Effects of Celebrity–Product/Consumer Congruence on Consumer Confidence, Desire, and Motivation in Purchase Intention

Abstract

It is well known that the consistency of Internet celebrity affects consumers’ purchase intention, but there are few studies on the logic behind the influence. This study investigates the impact of Internet celebrity congruence factors (celebrity congruence with the recommended products and celebrity congruence with followers) on followers’ purchasing psychological states and purchase behaviors. The research model is rooted in congruence theory and social-network theory and we conducted an empirical study on Tiktok and IG followers of celebrities. Then, we used partial least squares structural equation modeling (PLS–SEM) to analyze 297 valid questionnaires. The results show the congruence of Internet celebrity with both followers and products promotes follower product adoption; the congruence does so by enhancing followers’ perception that the recommendations are credible, in turn increasing the followers’ confidence in, desire for, and motivation to buy the product. This study explains that the congruence of Internet celebrities with followers and products must be strongly present in the recommendation, which leads to a high credibility recommendation. A high credibility recommendation greatly influences consumers’ pre-purchase psychological states and thus enhance their purchase intention. This study demonstrates to practitioners that when consumers’ purchase intention is lower than expected, it is necessary to consider the credibility of the recommendation.

1. Introduction

Purchase intention has always been a topic of concern to enterprises and scholars. With the emergence of social media and its various forms of information, how these affect consumers’ purchase intention has aroused more extensive and in-depth research. A prominent phenomenon on social media platforms is massive information presentation and their social persuasion. People are easily persuaded by social information or celebrities, who get a lot of attention on social media, especially when buying products. The celebrities who influence the attitudes or behaviors of others are called ‘influencers’; they persuade their followers to purchase products [1,2,3]. Influencer marketing as a marketing strategy often increases profits for sellers and is considered a new marketing frontier, one that marks a major shift from traditional offline marketing to novel online marketing. Not surprisingly, then, a great deal of research has addressed the reasons people are influenced by celebrities and the role of Internet celebrities in social media marketing. Much previous research [1,2,4] demonstrates that because Internet celebrities provide indirect purchasing knowledge to readers, the recommendations from Internet celebrities can significantly affect their followers’ attitudes toward various kinds of consuming targets (e.g., stores, products, and services); especially in the context of congruence, it enhances persuasion.

Various prior studies have postulated the factors of source credibility which can shape consumer cognition toward the recommendations and recommendation adoption, and many of them stem from congruence theory [1,5,6]. Much of the research discusses the issue of source credibility in social media and whether the congruence relationship (celebrity congruence with consumer/celebrity congruence with product) significantly affects consumers’ purchasing behavior [1,4,7,8,9,10]. All studies indicate that a congruence relationship improves source credibility; high source credibility makes persuasion more convincing. If congruence relationships can be the antecedent factors which affect followers’ cognition toward the recommendations, how do these two distinct congruence factors affect information readers’ cognition processes toward presented information and generate purchase intention? How does the reader’s psychological state [11] change between cognition and purchase intention?

While congruence is vital to explore the influence of Internet celebrity in social media marketing activities, researchers have not yet systematically investigated this issue. The aim of the present study is to fill this gap in the relevant literature by combining both celebrity congruence with products and celebrity congruence with consumers, and investigating their impacts on followers’ pre-purchase psychological states and purchase behaviors. In general, we believe that further exploring these unanswered questions is both interesting and crucial, as the work will help scholars further understand the application scope of congruence inference on this point.

As stated above, this study discusses and analyzes two key research questions:

(1)

Does the congruence relationship enhance followers’ perception that the recommendations are credible?

(2)

How does recommendation credibility increase followers’ confidence in, desire for, and motivation to buy the product?

We collected field data from the members of two famous social media platforms in Taiwan and mainland of China and conducted statistical analyses to examine the hypotheses stemming from the research questions.

2. Theoretical Backgrounds and Research Hypotheses

Internet celebrities are spokespersons for products, brands, and organizations because of online popularity in a particular cultural or social milieu. A characteristic that these individuals have in common is their rapid rise to fame, often through participation in some real or virtual events that attract widespread attention from netizens [12] and can realize social assets through self-promotion [13]. Other scholars describe Internet celebrities as people possessing a degree of expertise on a saleable product or service in a relevant field [6,14,15]. They are widely recognized in society for their personal attractiveness, garnering media attention, and having legions of devoted fans (i.e., followers); they influence fans to generate purchase decisions or purchase behaviors under their recommendation [14,15,16,17].

Internet celebrities create their original online content on social media [18]. The value that fans assign to the original content can be measured in the number of followers, likes, and re-posts that these social media accounts garner. Another measure of the value that fans assign to an Internet celebrity is the extent to which this person influences their perceptions and behaviors [1,12].

In short, celebrities become famous due to their appearance or personality attractiveness, professional skills, and the ability to provide valuable information to Internet users [13,14,19,20,21]. They are often invited by enterprises to endorse brands; one reason is that information from a celebrity helps consumers process advertising content [22]. The other reason is celebrities attract and keep an audience’s attention [21], and improve consumer trust in endorsed brands [22].

2.1. Celebrity Congruence with Products

The congruence theory proposed by Holland in 1973 became axiomatic among some research circles. At its core, congruence theory refers to one thing’s harmony with another thing or resemblance to something adjacent [23]. Since Holland’s proposals, researchers have put forward various principles on the basis of his axioms [23]. Some congruence theory literature is fleshed out with auxiliary theories and applied to empirical investigations into, for example, consumers’ behavior congruence with their self-image [24].

In marketing, congruence theory can help explain the increase of advertising benefits caused by the effective match between an endorser’s celebrity image and the endorsed product [25]. Here, the term ‘congruence’ refers generally to the fit, match, agreement, or similarity between two constructs [16]. Namely, certain characteristics associated with a celebrity must be consistent with certain characteristics of the endorsed products [26]. Congruence presents a coherent message and makes product information more interesting, visualized, and vivid. Visual information can enhance the perception of product message quality [9] and product information vividness can enhance the persuasiveness of an endorsement [27]. In addition, like vividness, credibility can strengthen potential buyers’ evaluation of the endorsed product [10,27,28]. The greater this congruence, the more effective the celebrity recommendation [16,29]. A rather obvious example of this kind of congruence is the use of attractive models to sell stylish clothing and accessories. McCracken (1989) [16] point out that celebrity congruence with a product is a process that, in the abstract, amounts to a “meaning transfer” [16,17,30]. It is the mean transfer of celebrity image from the celebrity to the endorsed product, and lastly the significance of celebrity endorsement is transferred to consumers through purchase and consumption of the product [10,16,17]. It goes without saying that the more popular the celebrity is, the clearer consumers are about the brand image and product value [26], and the greater the marketing impact of this meaning transfer is likely to be [16,17]. In addition, congruence can reduce consumers’ cognitive bias toward products before and after purchase, enhancing product credibility [31].

Consumers often purchase a product whose function is to improve personal self-image [32,33]. People use a particular brand name to reflect either their personality or the personality that they desire [34]. This process can make consumers feel good about their purchase of the product. When the characteristics of a celebrity endorser are consistent with the characteristics of the endorsed product, consumers can deepen their understanding of the product by applying what they already know—or think they know—about the celebrity to the product, improving the recognition of product credibility [7,10,35]. Thus, it was found to increase celebrity recommendation persuasion [36]. Put another way, consumers who are unfamiliar with a product but have favorable attitudes toward a celebrity are consequently predisposed to transferring these attitudes to the product [25,29,35]. It fully reflects the significant transfer effect of celebrity and product matching.

2.2. Celebrity Congruence with Consumers

Consumers have internal and external aspects. At the intersection of these two aspects is self-perception (The term ‘self-perception’ is interchangeable with such terms as ‘self-concept’, ‘self-notion’, and ‘self-image’). Self-perception concerns how people think about themselves [33,37]. From a marketing standpoint, we can say that consumers’ self-perception influences purchase intention [33]. Many consumers not only express their taste but also construct their self-image through the purchase of products [33,37]. When the product image matches the consumer’s actual self-image, the result is called self-consistency, and when the product image matches the consumer’s ideal self-image, the result is called ideal-image consistency. The central goal is to make one’s actual self-image mirror one’s ideal self-image [32,33].

The influence of celebrity spokespersons on consumer behavior depends on consumers’ perception of the given celebrity’s image. Celebrities are perceived to have desirable knowledge, skills, experience, and be physically/personality attractive, all of which add up to credibility [20]. This specialness attributed to celebrities can very powerfully shape how consumers perceive and evaluate their ideal self-image and their actual self-image [33]. Thus, a celebrity can represent an actual or ideal group of individual references [32]. To clarify for themselves what their ideal self-image is, consumers use ideal celebrities as a litmus test, gauging the consistency between the celebrities’ image and the consumers’ own self-image. After all, many consumers desire to resemble celebrities. Studies have proved that the more consistency between consumers’ ideal self-image and celebrities’ image, the more consumers imitate the behavior and dress of celebrities, and the greater the celebrity endorsement effect is [7,38]. This congruence comes on top of the fact that, as explained by social influence theory and self-esteem motive theory, consumers generally have a strong tendency to follow the advice of celebrities.

2.3. Source Credibility

Source credibility theory is widely used in the research of information persuasion in online and offline environments [39]. Source credibility refers to the trustworthiness of the source of information [10,36] that affects the degree to which information is accepted. Source credibility theory posits there is a direct effect on the causal relationship between informational factors and information credibility. One aspect of source credibility is the name recognition of the source [40]. Another critical aspect of source credibility is the source’s reputation [41]: is the celebrity individual punctilious, honorable, ethical, objective, authority, believability, honest, and judicious in decision-making [10,20]? The perceived expertise of a celebrity has a significant impact on consumers’ positive attitudes toward the endorsed products [2,30,36,41]. In addition, important for source credibility is the types of congruence (endorser–brand/product and endorser–consumer congruence) [42]. Past advertising-related research has proved that these two types of consistency enhance source credibility, making it easier to attract customers [7,19]. Quite obviously, an identifiable source is more credible than an unidentifiable one, and an Internet celebrity does enhance product and advertisement credibility when he/she is congruent with the product and consumer [42].

Research has demonstrated people’s behavior complies more with instructions from highly credible sources relative to low credibility sources [6,36,39]. In social media marketing, consumers can evaluate the celebrity’s credibility from a source credibility theoretical lens [22], using various cues [39], such as by scrutinizing the profile of the celebrity to get more detailed information (e.g., the celebrity’s other recommendations, his/her activities in this forum, as well as other members’ evaluation of this celebrity). Previous research [42] scrutinizes the types of congruence and finds that the types of congruence can directly form or change a consumer’s attitude, and that information provided by a highly congruent source produces a greater effect on perceived information credibility. Thus, the consumers are more inclined to adopt the viewpoint of the information [39].

Source credibility has a great influence on the persuasiveness of a message [10,36,39]: the contents of messages are more highly positive, and the impact on persuasion are more obvious [36,39,43]. Given the aforementioned findings about congruence (i.e., endorser–brand/product congruence and endorser–consumer congruence), congruence has a highly positive effect on recommendation persuasion, from the theoretical lens of source credibility.

In this way, the effect of a highly persuasive recommendation on its credibility is increased. From a less credible source, the persuasive credibility of the recommendation is lessened [7]. Therefore, we propose the following hypothesis:

Hypothesis 1 (H1).

The more congruence there is between an Internet celebrity and a product, the more credible the Internet celebrity’s recommendation of the product will be.

Hypothesis 2 (H2).

The congruence between Internet celebrity with consumer will have a positive effect on the recommendation credibility.

2.4. The Relationship between Recommendation Credibility and Psychological States

Briñol, Petty, and Tormala (2004) [44] suggest people’s confidence response to persuasive information is affected by the source credibility of the processed information; high source credibility generates positive thoughts and more favorable attitudes than low source credibility [45]. Confidence is defined as a sense of correctness in one’s beliefs either from personal experiences or from the quality and quantity of information they have [46]. Desire is also viewed as a single belief which guides individual actions and judgments [47], acting as a catalyst to transform motivations to actions [48], due to one’s yearning for an object [34,39], or as a driver of brand purchases. Motivation is defined as one’s will to pursue a goal [49]. Confidence, desire, and motivation are the consumer’s psychological states [11]. Therefore, we propose three new constructs: recommendation confidence, recommendation desire, and recommendation motivation. These constructs are related to recommendations and are affected by recommendation credibility. In this way, we propose that recommendation credibility has a positive effect on consumers’ three psychological factors of confidence, desire, and motivation. Thus, the positive effect of recommendation credibility on recommendation confidence, desire, and motivation is boosted.

Some previous research [5] shows that a highly credible source may strengthen the recommendation readers’ confidence in their judgment. Information consistency enhances information credibility and thus consumer confidence [50]. Furthermore, according to social comparison theory, people often make an upward comparison and choose a better (i.e., a more credible) target in order to improve their decision making [51]. Therefore, consumers seek credibility in a celebrity, always choosing one who commands a high degree of credibility [29] and associating their behavior with a credible source can improve confidence in their shopping performance.

The last two factors are desire and motivation. Regarding the former, recommendation credibility can shape a consumer’s desire, which in turn shapes the consumer’s behavior [52]. This complex process is especially compelling in social media communities, where social interactions involve a strong sense of belonging [53]. In our current study, in which desire is the conscious, stimulated pursuit by consumers of an ideal self-image, this complex process leading from recommendations to desires to behaviors is profound. One point on desire should be noted. As is widely known, the model of goal-directed behavior (MGB) is often used in the study of individual behavioral desire. MGB rests on the theory of planned behavior (TPB), but adds desire and emotion factors to the model. One form of behavioral desire is purchase desire, which has been shown to be significantly associated with consumers’ perception of celebrity credibility [34,54]. A celebrity recommendation of products in social media settings (the topic of our current study) involves such a formulation of desire: celebrities interact with their fans online, generating intense desire, “goal contagion”, and ultimately purchase intention [55].

Finally, recommendations perceived as credible can gain approval from consumers, thus boosting consumer confidence and consumer motivation [56]. Some studies indicate that trustworthiness has special influence on recommendation behaviors in the context of social interaction [18,43]. People may seek self-enhancement by projecting themselves as experts, and this expertise, when perceived by consumers as a trustworthy recommendation, can motivate them to engage in recommendation-based behaviors [27,43,57]. From these findings, we can reasonably infer that credible recommendations exercise a significant recommendation effect on consumers, whose beliefs and behaviors reflect celebrities’ recommendations.

Given the above research findings about how recommendation credibility stands in relation to confidence, desire, and motivation, we propose the following three hypotheses:

Hypothesis 3 (H3).

Recommendation credibility is positively associated with consumer confidence.

Hypothesis 4 (H4).

Recommendation credibility is positively associated with consumer desire.

Hypothesis 5 (H5).

Recommendation credibility is positively associated with consumer motivation.

2.5. The Relationship between Psychological States and Purchase Intention

Previous research [58,59] shows that consumer confidence in social media recommendations influences purchase intention because the consistency of information strengthens purchase intention by boosting confidence in the information [60]. Online information, in particular, seems to invite confidence. Proof of this is the fact that many people search online for an answer to a question that they may already have an answer to in their mind. This curious phenomenon suggests that, in some ways, searchers of information are less confident in their own knowledge than in online information. In addition, as noted, credibility translates into confidence: highly credible information, for example, generates confidence in the information [46]. Likewise, and as noted earlier, highly trusted celebrities who recommend a product generate consumer confidence in the product, in turn boosting purchase intention [59].

Belk, Ger, and Askegaard (2003) [61] characterized desire as the driving force behind consumers’ purchase behavior, which, by extension, involves purchase intention. Desire is very much like imagination insofar as they both involve a longing to satisfy an unsatisfied need, and in today’s world, many unsatisfied needs are products yet to be purchased; people always imagine a better life filled with purchased products, and because of this desirous imagination, purchase intentions take shape [49]. In social media, consumers can consult online reviews to act on these purchase intentions and to satisfy their underlying desire [62]. As we have seen, celebrity recommendations, when they enjoy congruence with products or consumers, can shape purchase intention. In sum, consumer imagination, online settings, celebrity recommendations, and congruence interact with each other to erect a causal bridge between desire and purchase intention.

The third psychological state, motivation, is a direct cause of behavior and incorporates willingness, interest, and readiness [49]. For the purpose of our current study, we have been considering motivation in relation to purchase intentions, particularly as they concern consumers’ motivation to realize an ideal self-image. Other motivations that guide purchase intention involve recreation, new trends, intellectual satisfaction, and physical stimulation [63].

Given the above research findings about how confidence, desire, and motivation stand in relation to the purchase intention of consumers, we propose the following three hypotheses:

Hypothesis 6 (H6).

Consumer confidence is positively associated with consumer purchase intention.

Hypothesis 7 (H7).

Consumer desire is positively associated with consumer purchase intention.

Hypothesis 8 (H8).

Consumer motivation is positively associated with consumer purchase intention.

To sum up, our research model is show in Figure 1 as follows:

3. Methods

3.1. Scale Development

All the measure scales in this study were adopted, developed, and finely modified on the basis of the survey scales presented in the previous literature. The items measuring construct of celebrity congruence with products (three items)/consumers (three items) were adapted versions of the measurement scale used by Japutra, Ekinci, and Simkin (2019) [8]. Based on analysis and derivation of the source credibility adopted from Chuan Luo et al. (2013) [5], we obtained five items of recommendation credibility. Same as above, the index used to measure purchase intention was revised from Wu and Wang (2011) [64] and Mikalef, Giannakos, and Pateli. (2013) [59]. Consistent with Japutra, Ekinci, and Simkin (2019) [8], respondents answered the items related to psychological states of purchase confidence (four items), desire (three items), and motivation (three items) in the recommendation credibility contexts that were adapted from the scales used by Raciti et al. (2011) [54], Wu, Ai, and Cheng. (2019) [11], and Loureiro, Plaza, and Taghian. (2020) [34]. Therefore,

Table A1. Measurement Items.

Construct	Code	Item	Adapted from
Celebrity congruence with product	CCP1	This celebrity is a mirror image of the product.	Japutra, Ekinci, and Simkin. (2019) [8]
	CCP2	The celebrity is fit for the goods.
	CCP3	This celebrity is consistent with how the goods would like to be.
Celebrity congruence with consumer	CCC1	This celebrity is consistent with how I see myself.	Japutra, Ekinci, and Simkin. (2019) [8]
	CCC2	This celebrity and I are in the same circle.
	CCC3	This celebrity is similar to me.
Recommendation Credibility	CB1	This celebrity recommendation is believable.	Chuan Luo et al. (2013) [5]
	CB2	The celebrity-recommended items are useful
	CB3	I believe that the information that I get from celebrity is reliable.
	CB4	The celebrity recommended is good.
	CB5	The celebrity recommending the goods is experienced.
Confidence	CD1	I am confident that the celebrity-recommended goods can satisfy everything I need.	Raciti et al. (2011); [54] Wu, Ai, and Cheng. (2019) [11]
	CD2	I am confident that I am good at purchasing the items that the celebrity recommends.
	CD3	I highly approve of the products recommended by celebrity
Desire	DR1	The recommendation of the celebrity enhances my desire to buy goods.	Wu, Ai, and Cheng. (2019) [11], Loureiro, Plaza, and Taghian. (2020) [34]
	DR2	I very much desire to buy the goods of celebrity recommendation in the near future.
	DR3	After celebrity recommendation, I imagine that I own the goods.
Motivation	MV1	I feel motivated to buy the goods of celebrity recommendation.	(Raciti et al. (2011) [54] Wu, Ai, and Cheng (2019) [11]
	MV2	I take a greater interest in the goods of celebrity recommendation.
	MV3	I am inspired by celebrity recommendation to buy the goods.
Purchase intention	PT1	Following celebrity recommendation, I would consider buying the product.	Wu and Wang, (2011) [64], Mikalef, Giannakos, and Pateli. (2013) [59]
	PT2	The probability that I would consider buying the product of celebrity recommendation is high.
	PT3	My willingness to buy the product is high.

(Appendix A) shows the detailed items of the measurement questionnaire, which used a five-point Likert scale ranging from 1 to 5.

To ensure the accuracy of translated items, we had a bilingual scholar translate the questionnaire from English into Chinese and afterward had another bilingual scholar check the translation.

3.2. Sample and Data Collection

To verify the universal applicability of the constructed model of this study, data from two different channels are integrated for analysis. Part of the data comes from the Tiktok platform in mainland China. Tiktok is a public and individual enterprise where users can upload and display their self-made short videos on the platform, thus attracting many netizens to participate in activity on Tiktok. Many individuals and companies hope to adopt the Tiktok platform to promote, sell, and reposition their company’s brands and products, and then to increase personal popularity, and divert traffic to their official websites for selling products.

We collected questionnaire data on Tiktok after celebrity Luo Yonghao’s live broadcast on April 1; the respondents are the group watching the live broadcast. Luo Yonghao was born on 9 July 1972. He was popular for his humorous teaching style and highly idealistic temperament, even creating a peculiar cultural phenomenon named “laoluo quotations”. He is an interesting person who is famous for his “tough life” and has written autobiographies: My Struggle and Life Goes On and On.

Luo Yonghao registered in Tiktok on 25 March 2020; within a week, the number of fans was up to 7.316 million, of which men accounted for 70.81%. The population of these fans is: 32.3% aged between 18 and 24 years old, 26.9% aged between 25 and 30 years old, 15.8% aged between 31 and 35 years old, and 10% aged over 36 years old. Luo Yonghao made the first live broadcast on TikTok to recommend products, with total sales of over 28 million dollars, over 900,000 orders, and over 48 million views.

Another part of the data was collected from the “Wooden fish” community on the IG (from Taiwan’s Instagram) platform. At first, the celebrity was just a narrator for animation and TV shows, sharing relevant background knowledge. Gradually, young people who like anime became fans of the commentator; they often pay attention to the commentator’s dynamic role, and the commentator became an online celebrity. She also shares her animation products on the IG platform, which are bought by the fans, promoting the celebrity to develop new products and sell them on the IG platform. Therefore, we conducted a questionnaire survey on community members.

In order to ensure the validity and integrity of the questionnaire response, only one questionnaire could be completed through the questionnaire link opened through the mobile phone, and the questionnaire is designed to be submitted only after all the items have been answered. Meanwhile, the cost of filling in the questionnaire is provided to the respondents. Each questionnaire was controlled at CNY 20 (USD 3.4), and had approximately 15 days to finish (between 1 April 2020 and 15 April 2020). Finally, we collected 150 valid questionnaires from Tiktok and 147 from IG).

The demographic characteristics of respondents are shown in detail in

Table 1. Demographic characteristics of the sample.

Measure	Items	Frequency	Percentage
Gender	Male	116	39%
	Female	181	61%
Age	Below 20	20	6.73%
	21–25	165	55.55%
	26–30	56	18.85%
	31–35	32	10.77%
	Above 36	24	8.1%
Have experience with Youtube or Tiktok	Yes	297	100%
	No	None	0%
Average time spent using YouTube or Tiktok “per day”	Under 1 h	208	70.04%
	Between 1 h to 2 h	75	25.25%
	Above 2 h	14	4.71%

. Interestingly, 116 respondents were male (39%) and 181 were female (61%). As for age, 165 (55.55%) were between 21 and 25, and 56 (18.85%) were between 26 and 30—most of the fans. All the respondents have experience with Youtube or Tiktok. In total, 70.04% responded in one hour, and 25.25% within 1 to 2 h. Judging from the demographic information provided, the questionnaire content is suitable for this study.

3.3. Measurements

3.3.1. The Indices of the Measurement Model

In modern research, scholars often use PLS (Partial Least Squares, a component-based structural equation modeling (SEM) technique) to examine research models and hypothesis testing. Due to its advantages, PLS is often recommended for small samples, especially for complex models and exploratory studies. In order to determine the relative fit of a structural model using the PLS procedure, researchers must examine the indicator reliability. The indicators for checking the reliability of a single item are the individual construct Cronbach’s alpha (α) value and composite reliability (CR) measures, which both exceed 0.70, with loadings exceeding 0.70. In addition, Fornell and Larcker’s (1981) [65] criteria are usually used to measure the discriminant validity of the model in PLS.

3.3.2. Structural Model

The structural model shows the causal and correlational links among latent variables in a theoretical model, and smart-PLS can also be used for analysis. Through the bootstrapping program to establish the path for 1000 times of statistical verification, the results verify the relationship between the model and the variables.

4. Results

The data analysis results are presented as follows: in this study, as shown in

Table 2. Reliability and Validity.

Construct	Variable	Factor Loading/Coefficient	Cronbach’s α	Composite Reliability (CR)	AVE
Celebrity congruence with Product (CCP)	CCP1	0.941	0.921	0.950	0.864
	CCP2	0.920
	CCP3	0.927
Celebrity congruence with Consumer (CCC)	CCC1	0.902	0.874	0.923	0.799
	CCC2	0.895
	CCC3	0.885
Recommendation Credibility (RCB)	RCB1	0.881	0.915	0.937	0.748
	RCB2	0.853
	RCB3	0.881
	RCB4	0.821
	RCB5	0.886
Confidence (CD)	CD1	0.881	0.906	0.941	0.842
	CD2	0.935
	CD3	0.935
Desire (DR)	DR1	0.855	0.891	0.932	0.822
	DR2	0.931
	DR3	0.931
Motivation (MV)	MV1	0.875	0.875	0.923	0.800
	MV2	0.904
	MV3	0.904
Purchase intention (PT)	PT1	0.790	0.787	0.876	0.702
	PT2	0.883
	PT3	0.838

, all factor loads (from 0.821 to 0.941), Cronbach’s alpha (α) value (from 0.787 to 0.921), and composite reliability (CR) (form 0.876 to 0.950) measures exceed 0.7, which indicates that the model is reliable and valid. The average variance extracted (AVE) of all the constructs needs to be above 0.5, and the research measured values ranging from 0.702 to 0.864, which meet the internal consistency criteria requirements.

Table 3. Discriminant Validity (Fornell–Larcker Criterion).

	CCC	CCP	CD	DR	MV	PT	RCB
CCC	0.894
CCP	0.711	0.929
CD	0.641	0.724	0.918
DR	0.596	0.615	0.741	0.906
MV	0.538	0.557	0.652	0.734	0.894
PT	0.516	0.534	0.619	0.670	0.740	0.838
RCB	0.621	0.682	0.813	0.679	0.603	0.533	0.865

Note: The logic of the Fornell–Larcker method is based on the idea that a construct shares more variance with its associated indicators than with any other construct, and this means the model has discriminant validity.

indicates that the discriminant validity was qualified.

Table 4. Model Fit Evaluation.

Discrepancy	Value	HI95	Conclusion
SRMR	0.049	0.106	Support
dULS	0.668	3.122	Support
dG	0.479	0.616	Support

Note: SRMR refers to standardized root mean square residual; dULS refers to unweighted least squares discrepancy, dG values mean geodesic discrepancy.

shows the following goodness of fit of the research model, when the SRMR value was below the recommended threshold of 0.08, and all discrepancies were below 95 % of the bootstrap quantile (HI95), which suggests a good fit between the research model and the data.

Figure 2 shows that the structural model explains 50.3% of the variance explained by the two dimensions of recommendation credibility. Similarly, the perceived recommendation credibility explains the variance of psychological states of 66% confidence, 46.1% desire, and 36.3% motivation. Lastly, psychological states explain the 59.2% variance of purchase intention. The estimation of causal path parameters from structural equation modeling enabled us to examine eight hypotheses (see

Table 5. Hypothesis Results.

Predictor Variable	Dependent Variable	Hypothesis
Celebrity congruence with product →	Recommendation credibility	H1—supported
Celebrity congruence with consumer →	Recommendation credibility	H2—supported
Recommendation credibility →	Confidence	H3—supported
Recommendation credibility →	Desire	H4—supported
Recommendation credibility →	Motivation	H5—supported
Confidence →	Purchase intention	H6—supported
Desire   →	Purchase intention	H7—supported
Motivation  →	Purchase intention	H8—supported

Note: “→” refer to the influence of predictor variable on dependent variable.

). Our analysis shows a positive correlation between online celebrities and product consistency for recommendation credibility (β = 0.486, p < 0.001). Similarly, Internet celebrities’ congruence with consumers is positively associated with recommendation credibility (β = 0.275, p < 0.001), supporting H1 and H2. We analyzed the relationship between recommendation credibility and consumers’ psychological states. It can be seen from Figure 2 and Table 3 that credibility is significantly positively associated with recommendation confidence (β = 0.813, p < 0.001), recommendation desire (β = 0.679, p < 0.001), and recommendation motivation (β = 0.603, p < 0.01), all confirming H3, H4, and H5. We hypothesized that confidence would be positively associated with purchase intention, and the results bear out our hypothesis: confidence is, in fact, significantly positively associated with purchase intention (β = 0.150, p < 0.05); that is, the results support H6. Furthermore, we observed that our results support H7: desire is positively associated with purchase intention (β = 0.190, p < 0.05). In H8, we hypothesized that motivation would be positively associated with purchase intention. The results back up our hypothesis (β = 0.502, p < 0.001), motivation is significantly positively associated with purchase intention.

5. Discussion

This research is meaningful for both scholars and practitioners, as the findings enrich the research in these fields. This study, bases on source credibility, establishes a new research model and examines how the influence of the two dimensions of consistency affect recommendation credibility, as well as how recommendation credibility affects psychological states and how psychological states have a mediating effect on purchase intention.

First, the hypothesis is that consistency improves recommendation credibility, which can realize product purchase by enhancing the followers’ confidence, desire, and motivation in recommendation credibility. The tested model is consistent with the study from [5], and they believe that consistency affects a consumer’s attitude toward products.

The results of H1 and H2 show that a relationship exists between, on one side, celebrity congruence with products/consumers and, on the other side, consistency’s effect on recommendation credibility. The results of H1 align with the results uncovered by Felix and Borges (2014) [28], who found that when consumers perceive consistency between celebrities and their promoted products, the promotions acquire more credibility. At the same time, it can reasonably conclude that H2 aligns with the results in Choi and Rifon (2012) [7], who confirmed two assertions: first, celebrity congruence with consumers can strengthen their favorable impression of the given celebrity’s product endorsement, and second, products recommended by celebrities have higher credibility and attract greater purchase intention than do products recommended by non-celebrities [12]. Consistency seems to enhance the credibility of recommendations.

In addition, this is the first time the consistency versus recommendation credibility questionnaire is used to study the psychological states in perceived high reliability for purchase in the context of social media. The results can be used as a reference for marketing managers to use celebrities as a marketing strategy.

The results tell us recommendation credibility is positively correlated with confidence (H3) and desire (H4). The conclusion is that the information receiver has confidence in the recommended products, which creates the desire to buy. These two sets of results are consistent with the research consequence of Wu, Ai, and Cheng (2019) [11]. It also found that perceived motivation (H5) is positively correlated with recommendation credibility. This observation is consistent with the research of Chandler and Chen (2015) [66] and Elvira et al. (2020) [4], which found that an increase in shopping motivation correlates to an increase in shoppers’ willingness to buy recommended products.

Our results suggest that reliability and attractiveness have positive effects on consumer’s purchase confidence, purchase desire, and purchase motivation, which is the same as that of [56,57,67]. A large number of previous studies have shown that consumers have a desire to purchase attractive internet-recommended products [68]. Scholars have verified that attractive internet celebrities arouse consumers’ intrinsic desire for the realization of their ideal self-image and generate purchase motivation [15,57].

Finally, regarding H6, we noted that confidence is positively correlated with purchase intention. If consumers have confidence in the products, it can be understood as the products recommended by celebrities are of high quality, which can effectively persuade consumers. Our study also supports that the psychological states of desire and motivation are positively correlated with purchase intention. All individuals instinctively seek to improve themselves, and consumers’ perception of reliable information transmission can help attach both desire and motivation to purchase intention. In fact, these two psychological states are critical to purchase intention, and our results are in keeping with those of Wu, Ai, and Cheng (2019) [11]. Finally, the study proves that psychological states can predict consumers’ purchase intention. This result promotes our understanding of the relationship between the premise of consistency and consumer behavior.

6. Conclusions

The purpose of this study has been to explore the effect that influencers who are recommending a product can have on consumers’ purchase intention. Regarding recommendation credibility, our results suggest that the persuasiveness of information affects consumers’ psychological states and consumers’ purchase intention. This finding is of theoretical significance and practical significance, particularly in the field of marketing.

Our study points to a positive correlation between celebrities’ consistency with a product and the credibility of the celebrities’ product recommendation. Our study also points to a positive correlation between celebrities’ consistency with consumers and the credibility of the celebrities’ product recommendation. We observed that the credibility of a recommender and the credibility of the recommendation are positively correlated with each other.

There are differences from previous studies on source credibility. Unlike them, we have examined the influence of recommendation credibility on the psychological states of consumers. Other social media research focuses on source credibility, information usefulness, information adoption, and purchase intention. Prior to our current study, none of the literature addressed the influence that source credibility has on psychological states. Moreover, few studies previously addressed the influence of psychological states on purchase intention. Our study fills these two major gaps in the research.

The third and final major contribution of this study to the literature is our analysis of the relationship between psychological states (namely confidence, desire, and motivation) and purchase intention. At present, very little research discusses these psychological states relative to source credibility as it exists on social media. Psychological states are of great significance to the study of purchase behavior and purchase attitudes and, consequently, purchase intention.

Additionally, our research has certainly hinted and argued for the selection of influencers. Negative news from influencers is frequent and even false and fraudulent, which greatly undermines consumer confidence. Therefore, the influencer marketing industry must consider how to choose the right influencer. The influencers we selected have certain characteristics, that is, their personal experience proves that they are a reliable person, and their personality attracts some followers, leading them to trust the products that they recommended. No matter the cooperation of enterprises or countries, the reliability and credibility of the influencer’s personality is the prerequisite and the basis of long-term cooperation. In order to purify the market environment of influencers and promote the orderly development of honest cooperation, it is necessary to formulate certain punishment policies for those who conceal fraudulent moral hazards.

7. Practical Contributions to Marketing

We firmly believe that our current study is applicable to actual marketing operations and serves as a roadmap for marketing environments. The celebrities we chose are not celebrities with a special background, but became a celebrity by attracting the attention of the majority of netizens through expertise or personal growth experience and personality characteristics. In becoming a celebrity, they can be presented in many countries and regions without regional specificity; therefore, the results of this study can be used across countries, regions or ages.

First of all, the research results strongly suggest that marketing departments should never underestimate the cruciality of social media to the relationship between brand and consumers. Marketing personnel should carefully design and manage product information that appears on social media. Brand and product information should be presented from highly credible sources; especially in the network world, high source credibility is extremely important. Our research proves that high source credibility of information enhances consumers’ purchase psychological states and purchase intention.

Second, an alliance between a firm and an influential celebrity substantially improve advertising’s contributions to products, sales, and brand awareness. Research shows that firms should give strong consideration to the possibility of collaborating with attractive celebrities for marketing purposes. To select the best celebrity collaborators, a marketing department should ensure that the celebrities “match” the recommended product. The brand spokesperson who is popular should be scrutinized in terms of reliability, professionalism, physical appearance, personality, and fan interaction. Celebrities have a good reputation and positive energy, which is especially important. Influencers’ personality determines who they are and how they are perceived by consumers. It is necessary to evaluate whether influencers match the product connotation from individual growth experience, moral and ethical cognition, and attitude toward money

In this way, product information and brand can be implanted into consumers’ minds through celebrity recommendation.

Third, marketing departments must not ignore the importance of pre-purchase psychological states. If the consumer’s purchase confidence, desire, and motivation are lower than expected, the marketer should first consider whether the recommendation credibility is low. Recommendation credibility affects consumers’ psychological states and purchase intention. The higher recommendation credibility is, the greater the impact on consumers’ purchase behaviors. At this time, the strategy is to improve the high credibility of product recommendation.

8. Research Limitations and Future Research Directions

There are two major limitations to this study. First, the number of questionnaires on which we relied limited the overall reliability and generalizability of our results. Furthermore, our study would have benefited from a wider array of sources for data. In this regard, Instagram would have been a logical source.

For future research, we will extend the current study’s research model to different social networks and expand the model’s scope. In addition, our future research will consider exogenous factors and the potential roles of intermediary variables in, for example, the relationship between psychological state and purchase intention.