‘We’ are more likely to endorse than ‘I’: the effects of self-construal and brand symbolism on consumers’ online brand endorsements

Abstract

Recent research increasingly highlights that consumers engage in online brand endorsements (e.g. Facebook likes) to signal their identity, but has failed to explain why different consumers use this type of signaling to differing degrees. This paper addresses this gap by looking at a culturally constructed individual difference variable, namely self-construal. Self-construal, which can be independent or interdependent, refers to the extent that people define themselves in terms of the relations they have with others. In four studies, this research shows that consumers’ self-construal is related to their intention to endorse brands online. In particular, high levels of interdependent self-construal positively affect consumers’ intention to endorse brands online (Studies 1A & 1B). This effect is mediated by an increased perception of brands’ symbolic value (Study 2). Moreover, this positivity bias toward symbolic brand cues is conditional upon consumers’ brand attitude (Study 3). These findings demonstrate that consumers’ identity plays a central role in their brand perception and brand-related social media use.

Keywords:

• social media
• self-construal
• online brand endorsements
• consumer identity
• consumer behavior

Introduction

If you are a Facebook user, you might have noticed that some of your friends often endorse brands by ‘liking’ brands or sharing brand-related information on Facebook. Others, however, hardly ever show this behavior publicly on Facebook. So what distinguishes your brand-endorsing friends from those who do not endorse brands on social media? In this paper, we argue that consumers’ online brand endorsements are a specific type of identity signaling (cf., Hollenbeck and Kaikati 2012) and that therefore the way consumers define their own identity in terms of their social environment (i.e. self-construal; Markus and Kitayama 1991) influences their willingness to endorse brands online.

Many brands are keen on receiving endorsements from their consumers (Wallace, Buil, and De Chernatony 2012; Schamari and Schaefers 2015), in the hope of sidestepping shortcomings of traditional advertising techniques, such as avoidance and resistance (Fransen et al. 2015; Kaikati and Kaikati 2004). Therefore, the question of why consumers would endorse brands online in the first place receives increasing attention from both industry and academic research. While industry research highlights the functional value of this behavior (e.g. ‘getting updates’, ‘getting discounts’, ‘participating in contests’ – Bosker 2013), recent academic research shows that these motivations have small explanatory value (Wallace et al. 2014). Instead, academic literature focuses on more identity-related drivers of this behavior, including self-presentation (e.g. Bazarova and Choi 2014; Hancock and Dunham 2001; Wilcox and Stephen 2013) and identity management (Hollenbeck and Kaikati 2012; Belk 2013).

Social network sites like Facebook provide a platform where consumers can easily use brands to signal their identity to a wide audience. On Facebook, people naturally spread identity-related information to their peers, like their daily emotions (Manago, Taylor, and Greenfield 2012), but they can also visibly interact with brands. Some users, for instance, like brand pages, post photos with branded content, or become part of brand communities. These two behaviors are often related to each other: recent research shows, for example, that consumers deliberately use brands on Facebook to manage and create their self-identities (Hollenbeck and Kaikati 2012; Bernritter, Verlegh, and Smit 2016).

With a global medium like Facebook, it is important to note that consumers around the globe differ widely in terms of how they view themselves. These differences may have important implications for advertising practice, especially on social media (see Okazaki and Taylor 2013, for an excellent review). In cross-cultural psychology, self-construal (SC) has been recognized as one of the major dimensions along which consumers differ, especially in the social domain. With regard to SC, a distinction is made between people with predominantly an independent (IND) or interdependent (INT) SC. INDs see themselves as unique and separate from others whereas people with an INT SC see themselves as being similar and closely connected to others (Markus and Kitayama 1991). These different views of the self go hand-in hand with different ways in which people express themselves and in how they perceive the world around them. INDs will seek ways to emphasize their uniqueness and focus their attention on focal objects and internal attributes whereas INTs will seek ways to highlight their connectedness and focus their attention on relations between people and objects (Markus and Kitayama 1991; Nisbett 2010).

Based on this notion, we suggest that consumers’ SC plays an important role in determining the likelihood that they utilize brands to communicate their identity on social media. Studying this construct in more detail will help us to understand some of the intricacies of social media advertising in a global world. In particular, we argue below that differences in SC affect the way a brand is perceived, specifically its symbolic value, and consequently determine consumers’ inclination to endorse brands online. We also posit that this SC-induced amendment of consumers’ brand perception depends on their brand attitude toward the endorsed brand. In doing so, we contribute to the literature in at least three important ways: first, we demonstrate how SC drives consumers’ online brand endorsements, which supports the notion that consumers differ in the way they use brands as communicators of their identity; second, we show that the way consumers see themselves (i.e. SC) affects their brand perception (i.e. brand symbolism). Finally, we demonstrate how consumers’ brand attitude affects the way SC influences brand perception, which reveals that the process that underlies consumers’ brand perception works differently for INT (vs. IND) consumers.

Conceptual framework

Consumers’ online brand endorsements

Consumers’ online brand endorsements are a very recent phenomenon, which gains increasing attention in the literature. A common example is the liking of brand pages on Facebook (cf., Bernritter, Verlegh, and Smit 2016). Obvious drivers of this behavior such as consumers’ informational goals or commercial incentives from brands to acquire endorsements from consumers have been found to explain only a small portion of consumers’ online brand endorsements (Wallace et al. 2014; Muntinga, Moorman, and Smit 2011). This suggests that other drivers of this behavior are at play.

An emerging body of research is dedicated to the role of consumers’ identity in their brand-related (online) behavior. Consumers use brands as means to communicate their identity to others (Belk 2013; Belk 1988; Berger and Heath 2007), and make inferences about others based on their consumption behavior (Reed et al. 2012). It has been recently demonstrated that these mechanisms are also at play in the online domain. Consumers deliberately use brands on Facebook as subtle cues to communicate their identity to others (Hollenbeck and Kaikati 2012) and self-expressive motives play a central role in consumers’ decision to like brands on Facebook (Wallace et al. 2014). It is suggested that self-expressive brand-related behavior might be particularly popular on social media because consumers realize that social media are more effective to communicate their own identity to others than offline brand identifiers since they allow to reach a considerably broader audience (Hollenbeck and Kaikati 2012). The idea that online brand-related behavior is more revealing than offline behavior finds support in recent work that demonstrates that consumers perceive online brand endorsements as more risky compared to offline brand endorsements (Eisingerich et al. 2014).

Identity-related motives are also reflected in the brand-related drivers of consumers’ online brand endorsements. Consumer brand identification (i.e. ‘a consumer's perceived state of oneness with a brand’; Stokburger-Sauer, Ratneshwar, and Sen 2012, 407), which leads to increased brand loyalty and advocacy, is considered to be an important antecedent of consumers’ online brand endorsements. The same goes for brand love (i.e. ‘the degree of passionate emotional attachment a satisfied consumer has for a particular trade name’; Carroll and Ahuvia 2006, 81), which is identified as an important predictor of consumers’ online brand endorsements as it enhances consumers’ brand advocacy, positive word of mouth, and engagement with the brand (Bergkvist and Bech-Larsen 2010). Another factor that is strongly related to consumers’ identity is what a brand communicates about its endorsing consumers. Bernritter, Verlegh, and Smit (2016), for instance, showed that warmth of a brand is a crucial driver of consumers’ intention to like brand pages on Facebook. It is argued there that consumers aim to communicate their warmth (i.e. that they are good people; Fiske et al. 2002) to others by means of their online brand endorsements.

Self-construal and brand endorsements

SC refers to the way we view ourselves in terms of our relationships with others, and has received most attention in the context of cross-cultural research (Markus and Kitayama 1991). With an IND SC, the individual is seen as separate from others and defined by a unique set of internal attributes. Conversely, with an INT SC the self is seen as connected to others and is defined in terms of one's social relations. Although originally most used to contrast western with eastern views of the self, it is increasingly recognized that both types of SC reside within every individual and that the one can become dominant over the other depending on the situational demands (Gardner, Garbiel, and Lee 1999; Oyserman and Lee, 2008; Singelis, 1994).

These different SCs go together with different motivational goals (Gardner, Gabriel, and Lee 1999; Markus and Kitayama 1991). An IND SC goes together with a focus on self-actualization, individuality, and personal success, whereas an INT SC leads to a focus on maintaining harmony, connectedness, and assimilation toward close others (Markus and Kitayama 1991). In expressing themselves, consumers with an IND SC might thus seek ways to emphasize their individual uniqueness; whereas consumers with an INT SC might seek ways to emphasize their connectedness with close others.

Indeed a variety of research has shown different effects of SC on consumers’ behavior in relationship to others. It has been demonstrated for example, that an INT SC is associated with higher degrees of impression management (Lalwani, Shrum, and Chiu 2009), an increase in self-disclosure on Facebook (Chang 2015), and that it positively affects consumers’ intention to engage in electronic word of mouth about brand communities (Lee, Kim, and Kim 2012). Earlier research suggests that consumers’ SC can influence their responses to advertising appeals. For people with an INT SC, advertisements that use INT group related appeals that communicate values such as harmony, in-group benefits, and family integrity, are perceived to be more persuasive. Contrary, for IND consumers, advertisements that communicate individual values such as personal success, independence, and individual benefits are more persuasive (Han and Shavitt 1994). Swaminathan, Page, and Gürhan-Canli (2007) showed that consumers’ SC determines how consumers form brand relationships. While INDs rather form relationships with brands based on an individual-level connection (e.g. self-concept connection), INTs tend to form brand relationships based on group-level connections (e.g. country of origin connection). Similarly, Escalas and Bettman (2005) found that a brand image that is consistent with an out-group has a stronger negative effect on consumers’ self-brand connection for people with an INT (vs. IND) SC. As affiliations with brands signal group-level connections (Escalas and Bettman 2003; Escalas and Bettman 2005), we suggest that consumers with an INT SC are more likely to endorse brands online than consumers with an IND SC.

H1:

An interdependent self-construal has a positive effect on consumers’ intention to endorse brands on social media.

Brand symbolism

Brand symbolism (i.e. the extent to which a brand symbolizes the person who is using it) has been shown to be essential in the construction of self-identity via usage of brands and the formation of self-brand connections (Escalas and Bettman 2005). In a slightly more abstract way, brand symbolism could also be described to be referring to the ability of a brand to signal consumer identity. It is not that much about the meaning of the brand for the consuming individual, but rather about the inferences that observers (can) make about consumers based on their usage of the brand. Brand symbolism has been demonstrated to increase the effects of individual (Bernritter 2016) and social identity signals (Escalas and Bettman 2005; Berger and Heath 2007). This is important as brand symbolism can communicate group membership.

INT SC motivates consumers to fit in with others (Markus and Kitayama 1991) and to seek out ways to signal group membership (Escalas and Bettman 2005). In line with research that demonstrates that goal-relevant cues in the environment do indeed receive more attention than cues that are not goal relevant (Aarts, Dijksterhuis, and de Vries 2001), we propose that this chronically active goal enhances INT consumers’ attention toward cues in their environment that may be used to signal group membership, such as symbolic brands. Next to enhanced attention toward group-level related cues, consumers with an INT SC might also attach more value to these cues, and consider them to be more ‘telling’ about a person. Research shows that people value objects based on their goal-relevance (Markman and Brendl 2000). The symbolic value of a brand, as communicator of group membership, might thus be higher for consumers with an INT SC, compared to those with an IND SC.

This notion finds support in cross-cultural research that has shown that Chinese, who in general are considered to be more INT, compared to Americans, who in general are considered to be more IND, have a tendency to classify and describe objects in terms of their category membership and internal attributes (A and B both have X so belong to category Y) as opposed to their functional or part–whole relationship (A is to B or A is part of B) (Chiu 1972; Nisbett et al. 2001; Ji, Zhang, and Nisbett 2004). The difference in focus on relationships versus attributes can also be found in studies comparing traditionally INT cultures with IND cultures in terms of self-description. When asked to describe themselves, Americans are more likely to list internal attributes such as tallness and stubbornness whereas East-Asians are more likely to describe themselves in terms of their societal roles, for instance as parent of a child and employee of the company they work for (Trafimow, Triandis, and Goto 1991; Bochner 1994).

We propose that this difference is also relevant when processing brand-related information. When thinking about brands one can focus on purely the internal attributes the brand stands for (e.g. Harley-Davidson as a brand producing motorcycles) or focus on the functional relationship this brand has with its users and larger society (Harley-Davidson as a brand used by those wishing to rebel against societal convention). As brand symbolism is closely related to observing the functional relationship the brand has with its consumers, we propose that INTs will, in general, perceive brands as more symbolic.

H2a:

An interdependent self-construal has a positive effect on consumers’ perceptions of brand symbolism.

As consumers with an INT SC have the chronically active goal to fit in with others (Markus and Kitayama 1991), their enhanced perception of brand symbolism might also affect their online brand endorsement behavior. Environmental cues have been shown to have the capability to activate and strengthen goals in consumers (for an overview, see Dijksterhuis and Aarts 2009). Symbolic brands might act as such cues, enhancing the accessibility of consumers’ goal to fit in with others. Within consumers with an INT SC, exposure to symbolic brands might thus enhance the accessibility of this goal and lead to behavior that is instrumental to achieve this goal: in this case, online endorsement of brands. Moreover, expression of actual and desired identities is one of the prime drivers of consumers’ endorsements of brands on Facebook (Hollenbeck and Kaikati 2012; Bernritter, Verlegh, and Smit 2016). Because brand symbolism represents the extent to which a brand represents a consumer's identity, and because INT SC consumers are more likely to perceive brands as symbolic, we propose that consumers with an INT SC are more likely to endorse brands on social media in order to express and communicate their identity to others. In other words, we suggest that consumers’ perceptions of brand symbolism mediate the positive effect of an INT SC on online brand endorsements.

H2b:

Perceptions of brand symbolism mediate the effect of interdependence on consumers’ intention to endorse brands on social media.

Overview of studies

We conducted four studies to test our hypotheses. The first two studies (1A and 1B) establish a positive correlation between a high INT level of SC and consumers’ intention to endorse brands online. In Study 2, we demonstrate that this positive association with interdependence is mediated by consumers’ perception of brand symbolism. Finally, in Study 3, we manipulated consumers’ SC to rule out causality concerns and show that the mediation by brand symbolism is conditional upon consumers’ brand attitude.

Study 1A

Method

Two hundred and nine students of a large Dutch university participated for partial course credits or financial compensation (M_age = 20.8, SD_age = 3.33; 80.9% female). Participants needed to be Facebook users in order to be allowed to participate. Seven participants were excluded from the analyses because they did not complete the questionnaire. The study was conducted online together with some unrelated tasks and consisted out of four blocks, which were presented in random order (i.e. one block per brand). Within each block, participants first were exposed to the logo of one brand, which was randomly selected out of a pool of eight brands in total. For each brand, participants had to indicate how likely they were to like it on Facebook on a 7-point Likert scale (1 = not likely at all/7 = very likely). Each participant got thus exposed to four brands in total. The brands were chosen based on a list of the 100 strongest brands in the Netherlands (BrandAsset Consult 2013). Instead of limiting our results by only focusing on one brand per analysis, and being susceptible to effects of individual brand characteristics, we were aiming for a measurement of consumers’ overall online brand endorsements (cf., Bernritter, Verlegh, and Smit 2016). Therefore, we calculated the average of consumers’ scores on their intention to like the different brands as dependent variable in this and all subsequent analyses. Cronbach's α for this measure was high (.85), which supports our suggestion that a general assessment of consumers’ willingness to endorse brands is a reliable way to examine this phenomenon. After indicating their intention to endorse the four brands, participants filled in Singelis’ (1994) scales for INT and IND SC's (α_INT = .65; α_IND = .62; M_INT = 4.45, SD_INT = 0.60; M_IND = 4.58, SD_IND = 0.62).

Results and discussion

Supporting Hypothesis 1, a regression analysis revealed a positive association between consumers’ level of INT SC and their intention to endorse brands on social media (β = 0.15, t(199) = 2.17, p = .031). Consumers with a high level of interdependence indicated higher intentions to endorse brands online. Consumers level of IND SC had no effect on their intention to endorse brands online (β = 0.06, t(199) = 0.816, p = .415).

In Study 1A, we found an effect of consumers’ INT SC on their intention to endorse but not of consumers’ IND SC. Singelis’ (1994) scale for INT and IND SC has, however, been criticized for several reasons (Levine et al. 2003; Grace and Cramer 2003). In Study 1B, we therefore decided to solely focus on consumers’ INT SC and aimed to replicate the findings using another scale, namely the Relational-Interdependent Self-Construal Scale (RISC; Cross, Bacon, and Morris 2000). To further enhance the generalizability of our findings, we also used a different set of brands.

Study 1B

Methods

Dutch students participated for partial course credit or financial compensation (N = 235; M_age = 21.7, SD_age = 4.31; 75.9% female). Eleven participants were excluded because they were not Facebook users. The procedure was identical to that of Study 1A, with three exceptions: consumers’ SC was measured using the RISC scale (Cross, Bacon, and Morris 2000;α = .824; M = 4.71, SD = 0.68) before measuring intention to endorse. Consumers’ intention to endorse was measured using a 101-point slider scale (‘what is the chance that you would like this brand on Facebook?’; 0%–100%; cf., Bernritter, Verlegh, and Smit 2016), and eight different for-profit brands were used.

Results and discussion

As in Study 1A, a regression analysis provided further support for Hypothesis 1 and revealed a positive association between consumers’ INT level of SC and their intention to endorse brands on social media (β = 0.21, t(222) = 3.25, p = .001).

The results of Study 1B replicate the findings of Study 1A using a different scale and a different set of brands. After establishing the positive association between consumers’ INT SC and their intention to endorse brands online, Study 2 examines the mediating role of brand symbolism that was proposed in our theoretical framework.

Study 2

Methods

Sixty-nine Dutch students voluntarily participated in a pen and paper survey (M_age = 22.8, SD_age = 3.37; 58% female). All participants were Facebook users. The procedure was similar to Study 1B except from that six different target for-profit brands were used, that we used the 7-point Likert scale from Study 1A to measure consumers’ intention to endorse, and that participants’ perception of brand symbolism was assessed. Brand symbolism was measured using the two-item Brand Symbolism scale (Escalas and Bettman 2005) and was included in between the RISC (M = 4.89, SD = 0.76) and the intention to endorse measurement. This brand symbolism scale (Cronbach's alpha > .70 for all brands) assesses consumers’ perception of brand symbolism by 7-point Likert scales on the two items: ‘How much does this brand symbolize what kind of person uses it?’ (1 = not at all symbolic/7 = highly symbolic); and ‘To what extent does this brand communicate something specific about the person who uses it?’ (1 = does not communicate a lot/7 = communicates a lot).

Results and discussion

A regression analysis revealed that consumers’ INT SC (β = 0.22, t(66) = 2.09, p = .041) and perceived brand symbolism (β = 0.45, t(66) = 4.19, p = .000) had a positive effect on their intention to endorse. Supporting Hypotheses 2a and 2b, a mediation analysis using Hayes’ (2013) PROCESS SPSS macro demonstrated that the effect of consumers’ INT SC was mediated by their perception of brand symbolism (indirect effect = 0.19, boot SE = 0.99, 95% BCBCI [0.03, 0.42]; direct effect = 0.35, SE = 0.17, 95% BCBCI (bias corrected bootstrap confidence intervals) [0.02, 0.68]). Thus, consumers who scored higher on interdependence perceived the brands as more symbolic, which in turn increased their intention to endorse these brands. These results replicate the findings of the first two studies and in addition show that these findings are mediated by INTs’ enhanced perception of brand symbolism.

Study 3

An INT SC is related to emphasis on how one is embedded in a social environment and one's connectedness to others. The focus of attention within INTs lies thus in their social environment. Contrary, an IND SC emphasizes the self and realization of the own needs (Markus and Kitayama 1991; Singelis 1994). This difference may have implications for the type of brands that consumers perceive as symbolic. For IND consumers, perceptions of a brand's symbolism might be closely related to their own preference for the brand (i.e. brands that are more liked are perceived as more symbolic). For INT consumers, this relationship is less likely, and all brands might have equal symbolic potential as they can serve as group-level identifiers (Escalas and Bettman 2003, 2005; White and Dahl 2006). Consequently, we hypothesize

H2c: Consumers’ perceptions of brand symbolism are dependent on their own brand attitude, but only for consumers with an independent self-construal.

In sum, we propose the following model (Figure 1):

Since the former studies were correlational, we aim to rule out concerns regarding causality of our findings by means of an experiment in which we manipulate SC by priming consumers with either IND or INT. Furthermore, we examine whether this process is conditional upon consumers’ brand attitude, as proposed in our Hypothesis 2c.

Methods

One hundred and sixty-four Dutch students (M_age = 21.3, SD_age = 2.25; 68.4% female) participated in a lab experiment for partial course credit or financial compensation. All participants had a Facebook account. Nine participants had to be excluded from the analyses, as they did not finish the experiment. The participants were randomly assigned to one of the two conditions of the between-subjects single factor design (INT vs. IND SC). We used a pronoun-circling task to activate either an IND or INT SC in our participants (Brewer and Gardner 1996; Gardner, Gabriel, and Lee 1999). The task consists of a paragraph describing a city trip with either only IND (I, me, etc.) or INT (We, us, etc.) pronouns. Participants are asked to circle all the pronouns. After completing this task, participants proceeded with the subsequent measurements, which were identical to that of Study 2. We additionally assessed participants’ brand attitude by a 3-item differential scale (good/bad; positive/negative; likeable/not likeable). Moreover, we again used four different for-profit brands and measured consumers’ intention to endorse by means of the 101-point scale from Study 1B.

Results and discussion

Moderated mediation analysis using Hayes’ (2013) PROCESS SPSS macro revealed that consumers that were primed with an INT SC perceived brands to be more symbolic than those primed with an IND SC (B = 3.79, SE = 1.28, 95% BCBCI [1.26, 6.32]). Supporting Hypothesis 2c, this effect was conditional upon consumers’ brand attitude (interaction: B = −0.77, SE = 0.26, 95% BCBCI [−1.28, −0.26]; Figure 2). That is, for consumers who were primed with an IND SC, their own brand attitude positively affected their perception of brand symbolism and by means of this their intention to endorse (indirect effect = 3.56, boot SE = 1.31, 95% BCBCI [1.52, 6.88]). Contrary, for consumers primed with an INT SC, the own brand attitude did not matter (indirect effect = −0.11, boot SE = 0.92, 95% BCBCI [−2.16, 1.46]). The effect of consumers’ SC on their perception of brand symbolism is thus dependent on their attitude toward the brand.

Figure 1. Proposed moderated mediation model. The dotted line implies that the direct effect of SC on online brand endorsements vanishes when testing the whole moderated mediation model.

Figure 2. Consumers’ perception of brand symbolism as a function of self-construal and brand attitude.

Conclusion and discussion

The present studies investigate to what extent consumers’ SC affects their propensity to endorse brands on social media. In doing so, they provide promising first evidence that online brand endorsements might be a phenomenon that considerably differs between individuals, namely in how they construct and express their identity. This research demonstrates that consumers with a higher level of INT have higher intentions to endorse brands on social media compared to consumers with a lower level of INT. We are furthermore providing first insight into the underlying process of this phenomenon by showing that it partially relies on INTs’ tendency to perceive brands as being more symbolic. Moreover, we show that perceptions of brand symbolism depend on consumers’ brand attitude for those with an IND SC.

The finding that consumers with an INT SC are more likely to endorse brands on social media than consumers with an IND SC supports the notion that brands can rather be seen as important group identifiers (e.g. Escalas and Bettman 2005; White and Dahl 2007). Brands unify their consumers since they all have at least one thing in common, the brand. That INT facilitates consumers’ online brand endorsement behavior might also suggest that online brand endorsements might be a phenomenon that not only differs between individuals but also between cultures, since SC is often associated with cultural differences (Markus and Kitayama 1991). This might eventually urge us to rethink our understanding of branding and social media marketing in a global context, which currently rather assumes that there is ‘the’ global consumer instead of various different kinds of consumers, which all differ in how and why they engage with brands on social media.

That consumers’ SC alters the perception of the symbolic value of brands adds to a body of literature that describes the impact of SC on values expressed in advertisement (van Baaren and Ruivenkamp 2007) and more broadly perception and cognition (Nisbett 2010; Kastanakis and Voyer 2014). To our best knowledge, however, we are the first to demonstrate that SC can alter consumers’ brand perception in general. INT consumers generally tend to perceive brands to be more symbolic than IND consumers. They recognize the symbolic value of brands, which seems to stay hidden for IND consumers. This suggests that the heuristics that consumers use in order to process brands and brand-related information differ dependent on their SC. While INT consumers tend to perceive brands to be symbolic because they can signal group membership in general (i.e. groups are very symbolic for INT consumers), IND consumers estimate the symbolic value of brands based on their own attitude toward the brand. In other words, INTs seem to try to see groups in brands, whereas INDs seem to rather look for reinforcement of the self in brands.

Limitations

This research also has some limitations. First of all, we measured online brand endorsements not by letting participants really like the brands, but assessed their intention to like the brand pages on Facebook. Although intention measurements are widely accepted and used in the field (e.g. Green, Krieger, and Wind 2001; Van den Poel and Leunis 1999; Gronholdt, Martensen, and Kristensen 2000), they might not always reflect real behavior.

Moreover, using student samples in all four studies limits the generalizability of the findings. Our sample might be more homogenous in terms of SC as the general population and because of the accompanying limited age range their interaction with social media could be generation specific.

Also, in the current studies, we used popular mainstream brands. One might argue that consumers, who score high on IND would be more likely to endorse niche brands than those mainstream brands because they might have a higher need for uniqueness (Burns and Brady 1992). Moreover, in the present studies, we assessed whether brands have symbolic value, but not what they symbolize. One might argue that, for instance, in-group versus out-group symbolism might form a boundary condition in this context.

Directions for future research

A possible direction for future research concerns the question of whether SC also affects the way other endorsers are perceived. Consumers make inferences about others on the basis of their consumption behavior (Belk 1988). Likewise, one might expect that consumers make inferences about other's identity on the basis of their public affiliations with brands such as online brand endorsements. As high levels of INT are related to increased perceptions of brand symbolism, one might assume that this extends to perception of others and that INTs make stronger inferences about other people's personality on the basis of their endorsements. On the other hand there is a large body of literature on attribution showing that INDs are more likely to make dispositional attributions compared to INTs who prefer to make situational attributions and are therefore more likely to always contextualize when inferring from behavior and thus less prone to the fundamental attribution error (Choi, Nisbett, and Norenzayan 1999; Lee, Hallahan, and Herzog 1996). Following from that, one might thus argue that it is more likely that INDs would make strong personality inferences on the basis of brand endorsements and that INTs would consider other factors that might have led the endorser to this behavior.

As a next step, it would also be very worthwhile to extend this research with cross-cultural data. We have looked at an individual difference variable that differs across individualistic and collectivistic cultures but have gathered data within one and the same culture (Markus and Kitayama 1991). Even though SC is an integral part of what makes these cultures different, other factors like language and technological advancements also differ between cultures and these might also affect consumers’ online brand endorsement behavior. Gathering cross-cultural data would not only provide insight into replicability but would also give us insight in whether these other differences should be considered.

Another interesting endeavor for future research might be to not only investigate the effects of SC on consumers’ online brand endorsements on the profile level, but also to examine how SC affects brand content diffusion on the message level. For instance, are the same processes at play when it comes to retweets on Twitter, or sharing of branded content on Facebook? While endorsing the brand (i.e. liking the brand's profile) can be considered to be an initial step to get exposure to branded content in general, it would be interesting to investigate whether different factors, such as the content of the message (Araujo, Neijens, and Vliegenthart 2015) play a role in the diffusion of branded content and the role of SC herein.

Practical implications

The current research has also important implications for the field of advertising and social media marketing. Assuming that the findings translate into cross-cultural differences, they might have great applicability for practitioners. The results suggest that using consumers’ online brand endorsements as marketing technique might be dependent on the predominant cultural mindset in the target market. Marketers addressing consumers across the globe as well as marketers aiming to solicit endorsements in one cultural group can use this knowledge to refine their social media marketing strategies. For markets with a predominant INT mindset, it might be beneficial to emphasize concepts like connectedness and harmony in order to facilitate consumers’ online brand endorsements. In markets with a predominant IND mindset, brands should rather focus on improving consumers’ attitude toward them to reach that goal.

As this paper shows, SC can also be manipulated in the situation and this might thus be used to guide content and form of marketing messages tailored to the target market, independent from the predominant cultural mindset. If brands, for instance, aim to acquire endorsements from their consumers on social media, they should address their consumers in a way that implies a sense of ‘we-feeling’ in order to trigger INT in consumers.

To conclude, these studies demonstrate that SC is an important predictor of consumers’ online brand endorsements. We demonstrate for the first time how individual differences can affect consumers’ tendency to endorse brands on social media and shed light on the underlying processes of this phenomenon. SC does not only affect consumers’ behavior on social media, but also affects how they perceive brands. By means of increasing perceptions of symbolism, priming consumers with INT might be a worthwhile strategy to facilitate their online brand endorsements.

Acknowledgements

The authors would like to thank the guest editor Martin Eisend and three anonymous reviewers for their valuable advice.

Disclosure statement

No potential conflict of interest was reported by the authors.