelings, the behavioral dimension refers to their actions, the cognitive dimension refers to their thoughts, and social media influence the decision dimension. These dimensions are interconnected and must be considered, as per Garcia et al. (2022). Aw & Chuah (2021) argue that the parasocial relationship can become even stronger if an influencer creates compelling content and demonstrates expertise.
According to Conde & Casais (2023), a strong relationship between influencers and consumers can be a powerful, persuasive tool, alongside strong opinion leadership, in influencing consumer behavior. In particular, parasocial relationships play a significant role in influencing consumer behavior, particularly purchase intention, as noted by Masuda et al. (2022). This suggests that the closer the relationship between influencers and consumers, the higher the likelihood of purchase intention. Fazli-Salehi et al. (2022) further explain that consumers who develop parasocial relationships with influencers tend to have a higher intention to purchase products recommended by these influencers. Previous studies, including those by Sokolova & Kefi (2020) 2022), have also demonstrated that purchase intention is influenced by parasocial relationships. Therefore, it can be concluded that: H2: Parasocial Relationship has a positive effect on purchase intention

Credibility on Parasocial Relationship and Purchase Intention
According to Shimp (2014), prior research has established that credibility and attractiveness play a significant role in the effectiveness of influencers. The TEARS model was developed to describe the influencer's capabilities, with TEARS representing five key attributes. Trustworthiness and expertise are dimensions of credibility, while physical attractiveness, respect, and similarity are the dimensions of attractiveness. Credibility pertains to an individual's tendency to believe in someone. When information is deemed credible, the consumer's attitudes can be altered through a psychological process known as internalization. This process occurs when the consumer aligns themselves with the influencer's stance on a particular issue. Trust and honesty are two critical sub-attributes of credibility, with trust referring to the source's honesty, integrity, and trustworthiness.
To gain the trust of consumers, influencers need to demonstrate their objectivity and avoid any attempts at manipulation. Building credibility is crucial, which can be achieved by showcasing expertise in a particular field. When influencers are seen as authorities in their area of interest, their opinions can sway consumer sentiment. Furthermore, consumers are more likely to trust an influencer's review if they have experience and knowledge of the product being discussed (Casaló et al., 2020). However, an influencer's credibility can be compromised if they appear to be promoting a product for financial gain rather than genuine interest (Belanche et al., 2021).
It is crucial to note the impact of credibility on consumer behavior, as previous studies have shown that it can increase purchase intentions (Sokolova & Kefi, 2020;Balaban et al., 2022). Kok Wei & Li (2013) also pointed out that trust plays a significant role in shaping consumer intentions. In addition, trustworthiness and expertise have been found to influence purchase intention (Masuda et al., 2022;Rungruangjit, 2022). Therefore, it is essential to consider these factors when seeking to influence consumer behavior.
The next attribute to consider is attractiveness. According to Kang & Herr (2006), this is a crucial factor in source-effect communication and can significantly influence consumer decisions. Attractiveness is divided into three categories in the literature: physical attractiveness, social attractiveness, and source attractiveness (Kok Wei & Li, 2013;Aw et al., 2023;Sokolova & Kefi, 2020). When consumers find something attractive in an influencer, they tend to identify with them, which can lead to persuasion. Consumers may adopt the influencer's attitudes, behaviors, interests, or preferences. Physical attractiveness, respect, and similarity are all key attributes in the TEARS model. Previous research has shown that physical attractiveness is particularly important, as good advertisements and brands are often supported by good physical attractiveness. Kok Wei & Li, 2013 suggest that brands can endorse celebrities with physical attractiveness to increase consumer motivation. Therefore, it can be concluded that: H3a: Attractiveness has a positive effect on purchase intention H3b: Attractiveness has a positive effect on parasocial relationships H4a: Trustworthiness has a positive effect on purchase intention H4b: Trustworthiness has a positive effect on parasocial relationships H5a: Expertise has a positive effect on purchase intention H5b: Expertise has a positive effect on parasocial relationships

Parasocial Relationship Mediating the Effect of Attractiveness, Trustworthy and Expertise on Purchase Intention
Effective communication relies on the information source's expertise, trustworthiness, and attractiveness, as supported by research conducted by Kang & Herr (2006) and Yuan et al. (2016). Expertise is particularly important in fostering parasocial influencer relationships that encourage consumer purchase intentions, as noted by Aw et al. (2023) and Chen et al. (2021). Additionally, both physical and social attractiveness can also impact parasocial relationships, highlighting the importance of both credibility and appeal (Masuda et al. (2022); Sokolova & Perez (2021)). Ultimately, the strength of parasocial relationships can influence purchase intentions, as found in studies by Masuda et al. (2022), Balaban et al. (2022), and Rungruangjit (2022). Therefore, it can be concluded that: H6: Parasocial relationship mediates the effect of attractiveness, expertise, and trustworthiness on purchase intention

Method
This research analyzes the followers of food bloggers Tanboy Kun and Farida Nurhan, who have over 2 million subscribers. The influencers were chosen based on their attractiveness to the studied variables. A questionnaire was used to collect data, and it was distributed online to 250 respondents who met the criteria of following both influencers. The study includes factors such as gender, age, education, length of time using YouTube, and following other influencer accounts, as done by Farivar et al. (2021). The indicators used in this study can be viewed in Table 1. The questionnaire was distributed over three months, from March to May, with 250 responses received. The reverse statement sentence was used to determine the seriousness of the respondents in answering the questionnaire. A total of 44 responses were disqualified, leaving 206 responses for analysis.
In data testing, Generalized Structured Component Analysis (GSCA) is employed. GSCA is best applied to structural models with a solid theoretical basis, and its primary function is to create a robust structural model for predictive purposes. To conduct GSCA testing, the SEM GSCA pro software was used in accordance with Ngatno's (2019) guidelines. The first tests performed were validity and reliability. The validity test in this study involved two factors. Firstly, the loading estimate (factor loadings) was used to test the convergent validity of a construct formed with reflective indicators. The loading estimate value is considered valid if each indicator is above 0.5 (Ghozali, 2008). Two Average Variance Extracted (AVE) values were used to test the discriminant validity of a construct. The required AVE value is 0.5 (Ngatno, 2019).
The reliability of the study is determined by alpha values, which are identical to the Cronbach alpha measurement (Ngatno, 2019). To ensure reliability, the alpha value must be at least 0.6 (Sugiyono, 2006). As part of the pilot study, 50 data sets were tested for validity and reliability using the procedure outlined by Farivar et al. (2021). The next step will be to test the structural model and establish a causal relationship between the variables, validating the proposed hypothesis. The t-statistic test will be used to assess accuracy, with a 95% confidence level, a commonly used approach by other researchers (Ngatno, 2019). Mediation testing will be performed with the Sobel test, comparing the z value to a significance level of 5% or 1.96. If the z value exceeds the significance level, then mediation is present, and vice versa. Additionally, this research also examined the suitability of the model through predetermined indicators in GSCA. The fit model comprises FITs, AFITs, FITs, FITm, GFI, SRMR, OPE, OPEs, and OPEm. The FIT value represents all variables (indicators and components) described in the research model, with a higher value closer to 1 indicating a better explanation of phenomena. AFIT is comparable to the adjusted R-Square in linear regression. FITs indicate the total variance of all components described by the structural model specifications, while FITm denotes the total variance of all components explained by a specific model specification's measurement. Higher values for both FITs and FITm indicate an improved model (Hwang & Choo, 2021). There were 206 respondents in this study, and the required GFI and SRMR values were GFI ≥ 0.93 and SRMR ≤ 0.85 (Cho et al., 2020). OPE demonstrates some ability to predict unseen observations (Hwang & Choo, 2021).

Result
The study's validity and reliability test results are illustrated in Tables 3, 4, and 5. Table 3 displays each indicator's estimated value over 0.5, implying that the indicators utilized in this research have good convergent validity. Table 4 showcases the AVE value of each variable, which surpasses 0.5, demonstrating that the variables used in this study have good discriminant validity. Additionally, Table 5 presents the alpha value of each variable, which is above 0.6, meeting Ngatno's (2019) criteria, ensuring that the constructs in this research are reliable. The outcomes of this analysis align with the tests conducted by Farivar et al. (2021) and Rungruangjit (2022), where the validity value exceeds 0.5 and the reliability is above 0.6. After building the model, the next step was to test it. Table 6 displays the results of this test. The FIT value of 0.634 indicates that the variables used can explain 63.4% of the model formed. Specifically, the research model can explain 63.4% of opinion leadership, parasocial relationships, attractiveness, trustworthiness, expertise, and purchase intention, while other variables explain the remaining 36.6%. This FIT value is close to 1, which is considered suitable for this study (Hwang & Choo, 2021). The AFIT value of 0.63 also suggests that the research model is good. It can explain 63% of the diversity of research variables, with the remaining 37% explained by other variables. The FITs and FITm values also show a value close to 1, indicating they are good. The GFI value of 0.992 and the SRMR value of 0.051 meet the criteria set by Cho et al. (2020). Therefore, the GFI and SRMR values in this study are satisfactory.
Structural model evaluation was carried out to confirm the proposed hypothesis, as presented in Table 7. To determine the significance of a hypothesis with a confidence level of 95, we use the indicator table r, also known as the norm distribution table. The r value for this study was set at 0.138, based on the criteria used by 206 respondents and a significance level of 5%. If the estimated value for each path is higher than the r value, then the hypothesis criteria are considered accepted.
The results of this study reveal that Leadership Opinion has an estimated value of 0.342, which is higher than 0.138. Thus, Opinion Leadership has a significant effect on purchase intention, and hypothesis 1 is accepted. Likewise, the estimated value of the parasocial relationship on purchase intention is 0.491, higher than 0.138. Therefore, hypothesis 2 is accepted. Based on the estimated values, it can be concluded that hypothesis 3b regarding the attractiveness of parasocial relationships is accepted, with a value of 0.543, which is greater than 0.138. However, hypothesis 3a regarding the attractiveness of purchase intention is rejected, as its estimated value is only 0.056, less than 0.138. Regarding the trustworthiness of parasocial relationships, hypothesis 4b is accepted with an estimated value of 0.273, greater than 0.138. However, hypothesis 4a is rejected since the trustworthiness of purchase intention has an estimated value of -0.088, which is smaller than 0.138. Both hypotheses 5a and 5b regarding the expertise on purchase intention and parasocial relationships are rejected. The estimated values for both are less than 0.138, with a value of 0.038 for purchase intention and 0.077 for parasocial relationships.
The mediation test was conducted using the Sobel test, and the results indicated that the variable of interestingness on purchase intention was partially mediated by parasocial relationships, with a z value of 3.58 > 1.96. Similarly, the mediation test for trustworthiness on purchase intentions showed that parasocial relationships partially mediate the relationship between attractiveness and purchase intention, with a z value of 2.59 > 1.96. However, the mediation skills test on purchase intention mediated by parasocial relationships yielded a z value of 0.93 < 1.96, indicating that parasocial relationships do not mediate expertise and purchase intentions. The statistical model results are presented in Figure 2.

Discussion
The purpose of this research is to focus on influencer marketing. The study aims to provide a better understanding of the various influencer attributes that can impact purchase intentions. Additionally, it contributes to the development of the source effect theory by combining opinion leadership, parasocial relationships, and credibility factors, such as attractiveness, trustworthiness, and expertise.
The research aims to understand how influencers' attributes affect people's intention to buy products. The research shows that influencers who are considered opinion leaders can have a significant impact on people's buying decisions. Consumers often trust the opinions of influencers and use them to decide whether to buy a product. High opinion leadership is advantageous because it sets influencers apart from others and makes them role models for consumers. This aligns with previous studies by Romero-Rodríguez et al. (2020) and Farivar & Wang (2022), which suggest that opinion leadership has the power to influence people's buying decisions.
The findings show that purchase intention is influenced by parasocial relationships. Social media is crucial in enabling influencers and consumers to communicate, establishing an imaginary bond that can drive consumer purchase intentions over time. This aligns with Conde & Casais's (2023) research which suggests that a strong relationship can significantly influence consumers. Table 2 reveals that the comfort indicator has the highest value among the four indicators. In other words, a parasocial relationship can make consumers feel at ease, leading to a strong interaction that ultimately results in purchase intentions. These results are consistent with Masuda et al. (2022) proposition that parasocial relationships impact purchase intentions.
This study has found that both opinion leadership and parasocial relationships have an impact on purchase intention. This means that influencers who possess high levels of opinion leadership and parasocial relationships can direct consumers toward generating a purchase intention. This is aligned with the findings of recent research conducted by Farivar et al. (2021), which highlights the importance of opinion leadership and parasocial relationships in influencing consumers. According to the estimated results in Table 6, parasocial relationships hold a higher estimated value when both opinion leadership and parasocial relationships are tested together. This indicates that although opinion leadership has an impact on purchase intention, consumers pay more attention to the closeness of their relationship with influencers. This closeness further encourages consumers to increase their purchase intention. Once again, these findings are in agreement with research conducted by Farivar et al. (2021), which suggests that parasocial relationships have a more significant impact on purchase intention than opinion leadership.
It is worth noting that attractiveness and trustworthiness impact purchase intention differently. Surprisingly, attractiveness does not have a significant effect on purchase intention. This finding is interesting because previous research has suggested a strong correlation between attractiveness and purchase intention. However, attractive influencers are not always the driving force behind purchase decisions. Consumers do not seem to view the attractiveness of food vlogger influencers as a compelling reason to make a purchase. Other factors, such as the quality of the product being reviewed, its location, or the cleanliness of the place, play a more significant role in influencing purchase intention. These results are consistent with a study conducted by Rungruangjit (2022), which indicates that attractiveness is not a primary consideration for consumers when making purchasing decisions.
The impact of attractiveness on purchase intention is not direct but rather through the development of parasocial relationships. This finding supports Aw & Chuah's (2021) argument that consumers are influenced by the long-term relationship built with the influencer, not just their physical attractiveness. Relying solely on physical attractiveness is not enough to persuade consumers to buy. Trustworthiness does not have a direct impact on purchase intention, which aligns with Rungruangjit's (2022) research. This could be due to consumers perceiving paid endorsements as less trustworthy. Taste and other factors also play a role in consumer trust. However, the study showed that parasocial relationships mediate the trustworthiness-purchase intention relationship, indicating that building a strong relationship with the influencer can lead to purchase intention, even if trust is not established directly. Expertise does not have a direct or indirect influence on purchase intention, according to  research. Consumers prioritize the character and charisma of influencers over their expertise, and being a food vlogger does not necessarily require high levels of expertise.

Conclusion
This study analyzed the impact of factors such as purchase intention, parasocial relationship, attractiveness, trustworthiness, and expertise on purchase intention. The findings revealed that opinion leadership and parasocial relationships have a positive impact on purchase intention. Attractiveness and trustworthiness, on the other hand, affect purchase intention indirectly through parasocial relationships. However, expertise did not have any direct or indirect effect on purchase intention. These results shed light on the influence of different attributes of influencers, particularly food vloggers, on consumers' buying decisions. Moreover, they contribute to the advancement of the source effect theory.
Marketers should take note of opinion leadership and parasocial relationships when working with influencers in promoting a product or brand. While influencers are currently at the forefront of marketing, not all can influence consumer purchasing decisions. To determine the effectiveness of an influencer, marketers should assess the trustworthiness of their suggestions or opinions, whether they serve as role models, and whether they stand out from other influencers. The presence of good opinion leadership skills and parasocial relationships are crucial factors to consider in identifying influencers who have the potential to be effective.
Another critical factor to consider is the parasocial relationship between consumers and influencers. Studies have shown that this relationship significantly affects purchasing decisions. The closer the relationship between the consumer and the influencer, the more comfortable the consumer feels, often considering the influencer as a close friend. This can benefit marketers as strong relationships can easily influence consumers to increase their intention to purchase. Additionally, the attractiveness and trustworthiness of influencers should also be considered when selecting whom to work with. While not a direct influence, an influencer who is physically attractive, persuasive, trusted, and has a strong connection with the consumer, can greatly impact purchase intention. This research has some limitations. This research only focuses on variables such as opinion leadership, parasocial relationships, attractiveness, trustworthiness, and expertise in purchase intention among food influencer vloggers on YouTube. In addition, the number of respondents in this study is still limited. To improve future research, future research is recommended to add other variables such as similarity, power, source visibility, identity, and influencer characteristics, which have been used in previous research based on source effect theory. Including these variables will help clarify the relationship between variables in source effects theory. Furthermore, future research should examine influencers on other objects, such as technology influencers, on different platforms. Finally, it is recommended to increase the number of research samples to get more accurate results.

Author Contribution
Author 1: Conceptualization, writing original draft, data curation, formal analysis, investigation, methodology