Examining the Antecedents of User Donation Intentions Toward Social Media Articles: Moderation Effects of Social Contagion

Abstract

Donations to articles on social media, as a new behavior, have been trending in recent years. Unlike donations to a charitable and nonprofit organization or victims, donations to social media articles have been accorded minimal attention from academic researchers. From the stimuli–organism–response framework, this study proposed a model to investigate the factors that influence the donation intentions of users on social media toward articles. Our results demonstrate that the credibility of the article determines users’ donation intentions. The results also indicate that the perceived value (usefulness and enjoyment) of the article mediates the effects of article credibility on users’ donation intentions. The social contagion nature of the article is also proven to moderate the magnitude of impacts on donation intention by users’ perceived usefulness, perceived enjoyment, and perception of article credibility.

Keywords

social media donation intention credibility social contagion social media donations

Introduction

Social media is the core of creating and distribution of information through user-generated content (UGC). The popularity of social media has attracted not only the participation and contribution of businesses but also researchers to delve into the motivation surrounding their use (Kapoor et al., 2018). Social media platforms possess dynamic functions, like the possibility for users to create channels and groups, and share contents rapidly with a significant number of an audience (Muller & Peres, 2019); these possibilities make social media attractive to users.

In China, some social media sites recently came out with a feature called Donation, which allows users to donate to UGC of interest. These donations provide the content creators with an incentive and, in turn, encourage UGC and opinion formation that are the primary goals of almost all social media (Wan et al., 2017). Users, after having access to an online content, make donations to the content on the basis of service and knowledge perceived, hence providing the content creators with monetary or nonmonetary reward. Although the donation feature is not accessible in social media websites globally, this functionality is gaining much popularity in the social media setting in China. WeChat (a famous social media platform in China) users subscribe to official accounts on WeChat that is responsible for publishing a variety of articles (stories, health topics, news, motivational speeches, comedy, and educational letters) from content creators. Upon accessing an article and finding it beneficial, users decide to donate to the article by clicking on a function at the bottom of the article called “Like the author”; this function redirects the user to a pop-up option with different amounts of money that users are willing to donate, as illustrated in Figure 1. It is therefore prudent for researchers and social media developers to understand the driving force responsible for donation to comprehend the entire range of social media possibilities (Wan et al., 2017).

Figure 1.

A WeChat article talking about life after the corona virus epidemic for people of Wuhan with 356 likes (donations).

According to Wan et al. (2017), users’ donation are mostly based on their attachment to the content creator. In addition, Y. Chen et al. (2019) argued that donation intentions (DIs) are influenced by social presence, trust, user attitude, and perceived behavioral control (Y. Chen et al., 2019). Contributing to the social media donation literature, L. Liu (2019) investigated perceived value as an influencing factor for users’ donation behavior (L. Liu, 2019). A study by Jia et al. (2018) also examined audiences’ reaction to user-generated videos on a YouTube-like UGC website designed to include social media functions, like user followership, user comments, and virtual reward money donations (Jia et al., 2018). Yet there is less attention on the credibility of a content as a motivational factor for users’ DIs. In addressing this phenomenon, the study solicits the stimuli–organism–response (SOR) model to examine how the credibility of an article influences users’ intentions to donate to such articles. The SOR model states that stimuli (characteristics of the article) always initiate a response (DIs) through an organism (the cognitive and internal evaluations of the stimuli available to the user).

A study by Leenders (2002) argues that individuals carefully consider the opinions and behaviors displayed by others, coupled with the consideration of other opportunities and constraints, to form their own behaviors and opinions. He termed this process “social contagion” (SC; Leenders, 2002). In the context of this study, an article may be perceived as credible, useful, and enjoyable by users; however, their willingness to donate might be determined by the strength of opinions and by behaviors displayed by others (SC; Tussyadiah, 2012). This study, therefore, specifically examines the impact of article credibility (AC) on DIs and the mediating roles of perceived usefulness (PU) and perceived enjoyment (PE), while examining the moderating effect of SC.

This research furnishes social media platform operators and content creators with practical understanding of the factors that play a significant role in users’ decisions to donate to articles. Donation in social media is a new feature and, thus far, several studies have focused on donation behaviors toward charity and nonprofit or victims.

Considering the reasons presented above, this study attempts to answer the following questions:

Research Question 1 (RQ1): What characteristics of social media articles influence users’ attitude formation?

Research Question (RQ2): How does attitude formation determine users’ donation behavior toward social media articles?

Literature Review

Donation in Social Media

Donation, as a general term, is the giving of charity, whether monetary or nonmonetary, to nonprofit organizations or victims who do not have a close relationship with the donor (Isaac et al., 2015). Numerous charity and nongovernmental organizations have utilized social media platforms to reach a great volume of potential donors (Q. Li, 2017). Social media platforms support charity organizations, for instance, kickstarter.com and Donorschool.org, by providing the ability to manage an online crowdfunding platform to raise funds for charity projects (Althoff & Leskovec, 2015). Social media has also promoted electronic philanthropy; this primarily refers to the generosity that happens in the cyberspace managed by charity organizations such as NetAid, HungerSite, and Giving.org.

Donation, in this study, is a feature in social media in the form of a “like button” with cash donation attached. It is intentionally placed below an article on social media, whereby readers or users who have accessed the article can click to donate an amount of money to the author of the article (Horwitz, 2015). These donations support the authors of such articles to create more contents of such sorts as the users find it enjoyable or useful (X. Liu et al., 2017). The donation feature in social media possesses identical characteristics as busking, whereby a performer performs an act in the street or a public place, and then the audience voluntarily gives an amount of money. Busking is characterized by a delivery of service (performance) by a performer followed by a donation in the form of money by the receiver, based on the performance; these two serve as a joint product (Amundsen, 2019). Similarly, donation as a feature in social media also revolves around a joint product: service and gifting, whereby an author publishes an article on a social media platform (service) and the users form donation decisions toward the author based on their internal evaluation of the article. (Schlosser & Levy, 2016). Wan argues that donation in social media subsumes two built-in attributes: charitable and consumptive attributes (Wan et al., 2017). From the perspective of consumptive attribute, donors pay for consumption of a specific service, for instance, an article published on social media by an author.

Research has distinguished individuals’ motivation to donate into two divisions: intrinsic motivation, whereby people donate to experience the satisfaction of positive emotions (warm effect) associated with helping people, and extrinsic motivation, whereby people donate to receive intangible rewards such as praise and recognition (Neumayr & Handy, 2019; Savary et al., 2015). Examples of extrinsic motivation are as follows: an individual may get front-row seats at an event for donating to an art organization or an individual who purchases building materials as a donation to a community project can have his name on a plaque as a donor. Wan et al. (2017) examined donation to content creators (involving both consumptive and charitable attributes) on live streaming social media platforms (yy.com), whereby live broadcasters’ login to dance, play video games, or sing while users stream live and donate based on their attachments to these broadcasters. The donations made in this case are in the form of virtual gifts like coins, stars, rockets, and so on.

This study considers the consumptive attributes of donation behaviors and posits that users’ intention to donate does not only depend on their attachments to the authors, but users may also be attracted to articles through their external characteristics, without even having any devotion to the author (Carlin, 2014). Donors donate to authors as payment for services consumed, especially on social media platforms where content creators are not broadcasting live (Y. Chen et al., 2019). It can, therefore, be argued that external factors like users’ perception of AC (drawn from users’ dependency on the medium and the informative nature of the material) can attract a user. These factors lead to the formation of attitudes by the user and, in turn, affect their DIs. To understand the effects of AC on DIs through consumers’ attitude formation, we employ the theory of persuasion and the SOR model for our examination.

AC and DIs

The credibility concept dates to the writings on rhetoric by Aristotle and, specifically, on his notions of pathos (appeal based on emotions), logos (based on logic or reason), and ethos (appeal based on the character of a speaker; Varpio, 2018). Recent accounts of credibility describe it as the believability of a source and it relies significantly on individual perceptions and the trustworthiness of the information source, as deciphered by the information receiver (Elhadidi, 2019; R. Li & Suh, 2015). The definition has guided research on credibility in diverse academic research areas. Studies in psychology and communication focus on source credibility and typically conceptualize credibility as the believability of a speaker, which is closely related to Aristotle’s notion of ethos. Information systems (IS) research has focused instead on the credibility of information, where the emphasis is placed on the believability of the message rather than the speakers, which is also closely aligned with Aristotle’s notion of logos (Varpio, 2018). Paying attention to these varying conceptualizations is crucial as determinants of credibility in the online realm may rest on information source evaluations, the message itself, or on a combination of the source and the message. Social media is consistently presenting remarkably inexpensive ways of disseminating information, usually buckled with incentives for doing so (Kakol et al., 2017).

DI is a behavioral construct that refers to a person’s likelihood or subjected probability that he or she will donate (Y. Chen et al., 2019). Studies on donation have examined the motives behind people’s willingness to donate by applying the theory of motivation. Such studies have classified motivation into intrinsic and extrinsic categories. Intrinsic motivations represent the positive emotions related to helping others; this motivation is similar to the “warm effect” (Kashif et al., 2015; Ye et al., 2015). Most people donate to charity organizations or victims, mainly due to the desire to help others and the feeling of satisfaction experienced when they perform a good deed. On the contrary, extrinsic motivation explains the willingness to give based on considerations of external factors (Erlandsson et al., 2018). Intentions to donate to articles, in this study, are strictly related to the extrinsic motivation category where users consider the characteristics and assess the usefulness and enjoyment of the article after engaging with it to determine their intention to donate.

According to the SOR model, stimuli are characteristics external to the individual and the individual initiates a response based on the exposed traits (X. Chen et al., 2017). This study conceptualizes AC as the stimuli in the model, which can initiate a response (DI) from an individual. When a reader of an article tags it to be credible, the more likely he is to form intentions to donate to such article (Chin et al., 2020; Thomas & Johnson, 2016). Thereby, the credibility of an article can be said to have a significant direct effect on users’ DIs. Thus, we hypothesize that

Hypothesis 1 (H1): DI is directly affected by the credibility of the article.

The Mediating Effects of PU on AC and DI

PU is the degree of personal conviction that involvement with activity will improve a user’s performance (Bonn et al., 2016). The usefulness of the event is related to the perception of the individual (Mou et al., 2017). Diop et al. (2019) posited that PU positively affects user intentions (Diop et al., 2019). Furthermore, Y. C. Cho and Sagynov (2015) studied the determinants of users’ intention in the online environment and argued that intentions are determined by users’ PU of a service received (Y. C. Cho & Sagynov, 2015). Drawing from the SOR model, PU represents the organism (emotive and cognitive processes) influenced by the stimuli and decides the response of the individual. To determine the usefulness of the article experienced by users, the extent to which the users consider the article useful in sharing ideas, improving their general productivity, and also boosting the performance of their daily activities are measured (Hamid et al., 2016). If a user perceives an article to be useful, regarding knowledge sharing, performance boosting, and improvement in overall production, they are more likely to donate to such articles. We, therefore, hypothesize that

Hypothesis 2 (H2): PU mediates the relationship between AC and DIs.

The Mediating Effects of PE on AC and DI

PE is a construct of individual differences generated from intrinsic motivations, and it is a significant factor that affects individuals’ decisions about online network behaviors (Aksoy & Basaran, 2017). Liao et al. (2008) defined PE as the degree to which users believe service to be exciting and associate the service with enjoyment; therefore, influencing their behavioral intentions toward such services (Liao et al., 2008). Chao (2019), after exploring the effects of PE in his study of understanding the factors that determine behavioral intentions, supports a significant positive impact of PE on users’ intentions (Chao, 2019). PE, as a construct, represents the organism (the psychological and cognitive evaluation of the article), based on the SOR model. The relationship between PE and users’ intentions is supported by many information systems (IS) studies (E. Cho & Son, 2019; Rouibah et al., 2016). PE is therefore considered as a necessary construct to investigate users’ intentions on social networking sites (SNS) as the use of SNS is usually viewed as enjoyable (Yang et al., 2017).

In addition, users consider articles that are interactive and relatable as enjoyable and therefore attach favorable judgments to such articles (Horning, 2017). This study employs the PE construct to explain the pleasure that users experience by interacting with the article and their DIs based on PE. We, therefore, hypothesize that

Hypothesis 3 (H3): PE mediates the relationship between AC and DIs.

The Moderation Effects of SC

SC is the spread of impact or behavior from one crowd participant to another, where the actions of acts of one person serve as a factor for the imitative responses or attitude formation of another (Ferrara & Yang, 2015). The definition possesses the merit of clarifying and focusing on the observable contagion phenomenon. SC in the context of social media refers to the social transmission of the idea, culture, and behaviors on a social media platform. Voluntary contributions are prevalent in online communities; this is manifested in product reviews from millions of people daily, product rating, writing comments, and clicking the “like button” under a product or service by users of the online platform who in one way or the other have engaged with such product or service. A revolutionary study by Fowler and Christakis confirms that altruistic behaviors can fluctuate through the social network (Fowler & Christakis, 2010). SC consists of two mechanisms: generalized reciprocity (GR), explaining that people who exhibit altruistic behaviors have benefited from kindness shown by a stranger, and third-party influence (TPI), revealing that individual’s altruistic behavior is as a result of witnessing a generous act of a third party (Vishwanath, 2015). In the context of social media donations, researchers consider the TPI mechanism. (Lin, 2011), who studied the influence of product contagion on users’ evaluations, show that consumers have a fixed impression about services after their evaluations but that these impressions are influences after being exposed to a third party’s attitude toward the same service, thereby affecting their final behavioral intentions toward such services (Lin, 2011). Ferrara and Yang (2015) studied contagion effects on twitter and found that a common mechanism of contagion regulated both negative and positive intentions of users toward a tweet.

SC has been used as a different term in previous studies. According to the theory of reasoned action (TRA) and theory of planned behavior (TPB), SC is referred to as subjective norms which reflect how users are influenced by their perception that society expects them to perform a given behavior (Hasan et al., 2016; Wei et al., 2016). Venkatesh and Davis (2000) examined an extended version of the technology acceptance model (TAM) and presented SC as “Social Influence” which is a predictor of technology adoption (Venkatesh & Davis, 2000). The authors argued that users’ behavioral intention based on perceived value (PU and PE) is affected by the behaviors displayed by others (SC; Venkatesh & Davis, 2000; Zhou & Feng, 2017). A study by Beldad and Hegner (2018) examined German users’ willingness to continue using a fitness app based on their PU and revealed that behaviors and opinions of others (SC) about the fitness app affected users’ willingness to continue using the app (Beldad & Hegner, 2018). Wang et al. (2020) tested the moderating effects of emotional contagion on social media and posited that positive emotional contagion has higher moderation effects on the relationship between the use of social network for teams and perceived value, whereas negative emotional contagion has no significant effect on the relationship (Wang et al., 2020).

In the context of donations to social media articles, users or readers will be willing to donate to an article based on the credibility and the perceived value (PU and PE) of the article. However, the behavioral reactions and opinions (article likes, comments, number of views, and number of donations made by others) displayed by others toward the article will affect users’ intentions to donate despite users’ perceived credibility, usefulness, and enjoyment of the article.

Contents shared online have some features that promote SC; common among them are the “share,” “like,” “view,” “comments,” and most importantly, “donate” features. When a user likes, shares, comments, or donates to an article on a social media platform, it makes it contagious. Other users see this attitude and build trust in the sense that third party’s positive attitudes toward the material affects their response to the article after evaluation. This study, therefore, posits that the SC state of a social media article will moderate the relationships between users’ PU, PE, AC, and DI. Hence, we hypothesize that

Hypothesis 4a (H4a): SC moderates the effect of PU and DI.

Hypothesis 4b (H4b): SC moderates the effect of PE and DI.

Hypothesis 4c (H4c): SC moderates the effects of AC on DI.

Method

Participants and Setting

This study solicited data from users of the “WeChat” software application by using convenience sampling method. The 2017 WeChat annual report shows that 60% of all users open the app more than 10 times daily, with 40% seeking content through the official accounts. In addition, 80% of all users are reported to be following official accounts (which are the hosts for authors’ articles) and 41% do so intending to get information (Brennan, 2017). WeChat was therefore selected for our study because it is a leading social media platform in China, with a successful business model that incorporates the donation feature. All WeChat users are members of one group or more (social and academic) where people with unified interest share information and chat regularly.

Data Collection Procedure

A link to the research questionnaire was sent out to these WeChat groups for a 3-month period (October 2019–December 2019) and members (normally made up of people aged 18 years and above) were asked to click on the link to respond to the questionnaire if the respondents have ever donated to an article. Respondents who have never donated to an article were not allowed to continue their responses after answering “NO” to the introductory question (Have you ever donated to an article on WeChat?). The valid obtained responses were 306 out of 372. Figure 2 is the conceptual framework representing the research hypothesis based on the SOR framework.

Figure 2.

The conceptual framework for the research hypothesis.

Measures

This study adopted multi-item perpetual scales based on a 7-point Likert-type scale possessing two anchors (strongly disagree and strongly agree) in measuring all constructs in the study. DI measured users’ willingness to donate to the content and likelihood of donating to similar contents in the future based on the credibility of the article and PE and usefulness (Wan et al., 2017; Ye et al., 2015). PE measured the extent to which users find the content entertaining, pleasant, and enjoyable (Zhou & Feng, 2017). Items that measured PE are reformed from Chao (2019) and Zhou and Feng (2017). PU variable was measured based on the idea sharing, productivity enhancement, and performance enhancement attributes of an article perceived by the user or the reader (Scherer et al., 2015). We reformed items that measured PU from Scherer et al. (2015) and Zhou and Feng (2017). The items weighing AC are based on Kakol et al. (2017) and R. Li and Suh (2015); the variable was measured based on the level of a users’ dependency on the WeChat platform to deliver resources, information, and logical contents needed in daily life (Kakol et al., 2017). Finally, we adapted elements measuring the moderator and SC from Ferrara and Yang (2015) and Zeng and Zhu (2019); the moderating variable measured the extent to which other users’ actions toward an article (number of likes gained by the article, number of subscriptions of the article source, and number of donations made by other users) affected a respondent’s intentions to donate (Zeng & Zhu, 2019).

Data Analysis and Results

The Kaiser–Meyer–Olkin (KMO) statistic was tested to ensure that the data were appropriate for factor analysis; the KMO values for each variable in the study were above 0.6 (DIs: 0.692, AC: 0.737, PU: 0.738, PE: 0.757, and SC: 0.761), indicating adequate sampling and data for factor analysis (Kaiser, 1974; Wan et al., 2017).

The model of the research was estimated using the traditional covariance-based structural equation modeling technique. A confirmatory factor analysis was conducted and all the indicators loaded on the expected factors and factor loadings exceeded 0.6, suggesting good convergent and discriminant validities (Wan et al., 2017), as presented in Table 1. We tested the common method bias using Harman’s single-factor test (Aguirre-Urreta & Hu, 2019). The most significant factor accounted for 43% of the variance and, although the value is below 50%, another test was conducted to ensure the absence of common method bias by using (Podsakoff et al., 2003) method. Two regression weights were estimated: first, by including a common latent factor (SR with common latent factor), and, second, by excluding the common latent factor (SR). By comparison, we found that the differences between all the estimates and all values fell in the accepted threshold of <0.20, as presented in Table 1 (Fuller et al., 2016). This indicates that there is an absence of measurement errors that could threaten the validity of conclusions about relations between measures.

Table 1.

Descriptive Statistics, Validity and Reliability, and Confirmatory Factor Analysis.

Constructs	Items	M	SD	Loadings	AVE	CR	α	SR-SR (CLF)
Article credibility					0.841	.941	.939
	AC1	5.482	1.336	0.874				0.002
	AC2	5.541	1.270	0.982				0.012
	AC3	5.622	1.234	0.893				−0.003
Perceived usefulness					0.854	.946	.945
	PU1	5.151	1.897	0.871				0.001
	PU2	5.162	1.894	0.944				−0.002
	PU3	5.224	1.834	0.962				0.002
Perceived enjoyment					0.800	.923	.923
	PE1	5.261	1.761	0.863				0.01
	PE2	5.342	1.687	0.894				0.011
	PE3	5.434	1.637	0.931				0.009
Social contagion					0.875	.954	.951
	SC1	5.391	1.634	0.872				0.11
	SC2	5.533	1.498	0.991				0.16
	SC3	5.562	1.436	0.942				0.105
Donation intentions					0.914	.955	.952
	DI1	5.893	0.912	0.901				−0.009
	DI2	5.861	0.929	0.988				0.072

Note. SR-SR (CLF) ≤0.2 represents unlikely existence of common method bias. AVE = average variance extracted; CR = composite reliability; CLF = common latent factor.

A validity and reliability examination of the measurement model presented Cronbach’s alpha, and composite reliabilities higher than .70 for all constructs, thus exhibiting excellent reliability. The average variance extracted (AVE) values were all above 0.5, showing that the data had an excellent convergent validity, as shown in Table 1 (Hamid et al., 2017).

Model Fit Indices

The results in Table 2 suggest that the structural model fits well with the data as all the goodness-of-fit statistics fell into acceptable ranges (χ² = 94.224, df = 67, χ²/df = 1.406, comparative fit index [CFI] = 0.994, Tucker–Lewis index [TLI] = 0.991, goodness-of-fit index [GFI] = 0.959, adjusted goodness-of-fit index [AGFI] = 0.935, normed fit index [NFI] = 0.979, standardized root mean square residual [SRMR] = 0.0266, and root mean square error of approximation [RMSEA] = 0.036).

Table 2.

Model Fit Indices.

Fixed index	Index value	Accepted value
χ²	94.224	N/A
df	67	N/A
χ²/df	1.406	≤3.00
RMSEA	0.036	≤0.05
CFI	0.994	≥0.95
TLI	0.992	≥0.90
GFI	0.959	≥0.90
AGFI	0.935	≥0.90
NFI	0.979	≥0.90
SRMR	0.0266	≤0.08

Note. RMSEA = root mean square error of approximation; CFI = comparative fit index; TLI = Tucker–Lewis index; GFI = goodness-of-fit index; AGFI = adjusted goodness-of-fit index; NFI = normed fit index; SRMR = standardized root mean square residual.

Hypothesis Testing

Using the (Hayes, 2015) process analysis, we investigated the direct effect of AC on users’ DI, the mediation effect of PU, and PE on the relationship between AC and DI. The result indicates that AC directly influences DI with a positive coefficient of 0.2147 (p < .05). The confidence interval (lower limit confidence interval [LLCI] → upper limit confidence interval [ULCI]) excluded zero, which means that there is a significant effect on DI by AC and therefore confirming H1. PE partially mediated the relationship between AC and DI positively, with a coefficient of 0.1345 (p < .001). PU partially mediated the relationship between AC and DI undoubtedly, with a coefficient of 0.1337 (p < .05). Hayes (2015) posited that the bootstrapped confidence interval levels estimate the presence of mediation (BootLLCI → BootULCI); the confidence interval from lower level to the higher level should exclude zero for there to be a mediation (Hayes, 2015). The result presented in Table 3 shows the bootstrapped confidence interval for both PU (0.0036 → 0.0679) and PE (0.0044 → 0.0711) excluded zero, supporting H2 and H3.

Table 3.

Mediated and Direct Effects of Constructs in Testing Hypothesis.

Paths	p value	BootLLCI	BootULCI	β (effects)
AC → PE → DI	.001	0.0044	0.0711	.1345
AC → PU → DI	***	0.0036	0.0679	.1337
		LLCI	ULCI
AC → DI	.0017	0.0813	0.3481	.2147

Note. Mediation = BootLLCI → BootULCI excluding 0 (at p < .05). LLCI = lower limit confidence interval; ULCI = upper limit confidence interval; AC = article credibility; PE = perceived enjoyment; DI = donation intention; PU = perceived usefulness.

***

p < 0.001.

The study, adopting the conditional process analysis by (Hayes and Rockwood, 2020), further examined the moderation effect of SC on the following relationships: PU and DI, PE and DI, and AC and DI. A visual representation of the direct, mediation, and moderation effects among the constructs in the model is presented in Figure 3.

Figure 3.

Direct, mediation, and moderation effects.

PU mediated the relationship between AC and DI at 13.37%. PE also mediated the relationship between AC and DI at 13.45%. Without the mediation variables, AC has a direct 21.47% direct effect on DI. In addition, introducing the moderator variable (SC) revealed that the extent to which AC, PU, and PE affect DI is specified by SC. The strength of effect of AC on DI is moderated at 21.47%. The strength of effects on DI by PU and PE are moderated at 21.27% and 22.53%, respectively.

An examination of the impact across the levels of the moderator (low, medium, and high) was conducted (Hayes & Rockwood, 2020). The result is as follows: SC significantly moderates the relationship between PU and DI negatively with an interaction p value <.05 (0.0061) and the confidence intervals excluding zero (–0.1287 → –0.0215), supporting H4a, as presented in Table 4. At the lower levels of the moderator, the effects on the relation is higher at a coefficient of 0.2073; at average levels, a mild impact at a coefficient of 0.0979; and a higher level of the moderator shows no impact on the relationship, the reason being that the confidence levels include zero at higher levels of the moderator (–0.0966 → 0.0738) as presented in Table 4.

Table 4.

Moderation Effects of Social Contagion on Construct Relationships in Testing Hypothesis.

Moderated path	β (effects)	p value	LLCI	ULCI
Moderated PU → DI
Perceived usefulness	.0979	.0001	0.0498	0.1461
Social contagion	.2127	***	0.1266	0.2688
Interaction effect	−.0751	.0061	−0.1287	−0.0215
Low-level moderator	.2073	***	0.1096	0.3051
Average moderator	.0979	.0001	0.0498	0.1461
High-level moderator	−.0114	.7920	−0.0966	0.0738
Moderated PE → DI
Perceived enjoyment	.1250	.0004	0.0570	0.1931
Social contagion	.2253	***	0.1350	0.3155
Interaction effect	−.0690	.0269	−0.1301	−0.0079
Low-level moderator	.2254	.0006	0.0968	0.3540
Average moderator	.1250	.0004	0.0570	0.1931
High-level moderator	.0246	.6010	−0.0678	0.1169
Moderated AC → DI
Article credibility	.1402	.0001	0.0717	0.2087
Social contagion	.2147	***	0.1407	0.2887
Interaction effect	−.1572	***	−0.1922	−0.1223
Low-level moderation	.3691	***	0.2944	0.4438
Average-level moderation	.1402	.0001	0.0717	0.2087
High-level moderation	−.0887	.0668	−0.1835	0.0062

Note. Moderation effect = significant interaction effect (p < .05). LLCI = lower limit confidence interval; ULCI = upper limit confidence interval; PU = perceived usefulness; DI = donation intention; PE = perceived enjoyment; AC = article credibility.

***

p < 0.001.

In addition, SC significantly moderated the relationship between PE and DI negatively with an interaction p value <.05 (0.0269) and the confidence intervals excluding zero (–0.1301 → –0.0079), thereby supporting H4b, as presented in Table 4.

At the lower levels of the moderator, the effects on the relation is higher at a coefficient of 0.2254; at average levels, the impact is mild with a coefficient of 0.1250; and a higher level of the moderator shows no impact on the relationship, the reason being that the confidence levels include zero at higher levels of the moderator (–0.0678 → 0.1169), as presented in Table 4.

Finally, SC significantly moderated the relationship between AC and DI negatively with an interaction p value <.001 (***). The confidence intervals excluded zero (–0.1922 → –0.1223), supporting H4c, as presented in Table 4. At lower levels of the moderator, the effects on the relation is higher at a coefficient of 0.3691; at average levels, the impact is small with a coefficients of 0.1402; and a higher level of the moderator has no consequences on the relationship as the confidence levels included zero at higher levels of the moderator (–0.1835 → 0.0062), as presented in Table 4.

Discussion and Implications

This study examined the provocations behind users’ donation behavior on social media platforms. We examined the relationship between characteristics of articles posted on Wechat and DIs. We conducted an investigation through the lenses of the SOR model. The outcome portrays a significant degree of explained variables as far as the dependent variable is concerned. The empirical results show the usefulness of explaining the donation behavior of users with the SOR model. Second, we found that users rely on PE and PU of the article when developing their DIs.

In addition, SC has a significant moderation effect on the relationships between PU and DI, PE and DI, and AC and DIs. To explain further, the rate of third-party interactions (number of views, number of donations already made) will affect users’ donation decisions toward an article despite the prior evaluations made. The result indicates that an article with high SC (high interactions by other users) will negatively affect users’ decision to donate based on the credibility, usefulness, and enjoyability of the article. On the contrary, an article with low SC will attract more donations based on AC, users’ PE, and usefulness. Thus, when a user sees that an article has already received a sizable amount of donation, they are not moved also to donate because they think the article has had enough, whereas an article with fewer donations made will influence the user to donate; after all, he or she thinks that the article has not received enough donations even though it is credible, useful, and enjoyable by the user in question.

Theoretical Implications

One significant contribution of this study is the expansion of the SOR model to include a moderator. The SOR model states that any set of conditions initially generates an emotional reaction, which in turn leads to a behavior. Introducing SC in the SOR reveals that the strength of relation between the emotional reaction and the behavior depends on the moderator (SC). Thus, users may be willing to donate to a content based on its PU and enjoyment; however, SC will specify when this relation will hold. In adapting the SOR framework to social media context, we incorporated PU and PE to be the organism (O) explaining the evaluation of perceived value involved in users’ attitude formation after experiencing the credibility of the article or the stimuli (S). Finally, users’ evaluated perceived value influences their DIs or response (R). However, an external variable (SC) moderates their DIs based on AC and users’ perceived value (usefulness and enjoyment).

The result of this study also fills a gap in the study of consumer behavior by analyzing how the characteristics of a free social media article will influence the attitude formation of consumers and how these attitudes influence their DI toward the article. The credibility of an article directly affects the DIs of a user. However, PU and PE are revealed to partially mediate the relationship between AC and DI, with PE possessing a slightly higher mediation effect as compared with PU.

Moreover, this study extends the understanding of factors that influence donation behavior on social media by introducing the SC nature of an article as a moderator that affects the relationship between consumers’ attitude formation and their intentions to donate. Studies have although shown the critical impact of social factors and attachment on DIs (Kogut & Ritov, 2015). Our study reveals the significant effects that the characteristics of a free article have on the donation behavior of consumers. The present study found that one primary attribute of an article (credibility), two factors of attitude formation (PU, PE), and one external moderating variable (SC nature of the article) are predictors of DIs toward the article on social media.

Practical Implications

The findings present several practical implications. They indicate that authors on social media platforms should pay more attention to the credibility of the article they make available on social media and also the attraction of a large audience to boost the SC nature of the article. For example, content creators can relate their articles to the daily life of users on the social media platform and also draw support from the platform that hosts the article (subscriptions on WeChat platform) to enhance their ability to engage a large audience (Saxton & Wang, 2014) effectively.

The findings show that users rely more on the credibility, enjoyment, and usefulness of the article. Simply put, Chinese users of the WeChat app are more interested in the credibility of an article they come across to decide whether or not to donate. Chinese users also form DIs more on their experienced enjoyment than experienced usefulness; while making reference to the article’s credibility, they look forward to experiencing enjoyment from an article they read. Therefore, authors should try to make articles both useful and enjoyable to increase users’ PE and usefulness of an article. In particular, authors of articles can draw support from platform operators to include gamification items that could make the articles both useful and enjoyable.

Limitations and Future Research

Although this research has significant implications both for researchers and practitioners, it possesses some limitations as well. First, there are other factors, given the scope of our study, which may affect DIs but are not featured in the model. Future research can, therefore, examine other significant antecedents that may predict the donation behavior of users on social media. Second, we exempted external variables, such as the level of income, education level, and gender, which may also present a more transparent comprehension of donation behavior. Future research can consider these factors.

In addition, our results may be related explicitly to China as we obtained our data from users of a Chinese social media application (WeChat) with 1.15 billion active users. As such, the generalization of the findings to other countries is not yet clear. Future research can, therefore, extend our study to different geographical cultures or contexts. Finally, this study did not examine the actual donation behavior in the model. Future research can focus on actual donation behavior or the amounts donated by the users.

Footnotes

Acknowledgements

This research article is made possible through the help and support from the author’s supervisor, colleagues, and family. Specifically, the author likes to dedicate his acknowledgment of gratitude toward the following significant advisors and contributors. First and foremost, the author would like to thank God for His unconditional guidance and wisdom throughout the research. Second, the author would like to thank his supervisor Professor Jiaming Fang from the Management and Economics department of the University of Electronic Science and Technology of China. His support, encouragement, dynamism, vision, and sincerity have deeply inspired the author. He has taught the author how to methodologically carry out research and to present the research work as clearly as possible. Finally, the author would like to thank his wife, parents, and colleagues for having provided him with the advice and motivation needed to complete this research. The product of this research article would not be possible without all of them.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Isaac Owusu Asante

Dennis Fiifi Darko

References

Ab Hamid

M. R.

Sami

Sidek

M. H

. (2017). Discriminant validity assessment: Use of Fornell & Larcker criterion versus HTMT criterion. Journal of Physics: Conference Series, 890(1), 012163.

Aguirre-Urreta

M. I.

(2019). Detecting common method bias: Performance of the Harman’s single-factor test. Data Base for Advances in Information Systems, 50(2), 45–70.

Aksoy

Basaran

(2017). The effect of perceived value on behavioural intentions. Pressacademia, 4(1), 1–16.

Althoff

Leskovec

(2015). Donor retention in online crowdfunding communities: A case study of DonorsChoose.Org. https://dl.acm.org/doi/10.1145/2736277.2741120

Amundsen

(2019). Out in the cold: Busking Copenhagen’s Nørreport station and the urban affects of music. Journal of Urban Cultural Studies, 6, 113–122.

Beldad

A. D.

Hegner

S. M.

(2018). Expanding the technology acceptance model with the inclusion of trust, social influence, and health valuation to determine the predictors of German users’ willingness to continue using a fitness app: A structural equation modeling approach. International Journal of Human–Computer Interaction, 34(9), 882–893.

Bonn

M. A.

Kim

W. G.

Kang

Cho

(2016). Purchasing wine online: The effects of social influence, perceived usefulness, perceived ease of use, and wine involvement. Journal of Hospitality Marketing & Management, 25(7), 841–869.

Brennan

(2017, April 25). 2017 WeChat user report is out! China Channel. https://chinachannel.co/1017-wechat-report-users/

Carlin

(2014). Working the crowds: Street performances in public spaces. In Brabazon

(Ed.), City imaging: Regeneration, renewal and decay (pp. 157–169). Springer.

10.

Chao

C. M.

(2019). Factors determining the behavioral intention to use mobile learning: An application and extension of the UTAUT model. Frontiers in Psychology, 10, Article 1652.

11.

Chen

Huang

Davison

R. M.

(2017). The role of website quality and social capital in building buyers’ loyalty. International Journal of Information Management, 37(1), 1563–1574.

12.

Chen

Dai

Yao

(2019). Donate time or money? The determinants of donation intention in online crowdfunding. Sustainability, 11(16), 4269.

13.

Chin

P. N.

Isa

S. M.

Alodin

(2020). The impact of endorser and brand credibility on consumers’ purchase intention: The mediating effect of attitude towards brand and brand credibility. Journal of Marketing Communications, 26, 896–912.

14.

Cho

Son

(2019). The effect of social connectedness on consumer adoption of social commerce in apparel shopping. Fashion and Textiles, 6(1), 14.

15.

Cho

Y. C.

Sagynov

(2015). Exploring factors that affect usefulness, ease of use, trust, and purchase intention in the online environment. International Journal of Management & Information Systems, 19(1), 21–36.

16.

Diop

E. B.

Zhao

Van Duy

(2019). An extension of the technology acceptance model for understanding travelers’ adoption of variable message signs. PLOS ONE, 14(4), Article e0216007.

17.

Elhadidi

M. A. F.

(2019). Facebook credibility: Evidence from online and offline political participation, political contribution, and platform efficacy. Global Media Journal, 17, 32.

18.

Erlandsson

Nilsson

Västfjäll

(2018). Attitudes and donation behavior when reading positive and negative charity appeals. Journal of Nonprofit & Public Sector Marketing, 30(4), 444–474.

19.

Ferrara

Yang

(2015). Measuring emotional contagion in social media. PLOS ONE, 10(11), Article e0142390.

20.

Fowler

Christakis

(2010). Cooperative behavior cascades in human social networks. Proceedings of the National Academy of Sciences of the United States of America, 107, 5334–5338.

21.

Fuller

C. M.

Simmering

M. J.

Atinc

Babin

B. J.

(2016). Common methods variance detection in business research. Journal of Business Research, 69(8), 3192–3198.

22.

Hamid

A. A.

Razak

F. Z. A.

Abu Bakar

Abdullah

W. S. W.

(2016). The effects of perceived usefulness and perceived ease of use on continuance intention to use E-government. Procedia Economics and Finance, 35, 644–649.

23.

Hasan

Ukkusuri

S. V.

Zhan

(2016). Understanding social influence in activity location choice and lifestyle patterns using geolocation data from social media. Frontiers in ICT, 3, Article 10.

24.

Hayes

A. F.

(2015). An index and test of linear moderated mediation. Multivariate Behavioral Research, 50(1), 1–22.

25.

Hayes

A. F.

Rockwood

N. J.

(2020). Conditional process analysis: Concepts, computation, and advances in the modeling of the contingencies of mechanisms. American Behavioral Scientist, 64(1), 19–54.

26.

Horning

M. A.

(2017). Interacting with news: Exploring the effects of modality and perceived responsiveness and control on news source credibility and enjoyment among second screen viewers. Computers in Human Behavior, 73, 273–283.

27.

Horwitz

(2015, August 13). A “like” button with cash attached: China’s WeChat offers the option to tip writers for posts—Quartz. Quartz Creative. https://qz.com/478732/a-like-button-with-cash-attached-chinas-wechat-offers-the-option-to-tip-writers-for-posts/

28.

Isaac

R. M.

Lightle

Norton

(2015). The pay-what-you-want business model: Warm glow revenues and endogenous price discrimination. Journal of Behavioral and Experimental Economics, 57, 215–223.

29.

Jia

A. L.

Shen

Chen

(2018). Predicting the implicit and the explicit video popularity in a user generated content site with enhanced social features. Computer Networks, 140, 112–125.

30.

Kaiser

H. F.

(1974). An index of factorial simplicity. Psychometrika, 39(1), 31–36.

31.

Kakol

Nielek

Wierzbicki

(2017). Understanding and predicting web content credibility using the content credibility corpus. Information Processing and Management, 53(5), 1043–1061.

32.

Kapoor

K. K.

Tamilmani

Rana

N. P.

Patil

Dwivedi

Y. K.

Nerur

(2018). Advances in social media research: Past, present and future. Information Systems Frontiers, 20(3), 531–558.

33.

Kashif

Sarifuddin

Hassan

(2015). Charity donation: Intentions and behavior. Marketing Intelligence & Planning, 33(1), 90–102.

34.

Kogut

Ritov

(2015). The “identifiable victim” effect: An identified group, or just a single individual? Journal of Behavioral Decision Making, 167, 157–167.

35.

Leenders

R. T. A. J

. (2002). Modeling social influence through network autocorrelation: Constructing the weight matrix. Social Networks, 24(1), 21–47.

36.

(2017). Research on impact factors for online donation behavior of bank customer. The Journal of Finance and Data Science, 3(1–4), 13–19.

37.

Suh

(2015). Factors influencing information credibility on social media platforms: Evidence from Facebook pages. Procedia Computer Science, 72, 314–328.

38.

Liao

C.-H.

Tsou

C.-W.

Shu

Y.-C.

(2008). The roles of perceived enjoyment and price perception in determining acceptance of multimedia-on-demand. International Journal of Business and Information, 3(1), 27–52.

39.

Lin

C.-H.

(2011). The influence of product contagion effect on consumer evaluation: The moderating role of cognitive style. African Journal of Business Management, 5(26), 10369–10375.

40.

Liu

(2019). Motivating user-generated content contribution with voluntary donation to content creators. In Lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics). LNCS (Vol. 11589, pp. 221–230). Springer. https://link.springer.com/chapter/10.1007/978-3-030-22338-0_18

41.

Liu

Burns

Hou

(2017). An investigation of brand-related user-generated content on Twitter. Journal of Advertising, 46(2), 236–247.

42.

Mou

Shin

D. H.

Cohen

(2017). Understanding trust and perceived usefulness in the consumer acceptance of an E-service: A longitudinal investigation. Behaviour and Information Technology, 36(2), 125–139.

43.

Muller

Peres

(2019). The effect of social networks structure on innovation performance: A review and directions for research. International Journal of Research in Marketing, 36(1), 3–19.

44.

Neumayr

Handy

(2019). Charitable giving: What influences donors’ choice among different causes? VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 30, 783–799.

45.

Podsakoff

P. M.

MacKenzie

S. B.

Lee

J. Y.

Podsakoff

N. P.

(2003). Common method biases in behavioral research: A critical review of literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. https://doi/10.1037/0021-9010.88.5.879

46.

Rouibah

Lowry

P. B.

Hwang

(2016). The effects of perceived enjoyment and perceived risks on trust formation and intentions to use online payment systems: New perspectives from an Arab country. Electronic Commerce Research and Applications, 19, 33–43.

47.

Savary

Goldsmith

Dhar

(2015). Giving against the odds: When tempting alternatives increase willingness to donate. Journal of Marketing Research, LII, 27–38.

48.

Saxton

G. D.

Wang

(2014). The social network effect: The determinants of giving through social media. Nonprofit and Voluntary Sector Quarterly, 43(5), 850–868.

49.

Scherer

Siddiq

Teo

(2015). Becoming more specific: Measuring and modeling teachers’ perceived usefulness of ICT in the context of teaching and learning. Computers & Education, 88, 202–214.

50.

Schlosser

A. E.

Levy

(2016). Helping others or oneself: How direction of comparison affects prosocial behavior. Journal of Consumer Psychology, 26(4), 461–473.

51.

Thomas

Johnson

(2016). Effect of celebrity credibility on consumer intention for behavioural change: A study on celebrity-endorsed social advertising. IIMS Journal of Management Science, 7(2), 145–155.

52.

Tussyadiah

I. P.

(2012). An assessment of contagion on social networking sites. In Information and communication technologies in tourism 2012 (pp. 25–35). https://link.springer.com/chapter/10.1007/978-3-7091-1142-0_3

53.

Varpio

(2018). Using rhetorical appeals to credibility, logic, and emotions to increase your persuasiveness. Perspectives on Medical Education, 7(3), 207–210.

54.

Venkatesh

Davis

F. D.

(2000). Theoretical extension of the technology acceptance model: Four longitudinal field studies. Management Science, 46(2), 186–204.

55.

Vishwanath

(2015). Diffusion of deception in social media: Social contagion effects and its antecedents. Information Systems Frontiers, 17(6), 1353–1367.

56.

Wan

Wang

Zhao

(2017). How attachment influences users’ willingness to donate to content creators in social media: A socio-technical systems perspective. Information and Management, 54(7), 837–850.

57.

Wang

Yuan

(2020). The influence of social networks on team performance: Moderating effect of emotional contagion. Open Journal of Social Sciences, 8(4), 553–565.

58.

Wei

Zhao

Zheng

(2016). Moderating effects of social value orientation on the effect of social influence in prosocial decisions. Frontiers in Psychology, 7, Article 952.

59.

Yang

Galak

(2017). When it could have been worse, it gets better: How favorable uncertainty resolution slows hedonic adaptation. Journal of Consumer Research, 43, 747–768.

60.

Teng

Wang

(2015). “What’s in it for me?” The effect of donation outcomes on donation behavior. Journal of Business Research, 68(3), 480–486.

61.

Zeng

Zhu

(2019). A model and simulation of the emotional contagion of netizens in the process of rumor refutation. Scientific Reports, 9(1), 14164.

62.

Zhou

Feng

(2017). Difference between leisure and work contexts: The roles of perceived enjoyment and perceived usefulness in predicting mobile video calling use acceptance. Frontiers in Psychology, 8, Article 350.