The Effect of eWOM Source on Purchase Intention: The Moderation Role of Weak-Tie eWOM

: Scholars around the world have studied electronic word of mouth (eWOM) heavily in the past decade. However, despite extensive knowledge of the effect of eWOM on consumer behavior, there is still a lack of understanding of how the eWOM source shapes those behaviors. There are many sources of eWOM, the inﬂuence of which may have different outcomes when it comes to consumers’ purchase intentions. Moreover, most studies have measured social tie strengths (strong/weak) as one variable. Weak-tie eWOM (W-tie) and strong-tie eWOM (S-tie) are always studied in contrast to each other. The versus characteristics of these two constructs have been considered in almost all studies; however, they may work together via different mechanisms. This study split social ties into two separate constructs, W-tie and S-tie, while adding another source of eWOM (celebrities eWOM), and investigated how W-tie moderated the effect of S-tie and celebrity eWOM on purchase intention, as well as the mediation effect of perceived value/perceived enjoyment on purchase intention. A total of 726 valid responses were analyzed using SPSS and AMOS. The results showed that all direct relationships in the model were supported. Moreover, perceived value mediated only the effect of celebrity eWOM on purchase intentions and did not successfully convey the effect of S-tie on purchase intentions. On the other hand, perceived enjoyment mediated both relationships. W-tie did not moderate the relationship between S-tie and perceived value, perceived enjoyment, or purchase intention. In contrast, W-tie moderated the positive relationship of celebrity eWOM with perceived value, perceived enjoyment, and purchase intention, while it also moderated the positive relationship of perceived value and perceived enjoyment with purchase intention. This research illuminates how different eWOM sources and social ties on social media platforms affect purchase intentions.


Introduction
As a result of the wide usage of social media, there is now a de facto omnipresence of material that has been developed and distributed by consumers [1], which has led to new communication phenomena [2]. People share their knowledge and experiences on social media sites. It is interesting to note that scientific research suggests that customers are inclined to have unconditional trust in material posted on social media [3]. On the other hand, in addition to the content that companies provide, there are a number of other sources, such as weak and strong ties, offering perspectives and information on goods and services available on social networking sites. Additionally, in today's society, celebrities and people who wield great power on social media play a significant role in determining how their followers feel about certain products.
There is nothing novel about the concept of using opinion leaders as a means of exerting influence over the decisions that customers make. People who have a stronger

Strong-Tie eWOM
The term electronic word of mouth refers to favorable or unfavorable online messages/posts about goods or services created or communicated by consumers and made accessible to a large number of people utilizing the internet [32]. The development of technology has brought about substantial shifts in the manner in which information is disseminated and has enabled it to transcend the traditional restrictions of word of mouth [33]. Consumers currently express their purchase experience with brands on social media [34], using a variety of channels available on the internet, including e-mail, chat rooms, online forums, community groups, and brand-specific communities [35]. eWOM is the term given to describe the phenomenon that has arisen due to the proliferation of word-of-mouth marketing due to the internet. This development removes the need for people to be physically present [36]. In addition, there is a substantial volume of online eWOM content online [37]. This does not mean that all eWOM information will influence consumer purchasing equally; there is likely to be some variance in the extent of the effect.
The term "influential marketing tool" has been attached to eWOM for quite some time [38]. Prior to making purchases of goods and services, customers are more likely to look for or investigate information that has already been made accessible to them. Customers do so to feel confident when making decisions on their purchases [39].
According to [40], those in a situation of strong ties are more inclined to interact on a more regular basis and put out more effort to preserve their relationships. People are encouraged to share information with one another via a range of communication platforms and to do so at a high frequency to foster the development of strong links built on deep personal relationships. Moreover, according to [41], strong ties are characterized by increased degrees of closeness, increased levels of self-disclosure, emotional and instrumental exchanges, reciprocity in exchanges, and increased levels of contact between the parties involved. Additionally, the authors of [42] argued that one may infer greater reliability and validity from sources with strong ties than from those with weak ties. In other words, knowledge obtained from reliable sources has a lower level of risk. Social exchange theory demonstrates that sources with strong ties are more likely to provide knowledge of greater economic value than sources with weak ties [43,44]. There is a better chance of better information coming from a trusted source than from someone you do not know very well. While customers trust strong-tie sources and the importance of the information supplied by strong-tie sources, consumers find it easier to assess whether strong-tie sources provide data about these products. In other words, customers feel that a reliable source can assist them in evaluating the quality and effectiveness of a product.

Hypothesis 1. S-tie has a positive effect on customers' purchase intention.
There are significant psychological and economic implications resulting from customers' ability to precisely predict their future experiences. An underestimation of future pleasure may result in forgone possibilities for the customer, as well as the marketer, while an overestimation of future enjoyment may lead to post-purchase disappointment and dissatisfaction [45]. People are not very good at formulating effective predictions, especially when it comes to enjoyment [46]. Forecasting mistakes are most often linked to inaccurate modeling of future experiences [47], and prescriptive counsel frequently seeks to improve the simulation process. According to [45], the function of eWOM in enhancing customers' capacity to anticipate the level to which they would appreciate the products and services offered in the market is becoming more important. The authors of [48] found that eWOM is a key predictor of perceived enjoyment. In addition, the authors of [45] studied how eWOM can predict consumption enjoyment and found that eWOM from strong ties can accurately help consumers forecast their own product enjoyment.

Hypothesis 2.
S-tie has a positive effect on customers' perceived enjoyment.
The authors of [49] analyzed the effect of user activities on YouTube and brand activities on Facebook, revealing that both have a favorable effect on the value of the brand. It is the main responsibility of the advertiser to deliver value to the consumer, and businesses may often modify the marketing mix to increase the value of their products to the target demographic or market group that they have in mind. The idea of customer-perceived value, which is sometimes referred to as perceived value or consumer value in the relevant literature, continues to play an important role in marketing [50]. Perceived value is examined through their "interactive relativistic preference experience," according to the findings of [51]. The academic community has continuously focused on gaining a better understanding of consumer perceptions of value, given the importance of this variable as a significant predictor of marketing results [52].
Researchers and brands often focus on the spread of eWOM because it is seen as a way to measure the results of marketing efforts or post-purchase behavior. [53]. Researchers have recently begun studying eWOM as an antecedent that affects consumer views and future behaviors (e.g., [37,54,55]).

Hypothesis 3.
S-tie has a positive effect on customers' perceived value.

Celebrities eWOM
Due to the rapid advancement of technology, consumers tend to change their preferences and desires toward a company's products and services more frequently [56]. The emerging concept of celebrity endorsement has grown concurrently as corporations need to attract consumers in the market and compete with their competitors in making sales and acquiring market shares. As a result, many companies from different industries increasingly employ celebrities as their brand spokesperson and representative [57]. Since celebrities can make commercials stand out from the surrounding clutter, they should be used wherever possible. It has been observed that products and services sponsored by celebrities have a higher degree of product recall, which helps such products stand out from the crowd and get more attention, while consumers shop for those products [58]. It is then undeniable that celebrity endorsement is an important marketing strategy, as this kind of brand representative can potentially influence consumers' attitudes [59].
The authors of [60] reported that the social influence of celebrities on an online platform can affect consumers' purchase decisions and behaviors. In addition, the authors of [61] argued that the person conveying the marketing messages of a brand can affect the customer, whereby more important role models have a greater effect on the customer. For this reason, in the process of advertising their goods, business owners employ celebrities to promote their products so that customers may accept their products and services.
Numerous studies have proven that celebrity endorsements in advertising are successful and have a favorable effect, notably on the endorser's integrity, message recall, support, and buy intent [62]. It is generally agreed that an individual's perceived enjoyment is an important factor in determining their emotional state or response [63]. The term "perceived enjoyment" refers to the extent to which a user believes that using the product or making use of the service will be pleasurable to them [64]. We are affected by the media in many different ways. It has an effect on the way we communicate, feel, and behave, including our views of reality and enjoyment. The use of celebrity endorsements in advertising is quite prevalent since it helps to develop trust with existing consumers and future customers, increases the probability of brand recall, and attracts a different kind of target [65]. The authors of [66] suggested that customers create connections with celebrities that enable them, via imagination, to flee the pressures and hassles of daily life by imagining experiences with the celebrity. This argument is based on the notion that consumers build associations with celebrities that allow them to fantasize about spending time with them. These escapes allow the celebrity to supply customers with pleasant emotions that help develop the bond between the two parties. Consumption of a celebrity-endorsed product for its hedonistic value revives these kinds of emotions for consumers. Hedsonic values are used to create an illusory face-to-face link with the audience in parasocial engagement [67]. Customers who place high importance on hedonic experiences interpret product content evaluations by celebrities as honest information regarding products [68,69].

Hypothesis 5.
Celebrities eWOM has a positive effect on customers' perceived enjoyment.
The strategies used in the marketing of a product have the potential to influence a person's attitudes, either positively or negatively, toward products [70]. Usually, an individual's conviction that there is some value connected with possessing the thing that is being promoted drives the creation and change of an individual's attitude [71]. Therefore, a buyer will not buy a brand's products and services if they do not think they have value [72]. This might be because of an emotional connection to the product, price, and quality, or even a sense of social acceptability that comes with buying services and products [73].
According to [74], advertisers have been using the assistance of celebrities for a long time to boost the value of a brand. Celebrities make an appearance in the media on behalf of a promoted brand, whereby they deliver information to the public, share their experience, and offer the audience a strong influence on the brands that they endorse. Celebrities gain a larger degree of attention and value than other types of endorsements [75]. Because the cost and quality of a product are vital factors in determining how much perceived value consumers give to a product, celebrity endorsements are a powerful tool for shaping consumer preferences [76]. Hypothesis 6. Celebrities eWOM has a positive effect on customers' perceived value.

Perceived Value and Perceived Enjoyment as Mediators
People's perceptions of value significantly influence whether they intend to make a purchase, and this influence may be utilized to forecast how consumers will behave when making decisions [77]. It has also been established that there is a direct and beneficial influence of perceived value on the behavioral intentions of consumers [9]. Moreover, it is considered a crucial antecedent for establishing customers' buying intentions in the existing body of research [10]. In addition, the authors of [11] discovered that customers' perceptions of product value had a beneficial impact on their purchase intentions.
On the other hand, perceived enjoyment refers to the degree to which customers experience amusement when making use of a product or service [78]. After constructing the TAM, perceived enjoyment was incorporated into the model in [12], and the associated findings suggest that it has a large effect on adoption intention. Additionally, buying intentions are influenced by perceived enjoyment [13]. Individuals are motivated to purchase for various reasons, including avoiding boredom, conforming to the expectations of their peers, and maintaining their social position [79]. The authors of [80] theoretically extended the S-O-R model initially developed by [81], which explains human actions by analyzing the internal responses affected by the environmental stimulus (Shah et al., 2020). In other words, the external environment stimulus (S) encourages the organism's internal response (O), thus encouraging a behavioral response (R). The stimulus is a trigger that provokes consumers [82]. The organism is a function of the cognitive and affective state of an internal state of an individual [83], which plays an intermediary role between stimulus and response [84]. Lastly, the response indicates the behavior of the consumer [85]. Researchers have widely used the S-O-R model in marketing [86]. The current study proposes that S-tie and celebrity eWOM can be considered external stimuli, while perceived value and perceived enjoyment are part of the internal response (organism), leading to purchase intentions as a behavioral outcome.
Hypothesis 9a. Perceived value mediates the effect of the S-tie on customers' purchase intention.
Hypothesis 9b. Perceived value mediates the effect of celebrity eWOM on customers' purchase intention.
Hypothesis 10a. Perceived enjoyment mediates the effect of S-tie customers' purchase intentions.
Hypothesis 10b. Perceived enjoyment mediates the effect of celebrity eWOM on customers' purchase intention.

Weak-Tie eWOM as Moderator
Consumers often try to find information whenever possible when looking to buy new products to reduce the risks associated with the decisions they make [39]. People are willing to listen to WOM, and it will impact their decision-making process as a function of their evaluation of what they perceived [30,37]. Although several source characteristics impact the effectiveness of WOM, the relationship closeness to the source is the most significant one [87]. Nonetheless, weak-tie eWOM is an important information source that helps consumers make their purchase choices [31]. Currently, it is an easy process for consumers to get information from different sources thanks to social media platforms on the internet [25]. Therefore, W-tie contributes to the flow of information, as it offers access to new information. Usually, consumers rely on weak ties when the number of choices increases [23], and they increase their knowledge after receiving recommendations from strong ties [31].
The authors of [31] argued that W-tie can influence the power of S-tie, as well as perceived value. They proposed that W-tie can reduce the effect of S-tie, resulting in a negative moderation effect. In other words, when consumers receive conflicting information between S-tie and W-tie, they become dissonant and uncomfortable regarding the dissimilarity of opinions, thus reducing the effect of W-tie recommendations. Their empirical findings revealed that the W-tie had a negative effect on the relationship between the S-tie and purchase intention. In contrast, they found that the W-tie had a positive moderation effect between the indirect effect of perceived value and purchase intention.
Other researchers have also studied the moderating effect of social ties on consumer behavior; for example, the authors of [28] found that social ties strengthen the influence of a price change communicated by WOM on information dissemination. In addition, the role of tie strength moderates the relationship between customers' involvement and WOM influence [29]. Furthermore, the authors of [27] found that social ties moderate the effect of WOM receptiveness (the degree of consumers relying on WOM in decision making) and intention to go to movies. On the other hand, tie strength does not always have a moderation effect, as the authors of [88] found that tie strength does not moderate the relationship between WOM source components (expertise, trust, and message richness) and purchase intention. Hypothesis 11h. W-tie moderates the effect of perceived enjoyment on customers' purchase intentions.

Conceptual Framework
The research model was developed on the basis of the work of [31] while adding two more variables found in the literature: celebrity eWOM as an independent variable and perceived enjoyment as the mediator (see Figure 1). Hypothesis 11a. W-tie moderates the effect of S-tie on perceived value. Hypothesis 11b. W-tie moderates the effect of S-tie on perceived enjoyment. Hypothesis 11c. W-tie moderates the effect of S-tie on purchase intention. Hypothesis 11d. W-tie moderates the effect of celebrity eWOM on perceived value.

Hypothesis 11e. W-tie moderates the effect of celebrity eWOM on perceived enjoyment.
Hypothesis 11f. W-tie moderates the effect of celebrity eWOM on purchase intention.

Hypothesis 11g. W-tie moderates the effect of perceived value on customers' purchase intentio
Hypothesis 11h. W-tie moderates the effect of perceived enjoyment on customers' purchase tentions.

Conceptual Framework
The research model was developed on the basis of the work of [31] while adding tw more variables found in the literature: celebrity eWOM as an independent variable a perceived enjoyment as the mediator (see Figure 1).

Methodology and Analysis
There are 6.84 million social media users in Jordan, which accounts for 61.5% of t Jordanian population. The most popular social media platform in Jordan is Facebook, w 5.5 million users (43.8% females and 56.3% males) [89]. Any adult has the authority to b products in Jordan, and this study aims to test eWOM sources on purchase intentio Therefore, the population of the study is any adult who has access to the internet and h an account on any social media platform.

Methodology and Analysis
There are 6.84 million social media users in Jordan, which accounts for 61.5% of the Jordanian population. The most popular social media platform in Jordan is Facebook, with 5.5 million users (43.8% females and 56.3% males) [89]. Any adult has the authority to buy products in Jordan, and this study aims to test eWOM sources on purchase intention. Therefore, the population of the study is any adult who has access to the internet and has an account on any social media platform.
This study employed a quantitative research approach by developing an online selfadministrative questionnaire on Google Docs and utilizing convenience sampling and snowball techniques. The URL was distributed on two widely used social media platforms (Facebook, WhatsApp, and Instagram) before asking peers to fill out the questionnaire (convenience sampling), as well as share the URL with their peers to get more responses (snowball sampling). In addition, the questionnaire was split into two sections. The first section covered demographic data, and the second section covered items related to the variables. Thirty-two items covering six variables from the literature were used to develop the model, and they were translated into Arabic. Thereafter, colleagues checked the questionnaire to ensure that the terminologies were correct and easy to understand. To reduce the nonresponse error, all fields were made mandatory prior to submitting the questionnaire and by comparing the characteristics of the sample with those of the population. The data collection was carried out in the period from February 2022 until the end of May 2022. After collecting all data, 726 questionnaires were found to be complete. Thereafter, both SPSS and AMOSv23 were used to validate the data and test the hypotheses. Table 1 shows the demographic data of the respondents.

Convergent Validity
To ensure the model's convergent validity, the threshold values in [90] were followed. Five items that cross-loaded into more than one latent variable were eliminated (Enj3, Enj5, Pval 5, Pval6, and Pval7) before reconducting the factor loading analysis. As shown in Table 2, all values met the criteria of factor loadings >0.6, composite reliability >0.7, and average variance extracted (AVE) >0.5. For the assurance of no multicollinearity between variables, the variance inflation factor (VIF) and factor correlation matrix were applied. The range of VIF values for all variables was 1.1-1. 39. In addition, all correlation values between variables in the factor correlation matrix were <0.7, as shown in Table 3.

Discriminant Validity
The procedure described in [91] was used to check for discriminant validity. Table 4 demonstrates that all correlation values were smaller than the square root of the AVE values, indicating that the model had appropriate discriminant validity.

Model Fit
The authors of [92] proposed using chi-square, RMSEA, SRMR, and CFI to test the model's goodness of fit. Table 5 shows that all values met the thresholds, indicating a good fit.  Figure 2 shows the structural model produced by AMOS. Furthermore, (Table 6) shows the direct-effect hypothesis-testing results. All hypotheses (H1-H8) presented a p-value > 0.05. Celebrities eWOM had a greater effect than S-tie on perceived value, perceived enjoyment, and purchase intention. In addition, perceived enjoyment (β = 0.376) had a greater effect on purchase intention than perceived value (β = 0.108). The R 2 value of purchase intention was 0.49, indicating that 49% of purchase intention was explained by variables in the model. Purchase intention is a sophisticated topic, with many other variables (not included in the model) influencing it.

Structural Model and Hypothesis Testing
shows the direct-effect hypothesis-testing results. All hypotheses (H1-H8) presented a pvalue > 0.05. Celebrities eWOM had a greater effect than S-tie on perceived value, perceived enjoyment, and purchase intention. In addition, perceived enjoyment (β = 0.376) had a greater effect on purchase intention than perceived value (β = 0.108). The R 2 value of purchase intention was 0.49, indicating that 49% of purchase intention was explained by variables in the model. Purchase intention is a sophisticated topic, with many other variables (not included in the model) influencing it.   Table 7 shows the interaction effect of W-tie with the variables of the study on the intended variables. The interaction between S-tie and W-tie did not affect the perceived   Table 7 shows the interaction effect of W-tie with the variables of the study on the intended variables. The interaction between S-tie and W-tie did not affect the perceived value, perceived enjoyment, or purchase intention, as all p-values were >0.05. Therefore, hypotheses H11.a, H11.b, and H11c were rejected.
In contrast, the interaction of celebrity eWOM and W-tie had an effect on perceived value, perceived enjoyment, and purchase intention, as all p-values were <0.05. As shown in Figure 3, W-tie strengthened the positive relationship between celebrity eWOM and perceived value (β = 0.89); thus, hypothesis H11d was accepted.  In contrast, the interaction of celebrity eWOM and W-tie had an effect on perceived value, perceived enjoyment, and purchase intention, as all p-values were <0.05. As shown in Figure 3, W-tie strengthened the positive relationship between celebrity eWOM and perceived value (β = 0.89); thus, hypothesis H11d was accepted.   As shown in Figure 4, W-tie strengthened the positive relationship between celebrity eWOM and perceived enjoyment (β = 0.074); thus, hypothesis H11e was accepted.    Figure 5, W-tie strengthened the positive relationship between celebrity eWOM and purchase intention (β = 0.84); thus, hypothesis H11f was accepted.  As shown in Figure 5, W-tie strengthened the positive relationship between celebrity eWOM and purchase intention (β = 0.84); thus, hypothesis H11f was accepted. As shown in Figure 6, W-tie strengthened the positive relationship between perceived value and purchase intention (β = 0.86); thus, hypothesis H11g was accepted. As shown in Figure 6, W-tie strengthened the positive relationship between perceived value and purchase intention (β = 0.86); thus, hypothesis H11g was accepted.   As shown in Figure 7, W-tie strengthened the positive relationship between perceived enjoyment and purchase intention (β = 0.105); thus, hypothesis H11h was accepted. As shown in Figure 6, W-tie strengthened the positive relationship between perceived value and purchase intention (β = 0.86); thus, hypothesis H11g was accepted.   The results of the direct effects revealed that S-tie and celebrity eWOM had a significant influence on purchase intention. Bootstrapping analysis with a 95% bootstrap confidence interval (Table 8) revealed significant indirect effects of celebrity eWOM on purchase intention through perceived value (lower level (LL) = 0.008, upper level (UL) = 0.047, p < 0.05), S-tie on purchase intention through perceived enjoyment (LL = 0.014, UL = 0.083, p < 0.05), and celebrity eWOM on purchase intention through perceived enjoyment (LL = 0.112, UL = 0.225, p < 0.05). Thus, hypotheses H9b, H10a, and H10b were accepted. Nonetheless, both perceived value and perceived enjoyment could be considered partial mediators of these relationships because the relationships between the independent variables and purchase intention were statically significant. On the other hand, S-tie had a nonsignificant influence on purchase intention through perceived value (LL = 0, UL = 0.021, p > 0.05); thus, hypothesis H9a was rejected.

Discussion
This study investigated how strong-tie, weak-tie, and celebrity eWOM work together to affect consumers' purchase intentions by studying the moderation effect of the Wtie and the mediation effect of perceived value and perceived enjoyment on consumers' purchase intentions. Currently, social media platforms play a pivotal role in all aspects of our lives, and consumers rely heavily on the recommendations and experiences of their peers and even strangers. Accordingly, brands spend billions of dollars on celebrities for access to their followers by promoting products and services [93]. The empirical results supported numerous hypotheses, and all the direct relationships in the model were supported. Specifically, our findings demonstrate that the S-tie has a significant effect on perceived value. Consumers share many common characteristics with their strong ties, usually hold a soft spot for them, and often consider them to be trusty advisors. This, in turn, affects their judgment regarding the perceived value of most products and services. These results are in line with previous studies in various contexts [31,[94][95][96][97]. A study conducted by [31] found that the S-tie affects perceived value. Likewise, the authors of [94] found that eWOM C2C conversations affect perceived value. In addition, similar results from the study by [95] noticed that eWOM between social ties on social commerce platforms affects platform customers' perceived value. In another context, the authors of [96] found that eWOM significantly affected tourists' perceived quality in Indonesia. Furthermore, the authors of [97] found that eWOM affected customers' perceived value in the altruistic service context.
Our findings showed that the S-tie affects perceived enjoyment. We believe that when consumers hear positive or negative experiences about products and services from their S-tie circle, either as direct online conversations or by reading reviews or ratings, they will begin to imagine how their experience will be when consumption occurs. The results of this research are aligned with similar results from previous studies. In a social commerce context, a study conducted by [98] found that eWOM affects perceived enjoyment. Similar results were found by [48] when they studied the influence of eWOM on purchase intentions in the Taiwanese laptop industry. Additionally, the authors of [45] found that eWOM accurately affects the future enjoyment of consumption for different product categories, especially when there are similarities between the sender and receiver. Furthermore, in the luxury hotel industry, [99] found that eWOM has a significant predictive effect on perceived enjoyment for Generation Y customers. However, the authors of [100] had contradictory results, as they found that positive eWOM does not predict perceived enjoyment. Nevertheless, they found that eWOM quality, such as in-depth reviews, had an impact on perceived enjoyment.
Another finding showed that the S-tie significantly affects purchase intentions. Consumers seek advice from trusted peers to reduce the risks associated with potential purchases. As previously mentioned, closeness, and trust between individuals and their S-tie peers will affect their judgment regarding products and services, as it will affect their purchase intention. Previous studies regarding eWOM in general terms, regardless of tie strength, showed mixed results regarding the effect of eWOM on purchase intentions. Research conducted by [99], they found that eWOM does not affect purchase intention in luxury hotels. Moreover, eWOM does not affect the purchase intentions of organic food products [101]. In contrast, eWOM was found to affect purchase intentions (e.g., [102][103][104]). Additionally, S-tie has been found to affect purchase intentions in several studies [19,25,31,105].
Furthermore, the findings showed that celebrity eWOM significantly affects perceived value. Consumers have an affinity for and admiration for celebrities. Consequently, the brand messaging that celebrities share becomes more credible, which leads to a higher perception regarding product value. The results of this study are in line with previous studies [76,106,107]. Additionally, celebrity eWOM significantly affects perceived enjoyment. Celebrities' messages trigger consumers to develop emotional engagement and imagine how they will enjoy the experience with those products or services, as supported by previ-ous studies [64,108,109]. Moreover, celebrity eWOM significantly affects purchase intention. The attraction that consumers hold for celebrities will drive them to attempt to mirror celebrities they admire, which includes endorsed products and services. This, in turn, will lead to the intention to buy such products to impersonate the celebrity. Surprisingly, these results contradict what the authors of [110] found when they studied the packaged coffee market. Their results highlighted that celebrities' endorsements did not directly affect purchase intention. Likewise, another study conducted by [111] for Vanilla Hijab found that celebrities' endorsements did not affect purchase intention directly. However, in both studies, they found that celebrities' endorsements affected other important antecedents of purchase intention, such as brand loyalty, brand attitude, and brand credibility. Still, many studies have found that celebrities' endorsements have significant effects on purchase intention [107,108,[112][113][114]. Interestingly, from a statistical point of view, the results of this research showed that the effect of celebrity eWOM was considerably higher than that of S-tie. For instance, when comparing the β-values for their effect on perceived value, celebrities eWOM β = 0.209, while S-tie β = 0.82. When comparing their effect on perceived enjoyment, celebrities eWOM β = 0.388, while S-tie β = 0.123. Lastly, for their effect on customers' purchase intention, celebrities eWOM β = 0.209, while S-tie β = 0.132. These results come as no surprise, as consumers view celebrities as role models. The authors of [61] argued that the celebrity conveying the brand's message affects purchasing behavior, especially for younger consumers.
Another aim of the study was to test whether perceived value and perceived enjoyment mediated the relationships between the independent and dependent variables of the model. The results showed that not all hypotheses were accepted. Both variables have a significant direct effect on purchase intention. When consumers value brand offerings and perceive that they will enjoy their future consumption, this, in turn, will increase their purchasing intent. These results are similar to those of previous studies, as the authors of [11,31,115,116] found that perceived value is an important variable in predicting purchase intention. Furthermore, previous studies have revealed that perceived enjoyment has an effect on purchase intention [64,108,109]. When it comes to the mediation effect results, perceived value only mediated the impact of celebrity eWOM on purchase intentions. In other words, perceived value did not successfully convey the effect of S-tie on purchase intentions, contradicting the results of [31], which found that perceived value mediates the relationship between S-tie and purchase intention. Furthermore, the authors of [117] did not find that perceived value mediates the relationship between celebrity endorsement and purchase intention. On the other hand, perceived enjoyment mediated both relationships. This result is in line with the results of [118]. They found that hedonic enjoyment mediates the relationship between eWOM and purchase intention.
The final aim of the study was to test the moderation effect of W-tie on eight proposed hypotheses. The results showed that W-tie did not moderate the relationship of S-tie with regard to perceived value, perceived enjoyment, and purchase intention. Thus, consumers who value S-tie eWOM do not have their intentions swayed by what is written on social media platforms. This result also contradicts the findings in [31], which indicated that W-tie negatively moderates the relationship between S-tie and purchase intention. This contrast may stem from differences between Jordanian and Chinese cultures. On the other hand, W-tie moderated the relationship between celebrity eWOM and perceived value, perceived enjoyment, and purchase intention, as well as the relationship between perceived value and perceived enjoyment and purchase intention. These results are in line with those of [31], which found that W-tie moderates the relationship between the mediation effect of perceived value and purchase intention. Our results indicate that the effects of celebrity eWOM, perceived value, and perceived enjoyment depend on the level of W-tie. When W-tie is high, the effect becomes stronger, and vice versa; when W-tie is low, it will reduce the effect of celebrity eWOM on perceived value, perceived enjoyment, and purchase intention due to internal conflicts and dissonance. The same applies to the effects of perceived value and perceived enjoyment with regard to purchase intention.

Theoretical and Practical Contribution
This research examined how different sources of eWOM (e.g., S-tie, celebrity eWOM, and W-tie) interact to affect customers' purchase intentions by studying the moderation effect of W-tie and the mediation effect of perceived value and perceived enjoyment on consumers' purchase intentions. Most studies have measured tie strength (i.e., strong/weak) as one construct. The versus characteristics of these two constructs were considered in almost all studies. Past studies have combined tie strengths at the same level of the model, either as an independent variable (e.g., [19][20][21][22][23][24][25]) or as a moderation variable (e.g., [26][27][28][29][30]) in different contexts. In their methodology, different groups (e.g., strong-tie groups and weaktie groups) were studied, or one group was exposed to different scenarios (e.g., strong-tie or weak-tie eWOM) before comparing the results. This research delivers new knowledge to the body of literature regarding the mechanism of eWOM sources by considering S-tie and W-tie separately. The results showed that W-tie moderated the positive impact of celebrity eWOM, indicating how brands may manage their efforts to increase the value of their products and services to consumers.
From a statistical point of view, the results revealed a stronger effect of celebrity eWOM than S-tie on consumers; thus, brands should utilize them in their marketing strategies. However, the impact of the S-tie on consumers must not be neglected. Easy access to the internet is resulting in stronger social influence than ever before, and marketers should be aware that consumers are relying on eWOM information that aids their decision-making processes. Marketers should have a presence wherever their customers are on the internet. This does not necessarily apply solely to official brand pages and forums, but rather to all third-party accounts, pages, and platforms where consumers are online. This will offer brands the opportunity to better understand their customers and adjust their future offerings. Additionally, brand managers should make it easy for consumers to be in touch with comments or questions. Neglecting will increase negative eWOM and damage brand value. On the contrary, sincerely listening to customers' complaints and finding ways to solve their problems will change negative views into positive ones. In addition, marketers should drive customers to boost positive experiences through comments and reviews. Running simple contests and rewarding customers who review is an adequate approach to obtaining help from customers and giving them a reason to share their brand experiences.
It is necessary for any organization to be sustainable in business. Therefore, brands should improve the customer experience and retention. Although technology provides consumers with various ways to communicate and learn about products and services, it also gives brands the opportunity to monitor and evaluate consumers' conversations, comments, and feedback posted on the internet. Harvesting this opportunity will ensure that brands can achieve favorable long-term results.

Limitations and Future Research
All studies have their own set of limitations that should be acknowledged. First, this study did not consider a specific product category in the market, and different categories may produce different results. Second, we studied perceived value as an aggregate construct, despite its many sub-dimensions, such as price, quality, emotional value, and social value. Third, the results of the study cannot be generalized due to the sampling method applied.
Future research could conduct a multigroup analysis according to gender and age group. Gender differences play a pivotal role in different aspects of consumer behavior, ranging from information processing to purchase intentions [119]. Moreover, gender differences impact decision making [120]. In a recent systematic review [121], it was concluded that there are more differences than similarities between genders in the online context. Furthermore, age is also an important element in purchase decisions [122], with variations in engagement and motivation according to age [123]. However, studies have revealed that the effects of age differences vary according to context [124].