Effects of University’s Social Media Presence on Students’ Organizational Media Use and Loyalty

ABSTRACT

Nonprofit and public sector organizations increasingly rely on social media platforms to strengthen their relationships with key stakeholders. Prior research, especially in the for-profit marketing field, indicates positive effects of customers’ engagement with brand-related social media content on their loyalty toward the brand. The present paper examines these effects in the experientially rich and transformative context of higher education institutions (HEI). Our conceptual model also comprises new potential antecedents relating to the specific context. The results from the structural equation modeling, based on responses from 764 students at a Swiss HEI, indicate significant yet small positive effects of students’ social media engagement (SME) on their loyalty. Positive word-of-mouth behavior and loyalty intentions are primarily influenced by identification with the university. Attitude toward the published content, perceived organizational reputation, and student integration significantly influenced SME, whereby the latter two antecedents showed only a weak effect. In conclusion, SME as a behavioral construct seems to be another expression of existing student loyalty rather than an effective tool to foster loyalty.

KEYWORDS:

• Higher education
• online communication
• relationship marketing
• social media engagement
• student loyalty

Introduction

Social media have become increasingly important for organizations in the nonprofit and public sectors to foster stakeholder engagement and fulfill their mission (e.g., Lovari & Valentini, 2020; Suh, 2022; Xu & Saxton, 2019). This development also holds for Higher education institutions (HEI): In times of growing competition, cultivating long-term relationships and relationship marketing are becoming increasingly vital (e.g., Helgesen, 2008), especially because the strengthened relationships are expected to lead to student retention and loyalty (de Macedo Bergamo et al., 2012). One way of cultivating relationships with current students is to create a sense of belonging to the university through social media (Garza Salgado & Royo Vela, 2019). These media channels are suitable for said purpose due to their relational and interactive nature (Azar et al., 2016). Since Facebook, Instagram, YouTube, and Co. have a central place in the lives of the current generation of students, HEIs have integrated social media into their communication strategies and professionalized their approaches (Alfonzo, 2021; Clark et al., 2017; Garza Salgado & Royo Vela, 2019).

The university community is formed by its members, namely professors, administrative and scientific personnel, as well as students. Those subgroups participate in a collective exchange of knowledge and a discussion process for gaining the latter. By enrolling, students officially join the university and become members. In exchange for tuition fees (which can also be seen as membership fees), the university offers various services, including curriculum, personnel, and material infrastructure. Moreover, students participate in “university governance by collective bodies such as universities senate and faculty councils” (Moraru, 2012, p. 54). Hence, the present discussion is also relevant for nonprofit organizations in general since they need to assess their social media activities’ effectiveness in engaging their constituents and, ultimately, inducing supportive behavior (Smith, 2018).

In higher education research, the subject of social media engagement (SME), its impacts, and possible drivers have become more prevalent over the last few years. Research has so far focused on questions about HEIs’ communication strategies on social media, the content and post characteristics as well as account-related (e.g., virtual lifetime) or organizational factors (e.g., size) and their effectiveness in engaging students in the form of likes, shares/retweets/repins, comments, etc. (e.g., Bonilla Quijada et al., 2022; Capriotti et al., 2023, 2024; Chauhan & Pillai, 2013; Condie et al., 2018; del Rocío Bonilla et al., 2020; Linvill et al., 2015; Peruta & Shields, 2016, 2018; Soares et al., 2022; Sörensen et al., 2023; Mai To et al., 2022; Wahid & Gunarto, 2022). Moreover, drivers of students’ engagement on HEIs’ brand pages such as university reputation (e.g., Brech et al., 2017), brand experience and brand interactivity (Farhat et al., 2021), as well as eWOM (mediated by brand equity) (Brandão & Ramos, 2023), or the relationship between SME and student recruitment/enrollment intention (e.g., Konstantoulaki et al., 2022; Nguyen et al., 2021; Zhu, 2019) as well as with HEI’s brand positioning (Perera et al., 2022) have been explored. Prior studies have also investigated the impact of SME on students’ perceived relationship quality with the university (Clark et al., 2017), on the identification with the HEI and its community (Fujita et al., 2017), and on loyalty (Garza Salgado & Royo Vela, 2019; Shah et al., 2021).

Social media channels serve as communication platforms for HEIs across the entire customer journey (see, e.g., Garza Salgado & Royo Vela, 2019), defined as “the student’s/participant’s journey with their educational service provider over the pre-admission, during the course and post admission” (Popli et al., 2023, p. 478). However, this article focuses on a specific stage of this journey: the “during the course” or “post-purchase” stage, as termed by Lemon and Verhoef (2016). The central question revolves around the impact of enrolled students’ engagement with university social media offerings (considered as brand-owned touchpoints) in guiding them toward the “loyalty loop” of the customer journey (see Court et al., 2009). Focusing on the psychological phases of this journey, Court et al. (2009) assert that the post-purchase experience significantly shapes customers' perceptions for future decisions within the same category, thus perpetuating the journey as a continuous cycle. In terms of loyalty, they distinguish between active loyalists, who advocate for the brand, and passive loyalists, who are susceptible to competitor offers (Court et al., 2009). The present study differs from previous research as it explores not only the effects of SME (conceptualized as a behavioral construct with a passive and active dimension) on students’ loyalty intentions but also action loyalty beyond the social media sphere. It also examines new potential antecedents of SME, like student integration. For HEIs as professional service organizations, knowing how their relationship marketing activities can contribute to the long-term well-being of the institution is crucial. They need to discern to what extent an organization can strengthen student loyalty through social media activities. Consequently, the central research question of this paper is: How does the students’ social media engagement affect their (intended) supportive behavior toward their university?

In order to elucidate the aforementioned scientific interest, a survey with a standardized questionnaire was conducted at a public university in Switzerland. Partial least squares structural equation modeling was then used to test the proposed conceptual model. As the results demonstrate, the influence of students’ engagement with the university’s social media content on loyalty is lower than expected.

Conceptual model and research hypotheses

Social media engagement

With the increasing importance of social media in students’ lives, HEIs today turn more frequently to those channels for relationship marketing purposes (Clark et al., 2017). As active communicators, users react to brand communication via social media with different behaviors (Piehler et al., 2019). Schivinski et al. (2016) speak of consumer’s engagement with brand-related social media content in this context and define it as follows: “A set of brand-related online activities on the part of the consumer that vary in the degree to which the consumer interacts with social media and engages in the consumption, contribution, and creation of media content” (p. 66).

In the higher education field, research on SME has become more prevalent in the last few years but is still relatively fragmented. How students’ SME (represented by following the brand page) affects the relationship quality with their university has been studied by Clark et al. (2017). Their results show a significant positive impact of engagement on perceived relationship quality. Fujita et al. (2017) found in a case study of international students at an Australian university that students’ SME enhanced their social identification with the university. Garza Salgado and Royo Vela (2019) showed strong positive effects of engagement with the HEI’s brand fan page and the usage intensity of said page on students’ brand loyalty toward their university. Shah et al. (2021) substantiated the positive impact of engagement with HEI’s social media activities (browsing, liking, sharing, and commenting) on students’ brand loyalty in the online sphere. The survey results from Pakistani public universities also show a significant positive effect of students’ general social media use on SME.

Regarding SME, two major research streams can be distinguished: 1) behavioral and 2) multidimensional (Yoshida et al., 2018). The multidimensional approach essentially assumes that SME includes a cognitive, an emotional, and a behavioral dimension (e.g., Dessart, 2017; Hollebeek et al., 2014). Hollebeek et al. (2014) define these three facets – always in relation to a specific encounter between a consumer and the brand – as follows: The cognitive dimension refers to the degree of thought processing the individual exerts concerning the brand; the emotional dimension comprises the level of positive affection she or he experiences with the brand; the behavioral dimension finally describes the activation, i.e., the amount of effort she or he invests in the brand. Researchers have been able to find positive correlations between the individual dimensions (Brodie et al., 2013; Habibi et al., 2014; Leckie et al., 2016 among others), which is why some authors – such as Hinson et al. (2019) – view engagement as a reflective second-order construct, which is composed of the aforementioned three first-order dimensions. The behavioral approach understands SME as concrete behavior (including Dolan et al., 2016; Muntinga et al., 2011; Schivinski et al., 2016; Tsai & Men, 2013). This perspective presumes that the latent cognitive and emotional processes of individuals are expressed and reflected in their behavioral engagement toward the brand (Khan et al., 2023). For instance, when a brand post evokes positive emotions, this affective engagement might manifest itself by the individual liking the respective post (Demmers et al., 2020). We follow the behavioral approach because it allows psychological constructs related to SME to be better delineated and classified (Yoshida et al., 2018). In accordance with the definition by Schivinski et al. (2016) cited above, this paper understand SME as measurable reactions of students to university-created social media content. It, therefore, draws on the COBRA framework (consumers’ online brand-related activities) introduced by Shao (2009) and further elaborated by Muntinga et al. (2011). It is an overarching behavioral construct comprising a broad spectrum of user activities with brand-related social media content. Consumption is the basic level of online activity and includes participation without actively contributing or creating (e.g., reading text posts, watching videos/photos). The middle level is called contribution and describes interaction with social media content. Examples include rating, sharing, or commenting on the posts of others. Creation represents the highest level of activity: the active production and publication of one’s own brand-related content (e.g., creation of posts). Therefore, the COBRA framework is a usage typology and not a user typology, which considers that an individual can perform different behaviors, simultaneously or also successively (Muntinga et al., 2011; Schivinski et al., 2016; Shao, 2009).

Student loyalty

A variety of possible consequences of customer engagement have been discussed in previous research. In the specific context of social media, Barger et al. (2016) and, with a specific focus on branding, Vander Schee et al. (2020), categorized potential outcomes of engagement identified in the scientific literature. Barger et al. (2016) discussed effects concerning the brand (e.g., brand awareness, brand loyalty), product (e.g., attitude toward product), consumer (e.g., social capital), content (e.g., re-sharing intention), or market (e.g. diffusion of information) (p. 270). Vander Schee et al. (2020) narrowed the perspective to brand outcomes, including status (e.g., preference, loyalty, usage intent), disposition (e.g., attitude, personality), attribute (e.g., image, satisfaction), connection (e.g., attachment, experience), affirmation (e.g., advocacy, fidelity), and aversion (e.g. avoidance) (p. 241). In higher education, studies have shown that students’ engagement on social media can positively influence their perception of relationship quality, characterized by satisfaction with the relationship, trust in the exchange partner, and commitment to maintaining the relationship (Clark et al., 2017), as well as strengthen their social identification with the HEI (Fujita et al., 2017). In the following, we focus on a specific outcome, namely loyalty.

Student loyalty is one of the primary objectives of HEIs, as it can elicit a competitive advantage in the growingly contested higher education sector. It is not bound to the actual period of study at the university but goes beyond graduation. Loyalty is of great importance for HEI: loyal students not only contribute to the quality of the education, but they also support their academic institution with tuition fees and, after graduating, through financial contributions, the willingness to recommend the university to others or some other form of cooperation. This study considers conative and action loyalty regarding the respective time period (during enrollment and after graduation). The construct of action loyalty encompasses the actual word-of-mouth behavior of enrolled students during their time at the university; conative loyalty refers to their intentions to continue the relationship with and financially support their alma mater after graduation.

Students’ SME is an essential intermediate step to student loyalty. HEIs can establish and maintain long-term relationships with their stakeholders through relationship marketing, fostering loyalty to the organization (de Macedo Bergamo et al., 2012). Social networks provide universities with an additional forum to build bonds with their community (Clark et al., 2017). Relationship marketing is connected to social exchange theory (SET), which forms the main theoretical basis for our research model and is discussed below.

Blau (1964, p. 91) defines social exchange as “voluntary actions of individuals that are motivated by the returns they are expected to bring and typically do in fact bring from others.” While his definition primarily identifies human entities as actors, social exchange theorists note that reciprocations can also occur between individuals (in this context, university students) and corporate groups acting as a unified entity (e.g., HEIs via their social media accounts) (Molm, 2006). These exchanges involve unspecified obligations, meaning the future consideration is not predetermined or negotiated but depends on the discretion of the obligated party (Blau, 1964). This implies that one actor’s reciprocation is contingent upon the behavior of the other (Cropanzano & Mitchell, 2005). Consequently, mutual trust plays a vital role and typically develops through successful or mutually satisfying exchange processes. Ultimately, friendly or social ties are established over time, which is one of the central functions of this type of exchange (Blau, 1964). Therefore, these interdependent transactions are capable of cultivating high-quality relationships (Cropanzano & Mitchell, 2005). The resources exchanged can be tangible (e.g., money or goods) or socioemotional (such as prestige, recognition, or loyalty), or both (Arnett et al., 2003; Homans, 1958). In their “Resource Theory of Social Exchange,” Foa and Foa (2012, p. 16) classify resources into six categories: love, social status, information, money, goods, and services. The long-term exchange processes under consideration are predicated on the principle of reciprocity (Bagozzi, 1995). What a person contributes represents a cost, and what they receive signifies a reward (Homans, 1958). Exchange partners aim for an equilibrium in these reciprocal transactions and may implement corrective measures if necessary (L. Hollebeek, 2011).

Through engagement with a brand’s social media content, users acquire additional knowledge about it and demonstrate favorable attitudes toward the brand (Fernandes & Castro, 2020). In the context of higher education, Clark et al. (2017), drawing on SET, argue that students with stronger SME have greater access to valuable resources and consequently feel a greater obligation to reciprocate (as improved relationship quality). Therefore, as students recognize the efforts of HEIs to maintain a good relationship with the student body by delivering value through their social media activities, they are expected to reciprocate this effort. Moreover, we anticipate that the reciprocal quality of customer engagement (L. Hollebeek, 2016), which extends to social media engagement, positively influences the relationship between students and the HEI, as it can be seen as a favorable evaluation of the exchanged resources. Jahn and Kunz (2012) maintain that users’ engagement with fan pages on Facebook is linked to an enhanced relationship with both the online brand community and the brand itself. Finally, to sustain their relationship with the brand, consumers develop loyalty intentions toward it (Dwivedi, 2015).

Various studies have already demonstrated significant effects of social media engagement on (brand) loyalty in different contexts (e.g., de Oliveira Santini et al., 2020; Garza Salgado & Royo Vela, 2019; Yoshida et al., 2018). Furthermore, focusing on diverse brand fan pages on Facebook, Fernandes and Castro (2020), as well as Jahn and Kunz (2012), found that both active and passive forms of SME positively influenced brand loyalty; in the latter case, loyalty includes brand commitment, word-of-mouth, and purchase behavior. Regarding which forms of SME are responsible for these effects, previous studies have shown varying results. For example, Mishra (2019) discovered that only creating brand-related content significantly influenced purchase intentions among young consumers. However, in their study in the luxury cosmetics sector, Cheung et al. (2021) demonstrated that creating and consuming (but not contributing) exerted a significant favorable effect on repurchase intention. Piehler et al. (2019) observed that only consumption significantly and positively impacted word-of-mouth behavior in the offline context while contributing or creating did not.

Based on those arguments and previous research findings, we expect students’ SME to have a positive effect on their word-of-mouth behavior:

H1a:

Students’ social media engagement has a positive impact on their action loyalty (positive word-of-mouth) toward the university.

As discussed above, students’ engagement with their alma mater may extend beyond graduation in various forms like attending professional development events or through donations/financial gifts, recruitment or volunteering for events (Garza Salgado & Royo Vela, 2019; Mael & Ashforth, 1992). Thus, following the previously mentioned arguments, we also assume that SME positively influences students’ intentions to continue supporting the university after graduation:

H1b:

Students’ social media engagement has a positive impact on their conative loyalty (behavioral intentions) toward the university.

Drivers of students’ social media engagement

In this paper, we explore three potential drivers of students’ SME, inspired by the gratification areas for fan page participation considered by Jahn and Kunz (2012). The first antecedent of SME in our model is students’ attitudes toward the social media content published by the HEI. In advertising research, Lutz (1985, cited in MacKenzie et al., 1986) defined attitude toward the advertisement as “a predisposition to respond in a favorable or unfavorable manner to a particular advertising stimulus during a particular exposure occasion” (pp. 130–131). In the context of viral video advertising, Huang et al. (2013) demonstrated that a positive attitude toward the ad significantly impacted the intention to share digital content online. Similarly, Kujur and Singh (2017) applied these considerations to social networking ads and found a significant positive impact of users’ attitudes toward brand posts on their participation and engagement on social networking sites. According to Foa and Foa (2012), information is a key resource in exchange processes. Additionally, Barger et al. (2016) stated that branded content itself exerts a non-negligible influence on consumer engagement in social media. It can be assumed that high-quality content is evaluated as a valuable resource in an exchange. Therefore, based on SET, when students perceive that the HEI publishes interesting and appealing content on social media – a judgment developed through experience – their favorable attitudes prompt them to reciprocate by interacting with this high-quality content. Consequently, it follows:

H2:

The more positive students’ attitudes toward HEI’s social media content, the stronger their social media engagement will be.

Another hypothesized driver of SME is the perceived organizational reputation of the HEI. Schwaiger (2004) distinguishes his attitudinal reputation construct into an affective (liking toward the organization) and a cognitive (perceived competence) dimension, which more strongly emphasizes subjective perception. We base our reasoning on the reflections of Swani and Labrecque (2020) on how self-presentation and brand relationship connections influence consumers’ SME. Impression management serves to cultivate a positive image in the public eye, driven by people’s basic need for self-enhancement (Leary, 2003). Social media is a convenient forum for this purpose, especially by intentionally curating and sharing content that supports one’s desired public image (Hollenbeck & Kaikati, 2012; Swani & Labrecque, 2020). In this context, Hollenbeck and Kaikati (2012) showed that consumers on Facebook publicly connect with brands (i.e., by liking their brand page) to convey their actual and desired selves to other users. Assuming that individuals aim to portray themselves positively, we expect them to interact primarily with brands they perceive as more reputable. Mei et al. (2022) found in their explorative study with representatives of Gen Y that concerns about one’s personal brand indeed affect their SME.

According to SET, prestige and social status are key exchange resources (Foa & Foa, 2012; Homans, 1958). Since HEIs represent brands to their stakeholders, they become targets for students’ self-presentation activities (Brech et al., 2017). Consequently, we hypothesize that an HEI’s perceived reputation influences how students publicly engage (e.g., through likes, shares, or comments) with the content to showcase their connection to the HEI and, in turn, benefit from this association.

H3:

The more positive the university’s reputation is perceived, the stronger the social media engagement of its students will be.

Within the framework of their relationship quality-based student loyalty model, Hennig-Thurau et al. (2001) show academic and social integration into the university context to positively influence emotional commitment to the institution, which had a substantial direct effect on student loyalty. The authors refer to Tinto’s (1975) theoretical model of student dropout behavior, distinguishing academic and social integration. While the former encompasses performance (academic grades and intellectual development), social integration looks at the individual’s relationship with others in the higher education community and how well the individual and this specific environment fit together. The resulting sense of belonging can be conceptualized as a social reward (Tinto, 1975). We follow the approach and hypothesize that the positive effect on loyalty can be transferred to students’ SME. It can further be assumed that students who are well integrated into university life are more interested in what is happening in this context and extend their exchanges with other university members into the online sphere. Social media platforms can satisfy the need for belonging “through facilitating and reinforcing online and offline relationships with both other consumers and brands” (Swani & Labrecque, 2020, p. 281). And as Muntinga et al. (2011) have argued, the motivations of integration/social interaction and information-seeking form essential driving forces behind SME. Based on SET, we thus hypothesize that the HEI’s content on social media is seen as a valuable resource to stay up-to-date and connected with the HEI and its community, which is reciprocated by engaging with it. Thus follows:

H4:

The better students are integrated into the university context, the stronger their social media engagement will be.

Identification with the HEI as a moderator

In the higher education context, with its shared user experiences, much of the interpersonal exchange occurs directly on-site (Fujita et al., 2018). Therefore, it can be assumed that the community nature of the educational setting already induces organizational identification among students in the offline context. The university (community) functions as a social category/group with which an individual identifies and from which she or he derives meaning (Ashforth & Mael, 1989). Studies have found a significant positive effect of identification with the university on brand loyalty and support (Fazli-Salehi et al., 2019; Palmer et al., 2016), on promotion/advocacy intentions (Abdelmaaboud et al., 2021; Mael & Ashforth, 1992; Stephenson & Yerger, 2014) as well as on donations to the alma mater (Arnett et al., 2003; Mael & Ashforth, 1992). Furthermore, social groups present themselves as a social context factor in media effects research, inter alia regarding the media selectivity paradigm (Valkenburg et al., 2016). Harwood (1999) argues that media users turn to specific messages supporting their social identities and offering a positive social differentiation to outgroups. Even more so for individuals for whom the portrayed characteristics are a key facet of their self-concept. He shows that individuals look for media portrayals that can reinforce their identification with a chosen social group or render it more positive. Selectivity also occurs during exposure to the message through selective perception: depending on their predispositions, individuals apprehend messages in an agreeable way that reinforces the latter and may facilitate media effects (Kim & Rubin, 1997). Those perceptions, cognitions, and sentiments are partly shaped by the groups individuals identify with (Merton, 1968).

We expect that students who highly identify with their HEI interact with the content through the lens of university-related social identity and thus pay more attention to positive narrative cues that differentiate them from the outgroup. This leads to a better perception of the social exchange relationship and finally to a better outcome.

H5:

Organizational identification strengthens the impact of social media engagement on student loyalty to the university.

The hypotheses presented above are summarized in Figure 1. The conceptual model shows the assumed antecedents and consequences of SME as well as the moderator and control variables.

Figure 1. Research model.

Methodology

Research context and sample

The data collection at a Swiss public university was conducted through a standardized survey using an online questionnaire in November 2019. All enrolled students were invited by e-mail to complete the questionnaire on LimeSurvey in German or French. Items originating from the research literature were translated from English into these two languages by native speakers and then compared to ensure consistency in meaning. To account for the study subject and cultural context, some modifications were made (discussed in the next section). Both survey instruments were then counterchecked with German and French native speakers. Subsequently, a preliminary test was conducted with a small cohort in a Master’s seminar, leading to minor linguistic modifications. The final items, retranslated into English after incorporating these adjustments, are presented in Table 1.

Table 1. Main effects model (without control variables).

Constructs	Mean (SD)	Range (min-max)	Loading	α	ρA	CR	AVE
Organizational identification (OI)		0–5		0.737	0.738	0.835	0.560
When someone praises this University, it feels like a personal compliment. (OI_1)	3.16 (1.20)		0.796***
I am very interested in what others think about this University. (OI_2)	3.31 (1.20)		0.704***
I consider the successes of this University to be my successes. (OI_3)	2.37 (1.22)		0.788***
Studying at this University means more to me than just having a degree. (OI_5)	3.16 (1.32)		0.700***
Student integration – academic (INT – acad)		0–4		0.695	0.698	0.832	0.623
I am an active member of several student organizations at this University. (INT_3)	1.08 (1.46)		0.797***
I frequently attend additional courses/lectures/meetings/colloquia outside of my curriculum. (INT_5)	1.82 (1.16)		0.734***
I am actively involved in University politics. (INT_8)	0.74 (1.09)		0.833***
Student integration – social (INT – social)		0–4		0.645	0.652	0.809	0.585
I meet regularly with my study group to work. (INT_4)	1.61 (1.27)		0.735***
I maintain intensive contact with many of my fellow students. (INT_6)	2.62 (1.15)		0.817***
I regularly participate in leisure activities at this University. (INT_7)	2.21 (1.31)		0.741***
Perceived Organizational reputation (POR)		1–5
From today’s perspective, I assess the value of my degree at this University in comparison to graduates of other Swiss universities in the same field as follows.	3.43 (0.77)		1.000***
Attitude toward social media content (ATT)		0–6		0.706	0.778	0.868	0.767
Boring – Interesting (ATT_1)	4.21 (1.26)		0.923***
Unappealing – Appealing (ATT_2)	4.88 (1.11)		0.825***
Social media engagement – Consumption (SME – cons)		0–4
I read the text posts or look at the pictures and/or videos posted by this University. (SME_1)	2.20 (0.91)		1.000***
Social media engagement – Contributing (SME – cont)		0–4		0.702	0.712	0.832	0.623
I rate/like the posts of this University. (SME_2)	1.31 (1.14)		0.808***
I share the posts of this University with my network. (SME_3)	0.56 (0.81)		0.786***
I write comments on the posts of this University. (SME_4)	0.35 (0.66)		0.774***
Positive word-of-mouth (pWOM)		0–4		0.787	0.797	0.876	0.701
I often advertise my studies at this University among my acquaintances. (pWOM_1)	2.38 (1.136)		0.858***
I often talk about my studies at this University in my personal environment. (pWOM_2)	3.15 (0.891)		0.795***
In personal conversations, I often point out the positive aspects of studying at this University. (pWOM_3)	2.64 (1.075)		0.858***
Conative loyalty (LOY) From today’s perspective, can you imagine …		0–3		0.643	0.648	0.800	0.571
… to attend a continuing education program at this University for your professional function? (LOY_2)	2.12 (0.696)		0.737***
… to attend a continuing education program at this University for your personal development? (LOY_3)	1.92 (0.779)		0.768***
… supporting this University, a department, or a specific research project financially? (LOY_4)	1.16 (0.785)		0.762***

Significance test based on 5.000 bootstrap runs with 674 cases.

Level of significance: ***p < .001.

SD = Standard deviation, α = Cronbach’s α, ρA = Reliability coefficient, CR = Composite reliability, AVE = Average variance extracted.

A total of 2,688 (fully or partially) completed questionnaires were returned, corresponding to a response rate of around 26%. For the present study, a sub-sample was drawn, consisting of 674 current Bachelor and Master students who follow the university’s official brand page/profile on at least one social media channel and encounter the HEI’s postings in their feed. Those followers form a virtual community around the university brand, in which, however, the community character may be less salient compared to consumer-administered brand communities (Zaglia, 2013). Bachelor students make up the larger proportion with 63%, and 37% are Master students.

Operationalization

The variables of the research model were taken mainly from the scientific literature and supplemented by specific items tailored to the context.

Social media engagement comprises four items based on Muntinga et al.'s (2011) COBRA typology, aligned with the Schivinski et al. (2016) scale. Individual items were translated and reworded to fit the research context, primarily to establish a more explicit connection to the sender of the messages. References to specific social media platforms are intentionally relegated to the background to provide a general assessment of the level of interaction with university postings. One item measured the consumption level (reading text postings/looking at pictures or videos published by the HEI), and three items represented the contribution level: liking, sharing, and commenting on the HEI’s postings. The highest level was not taken into account because of the focus on university-created content and the interaction with it. The items were measured on a 5-point Likert scale from “always” to “never.”

Attitude toward organizational social media content was measured on a 7-point semantic differential scale following MacKenzie et al. (1986), who used the two scales “favorable/unfavorable” and “interesting/boring” to measure one’s attitude toward an advertisement. We replaced the first dimension with another item capturing the presentation/editing of the content (appealing/unappealing).

Perceived organizational reputation was queried in the present study based on a single item: “From today’s perspective, I assess the value of my degree at the University of […] in comparison with graduates of other Swiss universities in the same field as follows” with a 5-point scale from “significantly superior” to “significantly inferior.” It only considers the cognitive dimension or the subjective perception of competence. The value of an HEI’s degree (an indication of the university’s competence) in the labor market is determined in proportion to other universities.

The eight items of the higher-order construct student integration are based on the scale by Hennig-Thurau et al. (2001), measured on a 5-point Likert scale (“fully applies” to “does not apply at all”), and supplemented with items by Iskhakova et al. (2016). Both cited articles distinguish between academic and social integration (with two separate scales), representing the two lower-order constructs in the present article. Since we surveyed enrolled students, we formulated the items in the present tense (as Iskhakova et al. (2016) did). An example of social integration is: “I regularly participate in leisure activities at the University of […] (e.g., sports activities, anniversaries, events);” and one example for academic integration is: “I frequently attend additional courses/lectures/meetings/colloquia outside of my curriculum.” To take greater account of academic integration, understood as adherence to explicit standards in this system (Tinto, 1975), two items related to university teaching were added to the scale (e.g., “I strive to participate as actively as possible in class”).

Students’ loyalty to their university comprises two latent variables: action and conative loyalty. The former refers to pWOM and comprises three items derived from the variable “promoting” by Arnett et al. (2003). The adapted wording emphasizes the individual facets and highlights the varying intensity of the items (e.g., “In personal conversations, I often point out the positive aspects of studying at the University of […]”). All three items were measured on a 5-point Likert scale from “fully applies” to “does not apply at all.” Four items measured on a 4-point scale (from “yes, definitely” to “certainly not”) form the conative loyalty construct (e.g., “From today’s perspective, can you imagine supporting the university, a department, or a specific research project financially?”). These items were derived using the survey instruments Hennig-Thurau et al. (2001) and Sung and Yang (2009) developed.

The well-known scale developed by Mael and Ashforth (1992) was used and adapted to the Swiss context to measure organizational identification. Four of the six original items were retained (including “When someone praises the University of […], it feels like a personal compliment”), while two were omitted due to redundancy and the different context. A fifth item – borrowed from Arnett et al. (2003) University Identity Salience and adapted from alumni to students – completes the present scale (“Studying at the University of […] means more to me than just having a degree”). To avoid a tendency toward the middle, a 6-point Likert scale was used (“strongly agree” to “strongly disagree”).

Regarding SME, general social media use (3-point scale: “intensively;” “occasionally;” “not at all”) was controlled for. The duration of studies (number of semesters completed as a free numerical specification) was assessed concerning the outcome variables of behavioral and conative loyalty.

Method

To test the conceptual model outlined above, we relied on partial least squares structural equation modeling (PLS-SEM) using the SmartPLS 4 software package (Ringle et al., 2022). Data were analyzed in a two-step procedure, first measuring the outer models to assess measurement quality, and in the second step, we estimated the inner (structural) model (Hair et al., 2022). The hierarchical component models (SME and INT as reflective-formative-type models) were specified using the embedded two-stage approach with Mode B estimation setting for step 2 (Sarstedt et al., 2019). As recommended by Hair et al. (2017), the path weighting scheme was employed (stop criterion of 1.0E–7, maximum of 300 iterations), and bias-corrected bootstrapping served for significance testing (5.000 bootstrap samples, two-tailed test). With regard to the moderation analysis, the two-stage approach has been chosen since the exogenous constructs (SME and INT) are measured formatively (Chin et al., 2003). The missing values per item are generally between zero and one percent, except for three items with up to 7.3%. Those three items represent personal perceptions, which might have been (especially for perceived organizational reputation with most missing values) more challenging to determine for some participants than actual behavior. Mean replacement was employed for the statistical analyses.

Analyses and results

Test on unidimensionality and common method bias

Prior to analyzing the hypothesized model, Principal Component Analysis (PCA) in IBM SPSS Statistics v.28.0.0.0 (IBM Corp, 2021). has been conducted. This procedure was performed using varimax rotation with Kaiser normalization (KMO = 0.810). Ten different factors were identified, accounting for about 58% of data variance (see Appendix A).

As the present study relies on self-reported data from the same person for both the independent and the dependent variables, we controlled for common method variance. The technique of choice was the widely used Harman’s single-factor test (Podsakoff et al., 2003). The PCA (unrotated solution) results revealed ten factors with an eigenvalue greater than 1, negating a single-factor solution. Furthermore, the largest single factor accounted for 17% of the total variance, which is below the threshold of 50%. Hence, common method bias is not an issue in the following analysis (Podsakoff et al., 2003).

Measurement model assessment

The first step consists of assessing indicator reliability. Three reflectively measured constructs showed outer loadings smaller than the threshold value of 0.708 (Hair et al., 2022). Even though this is a common situation for estimated models (Hulland, 1999), we proceeded with the relevance testing for all those indicators, as Hair et al. (2022) suggested. As a result, general social media use was transformed into an index, measured as the mean of the five initial items. One item was deleted in the case of organizational identification and one item in the case of conative loyalty – which in both cases led to an improvement in various evaluation criteria. Finally, all outer loadings show values of 0.700 or higher, which is acceptably close to the threshold value (Hair et al., 2022).

Concerning SME and INT, two steps need to be accounted for when assessing higher-order constructs (HOC): evaluation of the measurement models of 1) the lower-order constructs (LOC), and 2) of the HOC (Sarstedt et al., 2019). Regarding SME, we retained two LOC from the first step: consuming with one reflective and contributing with three reflective indicators (liking, sharing, commenting). No collinearity issues among the LOC had been found (VIF <3). Both LOC, consuming and contributing, representing formative indicators of the HOC, showed significant and (more or less) relevant outer weights (consuming: 0.757; and contributing: 0.407, respectively) and were retained. Also both LOC (academic and social integration) were retained for student integration with three items each. The two additional self-formulated items for academic integration were deleted. Again, no collinearity issues among the two LOC had been found (VIF <3). Both LOC (formative indicators) showed significant (social integration on a 10% level) and more or less relevant outer weights (academic: 0.769; social: 0.380). Hence, the two LOC were retained (Sarstedt et al., 2019).

In the second step, we evaluated the internal consistency reliability, assessed by Cronbach’s alpha as the lower bound, the reliability coefficient ρ_A, and composite reliability (CR) as the upper bound (Sarstedt et al., 2017). The Cronbach’s alpha and ρ_A values for all constructs but three are acceptable (over 0.70) (Hair et al., 2022). Academic integration (CA = .695) falls just below the threshold value, social integration (CA = .645) and conative loyalty (CA = .643), with three items each, are lower. We consider those values acceptable since Cronbach’s alpha represents the lower bound and CR’s upper bound (which meets the criteria). CR values should fall between 0.70–0.90 (Hair et al., 2019). This criterion is met since all CR values range between .800 and .876.

Convergent validity is measured by the outer loadings/indicator reliability (which we already established) and by the average variance extracted (AVE), with a threshold value of 0.50 (Hair et al., 2022). As shown in Table 1, all latent constructs in the model meet this criterion.

In the last step, the measurement models were assessed for discriminant validity. The HTMT ratio did not emphasize any issues (see Table 2), even according to the more restrictive threshold of 0.85 (Henseler et al., 2015), nor did the Fornell-Larcker criterion (Hair et al., 2022). Since the LOCs of the SME and INT constructs do not need to show discriminant validity toward their HOC, they are not taken into account here (Sarstedt et al., 2019).

Table 2. Discriminant validity of the measurement models – HTMT ratio.

	1	2	3	4	5	6	7	8	9	10	11	12	13
ATT(1)
LOY(2)	0.353
SME – cons(3)	0.427	0.251
SME – cont(4)	0.300	0.281	0.488
DUR(5)	0.059	0.032	0.010	0.047
INT(6)	0.249	0.229	0.178	0.238	0.077
OI(7)	0.316	0.550	0.306	0.380	0.049	0.336
POR(8)	0.174	0.226	0.177	0.159	0.055	0.086	0.335
SME(9)	0.414	0.328	0.798	1.344	0.042	0.264	0.430	0.200
gSMU(10)	0.047	0.047	0.066	0.184	0.159	0.047	0.155	0.024	0.175
INT – acad (11)	0.192	0.233	0.192	0.206	0.081	1.192	0.336	0.105	0.243	0.021
pWOM(12)	0.282	0.483	0.249	0.239	0.065	0.360	0.728	0.298	0.294	0.015	0.327
INT – social (13)	0.256	0.176	0.123	0.222	0.056	1.223	0.264	0.047	0.229	0.064	0.671	0.317
OI x SME (14)	0.069	0.020	0.057	0.146	0.005	0.077	0.073	0.055	0.140	0.061	0.079	0.037	0.059

LOY = Conative loyalty, SME = Social media engagement, cons = Consumption, cont = Contribution, INT = Student integration, DUR = Duration of studies, OI = Organizational identification, POR = Perceived organizational reputation, ATT= Attitude toward organizational social media content, gSMU = General social media use, pWOM = positive word-of-mouth.

Structural model and hypothesis testing

After substantiating the quality of the outer models, we evaluated the structural model, following the procedural steps suggested by Hair et al. (2022) and Sarstedt et al. (2017). We hereby compare the full model with the interaction effects (Model 2) to the basic model with only direct relationships (Model 1).

Since all VIF values were below the threshold value of 3, collinearity is unlikely to bias the regression results (Hair et al., 2019). The assessment of the structural model furthermore considers the coefficient of determination (R² value) as a means of the model’s explanatory power as well as the significance and relevance of path coefficients. Even though model fit measures for PLS-SEM should be taken with caution (e.g., see Hair et al., 2019) – Hair et al. (2022, p. 113) even “advise against the use of such statistics in the context of PLS-SEM” – we took into account the standardized root means square residual (SRMR = 0.058) as a guideline to get a first assessment. The approximate model fit index meets the criteria (SRMR <0.08) discussed by Henseler, Hubona, et al. (2016, p. 12).

In the second step, we evaluated the significance and strength of the path coefficients. Concerning the effects of SME on student loyalty, the results show a significant but smaller positive effect on both pWOM and conative loyalty. The hypotheses H1a and H1b are therefore accepted, although the influence is relatively weak. On the other hand, students’ identification with their HEI does not show a moderating effect on the relationship between SME and student loyalty (see Table 3) but a strong and significant direct effect on pWOM and LOY instead. Hypothesis H5 is therefore rejected. The comparison between Model 1 and Model 2 shows that by introducing OI, the effect of SME on student loyalty decreases. About the antecedents of SME, the student’s attitude toward social media content has a significant and strong positive effect, leading to the acceptance of hypothesis H2. Furthermore, perceived reputation of the HEI and student integration also have a significant and positive effect on students’ propensity to interact with the university’s social media content, but their impact on the latter is smaller compared to the attitude construct. Hypothesis H3 and H4 are both accepted.

Table 3. The structural models.

	Model 1 – only direct effects			Model 2 – with moderator
Path	β	Std. err.	t-Value	β	Std. err.	t-Value
ATT → SME	0.329	0.037	8.795***	0.329	0.037	8.795***
POR → SME	0.130	0.041	3.179**	0.130	0.041	3.179**
INT → SME	0.127	0.037	3.420**	0.127	0.037	3.420**
SME → pWOM	0.243	0.038	6.459***	0.075	0.037	2.022*
SME → LOY	0.237	0.037	6.489***	0.122	0.037	3.322**
OI → pWOM				0.538	0.031	17.317***
OI → LOY				0.367	0.038	9.630***
OI x SME → pWOM				0.001	0.028	0.022
OI x SME → LOY				0.012	0.032	0.391

Level of significance: ***p < .001; **p < .01; *p < .05.

ATT= Attitude toward organizational social media content, SME = Social media engagement, POR = Perceived organizational reputation, INT = Student integration, OI = Organizational identification, pWOM = positive word-of-mouth, LOY = Conative loyalty, gSMU = General social media use, DUR = Duration of studies.

The f² effect size identifies if the removal of a specific predictor has a substantive impact on the respective endogenous variable (Hair et al., 2022). As the analysis for Model 2 shows, only attitude toward social media content in the case of SME and OI for the two loyalty constructs have a substantial impact, emphasizing the findings outlined above.

Thirdly, following Hair et al. (2017), in Model 2, the R² of pWOM can be classified as moderate (0.321). Conative loyalty shows an R² value of 0.177, and social media engagement has a value of .171. Hence, they both fall below 0.25 and thus are pretty weak. The R² value levels of pWOM and LOY are mainly due to the direct effect of OI. In Model 1 (without the moderating effect), they are much lower, explaining about 6% of the variance in these constructs. Therefore, the explanatory power of SME is very weak.

Following Yoshida et al. (2018), we verified the robustness of the hypotheses by including two control variables in our model. The results show that only general social media use has a significant and positive impact on SME (β = 0.104, p < .01) and slightly increased the R² value of the construct (+0.009). In contrast, the duration of studies had no significant effect on pWOM or LOY (Model 1: β = −0.036, β = 0.022 respectively; Model 2: β = −0.018 and β = 0.035 respectively, p > .05 for both relationships). Overall, the hypothesized paths endured the inclusion of the control variables and proved to be sound.

In order to check for possible heterogeneity in the data that might alter the relationships between constructs, multigroup analysis (MGA) has been conducted for Model 2 with OI as a moderator (Henseler et al., 2009). As a proxy for age difference – age has been explored in earlier studies regarding social media use and engagement, e.g. by Laor (2022) –, the level of studies (Bachelor’s or Master’s degree) functioned as a grouping variable. Furthermore, a comparison between social media channels has been carried out, since Voorveld et al. (2018) argued for taking the context/respective platform into account. To ensure meaningful group comparisons, the invariance of the measures used in the model has been confirmed following the three-step MICOM procedure developed by Henseler et al. (2016). According to the results of Henseler’s bootstrap-based MGA (Henseler et al., 2009), the path coefficient from SME on pWOM is significantly higher for BA students than for MA students (β = 0.142, p < .01; and β = -0.029 respectively, p > .05). In addition, a difference between the two groups appears regarding OI as a moderator on the relationship between SME and pWOM: In the case of MA students, OI has a significant positive but small effect (β = 0.083, p < .05). The effect is insignificant for BA students (β = -0.055, p > .05). Regarding differing social media channels (Facebook, Instagram, and LinkedIn), pairwise comparisons showed no significant differences between SME and its antecedents or consequences.

Finally, the weak explanatory power of SME on student loyalty opens up the question of whether becoming a follower by liking the HEI’s social media page already induces a higher level of loyalty. We conducted a Kruskal-Wallis test in IBM SPSS Statistics v.28.0.0.0 (IBM Corp, 2021). to examine the differences in student loyalty between students who follow the university on social media, students who do not follow but encounter university postings from time to time, and students who never encountered university postings in their newsfeed before (N = 1.078). Concerning pWOM and conative loyalty, the test showed significant differences among those three groups (Kruskal-Wallis H = 22.240, p < .001, and Kruskal-Wallis H = 18.882, p < .001). Subsequently, post-hoc (Dunn-Bonferroni) tests were performed. Followers show a significantly higher level of pWOM than students who never encounter university postings in their feed or only occasionally when friends like or comment on something. The same pattern reveals itself about conative loyalty, where followers are more willing to support the HEI after graduation, even if the difference is only significant for students who never encounter university updates in their feed. The two groups of non-followers do not differ significantly in their loyalty. It can be concluded that the major leap in loyalty is to be found in the activity of following the HEI.

Summary of the hypotheses tests

Although our results show a significant positive but small effect of SME on conative loyalty and pWOM, the model’s explanatory power is very weak. The two aforementioned endogenous variables are foremost influenced by students’ identification with their university (modeled initially as a moderator). Still, SME and OI together only account for about 32% of the variance in pWOM and about 18% of the variance in LOY. Instead, a comparison between followers of the HEI’s social media pages and non-followers showed that the former group displayed higher levels of loyalty than the latter, indicating the importance of liking the organization’s social media page. With regard to the predictors of SME, the attitude toward the published content is – as expected – the critical antecedent in this model. In addition, perceived organizational reputation and student integration both have a positive but small influence. Overall, the hypothesized predictors only explain 17% of the variance in SME, implying other more significant factors we did not account for. Table 4 summarizes those findings with regard to the formulated hypotheses of this study.

Table 4. Summary of hypotheses.

Hypotheses	Hypothesis result
H1a: Social media engagement → Positive word-of-mouth	accepted
H1b: Social media engagement → Conative loyalty	accepted
H2: Attitude toward org. social media content → Social media engagement	accepted
H3: Perceived organizational reputation → Social media engagement	accepted
H4: Student integration → Social media engagement	accepted
H5: Moderation effect of organizational identification	rejected

Through multigroup analysis, it further emerged that age (represented by level of study) as a specific context factor should be included in the model as a moderator variable. It appeared that SME has a stronger direct impact on pWOM within the group of younger students. On the other hand, older students show a significant but small positive moderating effect of OI on the relationship between SME and pWOM.

Discussion

Since social media platforms have become widely used tools for marketing purposes in HEIs, questions regarding their effectiveness in fostering long-term relationships with current students have arisen. Based on social exchange theory, the present study aimed to explore the effect of students’ social media engagement on their (intended) supportive behavior toward the HEI and to identify drivers of SME. Data for this study were collected through an online survey conducted at a Swiss public university and were subsequently analyzed using PLS-SEM.

Our study partially substantiates prior research findings. Social media engagement appears to play a less critical role in generating loyalty among current university students, similar to the variables of perceived organizational reputation and student integration with respect to SME. On the other hand, the significant positive effect of attitude toward the university’s content on SME was unsurprising and consistent with prior findings in advertising research (e.g., Huang et al., 2013; Kujur & Singh, 2017). In contrast to former research in the business sector (see, e.g., Cheung et al., 2021; Mishra, 2019) and higher education (e.g., Garza Salgado & Royo Vela, 2019), our study identified only a vanishingly small impact of SME on student loyalty. Instead, organizational identification (initially modeled as a moderator) exhibits a comparatively strong direct effect on the latter, reflecting earlier research in this setting (e.g., Abdelmaaboud et al., 2021; Fazli-Salehi et al., 2019). As SME explains only about 6% of the variance in student loyalty, other forums of interaction and exchange seem to be more important. This conclusion aligns with Fujita et al. (2018), who argue that the majority of the relational exchanges in HEIs occur offline, with social media activities supplementing those. SME with the HEI’s digital content cannot compensate for the influence of other touch points on its own. Instead, SME largely depends on the relationship formed beforehand. We conclude that – in the case of enrolled students – SME is more an expression of student loyalty than a tool to foster it. However, discrepancies between our findings and previous research may also be attributed to differences in measurement. Some earlier studies in the business context analyzed SME levels separately and found that not all levels affected facets of brand loyalty (e.g., Cheung et al., 2021; Mishra, 2019). In contrast, we tested the relationship with engagement as a higher-order construct, encompassing both consuming and contributing behaviors, which could have overruled significant effects by one of the two levels. The same applies to the case of pWOM: Piehler et al. (2019) found that only consumption of social media content impacted the former; however, our study revealed only a weak effect when testing with the HOC.

Conclusions

Theoretical contributions

This study contributes to the literature on social media engagement within the higher education context in several ways. It examined the drivers and consequences of SME from the for-profit in the higher education/public sector, introducing new potential influencing factors. Social exchange theory proved to be a valuable theoretical approach, highlighting that, in addition to information, the social media activities of HEI provide additional exchange resources for students, motivating them to engage. For instance, student integration was explored as an antecedent of SME. Although this variable demonstrated only a small effect on SME, primarily due to the dimension of academic integration, the construct should be considered for future studies on SME. This is particularly pertinent as immersion in the HEI’s offline community is characteristic of this service context and has been shown to directly affect loyalty in Russian universities (Iskhakova et al., 2016). Additionally, the study investigated perceived organizational reputation as a driver of SME beyond merely liking the brand page. Contrary to expectations, this variable exhibited only a small significant effect compared to other research on student loyalty (e.g., Helgesen, 2008; Kaushal & Ali, 2020). This finding is surprising given the critical role of reputation in the higher education context in general. A possible interpretation could be that students’ social media engagement is more about communal exchange within their HE community and not foremost about personal branding or differentiation from other universities. Hence, in the HE context, there also seems to be a social reality beyond reputation that we were able to map out in this study. However, this finding might also be linked to the operationalization of SME in our study, where liking the brand page was initially excluded. This assumption aligns with Brech et al. (2017) findings, which indicated that university reputation positively influences the number of social media followers. The same distinction between liking the brand page and other forms of engagement was also evident in its relationship with student loyalty. Another contribution of our study was comparing three groups of students with varying levels of contact with the HEI’s content on social media, revealing that followers also show a higher level of loyalty than non-followers. Further contributions were made through multigroup analysis. By considering social media channels as varying contexts, we demonstrated that engagement on those different channels did not have significantly different impacts on student loyalty. Thus, our results challenge Voorveld et al. (2018), who advocated for this differentiation. The final theoretical contribution of this study pertains to the operationalization of the SME construct itself. Our analysis using PLS-SEM supported that SME is a formative higher-order construct, which contrasts with other studies (Men et al., 2020; Schivinski et al., 2021). If different levels of engagement form distinct groups of activities demanding comparable effort, they are not interchangeable and, therefore, are formative rather than reflective. Accordingly, we suggest that future research should investigate the operationalization of SME, taking these considerations into account.

Practical implications

In addition to its theoretical contributions, this study also entails implications for higher education marketers. Notably, student loyalty in the “during the course” phase of the customer journey (Popli et al., 2023) is profoundly influenced by offline experiences and the bonds formed with the university community, as these are transformational experiences (e.g., discussed by Fujita et al., 2018). This indicates that in organizations with numerous offline touchpoints with their members – in this case, current students – SME constitutes a part of the larger whole and serves a supportive function. Therefore, it is crucial for HE marketers to create a cohesive and integrated (communication) experience across all university-owned touchpoints they can actively influence (see also Palazón et al., 2022). According to our results, students following the social media profiles of their HEI and thereby receiving updates already indicate an increased level of loyalty. This step is important in the long term for the transition into the alumni stage, as it helps maintain contact with the HEI community, thereby strengthening the network and ultimately enhancing loyalty (bringing them into the “loyalty loop” as described by Court et al. (2009)). Activities facilitating this step are, therefore, an important task for marketers. Clark et al. (2017), for example, suggest that universities should specifically encourage students to become social media followers during orientation day for new students as well as at other university events.

The analysis of specific antecedents of SME further reveals actionable opportunities – or exchange resources – for marketing professionals. Firstly, cultivating a positive attitude toward the HEI’s social media posts by consistently sharing interesting and appealing content is essential. Marketers must identify what content resonates with students and understand the characteristics that make posts appealing. For instance, a study by Capriotti et al. (2024) demonstrated that institutional content, such as information about the operational processes of an HEI, induces higher engagement levels compared to functional content like details on teaching and research activities. Additionally, research by Sörensen et al. (2023) emphasized the importance for HEIs to “tailor their social media communication to the respective affordances of different platforms” (p. 1) – including the content. Therefore, marketers need to balance necessary communication content with what students desire, utilizing the most appropriate social media channels (e.g., see Alfonzo, 2021). Regarding post format, Peruta and Shields (2018) provide valuable insights. Their study found that Facebook posts featuring user-generated content contributed to engagement, whereas posts containing hyperlinks or calls to action exhibited lower levels of engagement compared to those without such elements. These considerations enable higher education marketers to foster engagement across all social networking sites with strategies tailored to strengthen the intention to support the alma mater after graduation. Moreover, our study’s insights underscore that an excellent reputation enhances engagement, suggesting that marketers should actively communicate the strengths of their educational programs across various channels. Furthermore, the findings suggest that improving students’ academic integration (e.g., extra-curricular activities) could bolster their SME. If marketers succeed in encouraging students to participate in extracurricular activities through appropriate advertising and communication efforts, it can positively impact engagement with the HEI’s social media content.

In conclusion, higher education marketers should view SME as supportive behavior to establish a lasting digital network among students that continues beyond graduation. With the reduction of many offline university-owned touchpoints in the “post admission” stage, social media becomes a more vital channel for maintaining relationships with alumni.

Limitations and future research directions

Despite the contributions of this study, it is not without limitations. First, the survey was conducted at a single public university in Switzerland, limiting the findings’ generalizability. The results could primarily be due to the specific social media strategy of the HEI under scrutiny, which we did not account for in our model. Depending on the goals of the institution’s digital presence and the content accordingly published on social media, the intensity of engagement and the resulting effects may differ. Furthermore, our sample contained only students who follow the university on one or more social media channels; non-followers who deliberately accessed the university’s profile were not considered. Since accessing the brand page is more intentional than coming across the HEI’s update in the newsfeed, a more substantial effect of the engagement with the online content on loyalty could be assumed. Another limitation concerns the variables of SME and INT. Since they are conceptualized as formative higher-order constructs, their convergent validity is determined utilizing a redundancy analysis. As we lacked an alternative measure of the respective concepts to analyze the correlation (Hair et al., 2019), we could not perform this test. Hence, the formative indicators’ validity was established solely by relying on existing literature. Concerning the output variables, loyalty toward the HEI was conceptualized through pWOM as well as the intention to donate and choose continuing education at the same institution. Mael and Ashforth (1992) showed that alumni can also exhibit other forms of supportive behavior toward their alma mater – some of which could probably be more accurate for the Swiss context (e.g., participating in or volunteering for events held by the HEI, being active in alumni groups, recruiting graduates).

Future research endeavors should include more HEIs across different higher education systems as well as public and private institutions. Especially a comparison between HEIs with presence mode and others with distance learning would be interesting since, in the second case, the on-site exchanges are more limited, which could increase the importance of social media interactions. This approach would allow determining the relative importance of SME for student loyalty across cultures and types of HEIs, potentially uncovering further moderators. Concerning the sample, future studies should expand it and compare the effect of SME on student loyalty between groups of users having various points of contact with the specific content, implicating differing levels of relationship to the institution. In view of the predictors in our model, the relationship between the HEI’s reputation and SME needs to be investigated more closely, as the results of our study surprisingly showed just a small effect from perceived reputation to engagement. Furthermore, complementing the endogenous construct of loyalty intention with new items like the examples cited above, considering the respective HE context, could better depict different realities. Lastly, to what extent the different SME levels influence resulting loyalty needs further investigation to clarify their respective impact.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are openly available in “figshare” at http://doi.org/10.6084/m9.figshare.19317815.

Appendix A

Table

	Component
	1	2	3	4	5	6	7	8	9	10
pWOM_1	.752
pWOM_2	.726
pWOM_3	.762
POR	.496
SME_1		.430				.516
SME_2		.686
SME_3		.776
SME_4		.804
INT_3			.684
INT_5			.609
INT_8			.717
gSMU_4			−.414
INT_4				.692
INT_6				.704
INT_7				.610
OI_1					.647
OI_2	.326				.575
OI_3					.671
OI_4					.710
OI_5	.499				.283
ATT_1						.793
ATT_2						.767
LOY_1			.453				.301
LOY_2							.833
LOY_3							.782
LOY_4	.257						.452
DUR								.734
gSMU_1								.699
gSMU_5								.524
INT_1									.802
INT_2									.587
gSMU_2										.596
gSMU_3										.772

Rotated component matrix from PCA for all items of the conceptual model.

Notes: Extraction method: principal component analysis.

Rotation method: varimax with Kaiser normalization.

KMO-test = 0.810; Bartlett’s test of sphericity = 4567,602, df = 528, p < .001.

pWOM = positive word-of-mouth, SME = Social media engagement, INT = Student integration, OI = Organizational identification, ATT= Attitude toward organizational social media content, LOY = Conative loyalty, gSMU = General social media use, DUR = Duration of studies, POR = Perceived organizational reputation.