Social Media and Me: How Community Identity Influences Click Speech

ABSTRACT

Social media has become an integral part of our society. So did participation in social media communities. Click-speech, the simple expression of opinions via a click in the form of likes and shares, emerged as a dominant feature of various social media platforms. In this research, we investigate how social media users use click-speech within social media communities through the lens of community identity and how it shapes click-speech behavior. Using a scenario-based study, we show that community identity strongly influences click-speech to the point where content is liked and shared even when the community member disagrees with the post or content. We contextualize the construct of fear of isolation from the spiral of silence theory within identity theory. In this context, the fear of being isolated from the community is salient, and click-speech evolves from a function of endorsing content to a socially driven behavior of fitting in.

KEYWORDS:

• Social media
• community identity
• fear of isolation
• click-speech

Introduction

The storming of the Capitol on January 6, 2021, in the United States demonstrated how inflamed narratives in online communities can cause harm in the real world. Online social media communities played a significant role in the mobilization and eventual storming of the Capitol, resulting in deaths and deep shock for the American people.^1,2 This upsetting event calls for a more profound knowledge of how harmful opinions are leveraged among community members to the point where they begin to take real-life actions with real-life consequences.

Social media platforms like Facebook and Reddit allow users to form online communities. On Facebook, this takes the form of private groups. On Reddit, it takes the form of SubReddits. To understand how online communities work on social media, we first must understand the primary mechanism that drives social media interaction. Click-speech, or liking and sharing, is a nonverbal communication implemented by various social media platforms.^3,4 It is a form of effortless attitudinal expression toward content presented on social media. Click-speech is a popular metric to evaluate whether the content is famous and accepted by social media users.⁵ In other words, it indicates how well social media posts are received. In this context, harmful opinions may be exploited by the prevalence of high numbers of likes or shares.

Previous research has examined the circumstances under which click-speech is used. Social media users click like or share buttons for reasons including self-presentation to peers,⁶ to show social support,⁷ or to signal to others that a post has been seen.⁸ Thus, click-speech is a way to show a “virtual presence” and to keep other users thinking that one is part of the social network or community.

Click-speech, a behavior used to create bonds with other community members rather than endorse specific content, raises the question of why members of social media communities engage in this behavior. Since click-speech is strongly tied to the social aspect of the community, we propose community identity as a driving factor for click-speech. Community identity, which is the feeling of belonging to and behaving following the community, provides a concept of how individuals see themselves in communities.^9,10 Community identity leads to depersonalization, in which information is viewed through the lens of the community,^11,12 which diminishes one’s point of view in favor of that of the community. Consistent with click-speech, which focuses on the social aspect and not specifically on the content of posts, community identity also focuses on in-group dynamics and related behaviors.

With community identity as part of the identity, members find pleasure and pressure in behaving within the social media community. Failure to live up to community expectations (i.e., supporting other members) may lead to fear of isolation from the community. We argue that because community identity is an integral part of the self, individuals act in ways that prevent the potential loss of such community identity, i.e., fear of being isolated from the community. Fear of isolation is the primary mechanism of the spiral of silence theory, which explains people’s tendency to behave socially and not to be isolated from the community.^13,14

Literature suggests that when observing behavior on social media, we should investigate in a more detailed manner.¹⁵ Research already provides a good overview how social media users engage within the public space of social media, but less so in more clearly defined communities as it is the case of Facebook groups or SubReddits. We argue, in order to understand the dynamics within such communities we have to turn to a more socially driven approach to explain behavior. Therefore, with the spiral of silence theory and community identity lens, we account for this unique communication environment in order to draw a clear picture of the socially driven behavior of opinion disclosure. We explain click-speech with community identity as the main driver and fear of isolation as the mediator, which includes the negative emotion associated with community identity. We do this in the extreme situation of community opinions that do not match one’s own opinion to show the irrelevance of content and how strong communities can influence the behavior of their members. Thus, we pose the following research question:

RQ:

How does online community identity on social media influence click-speech against the own opinion?

We use the spiral of silence theory¹³ to address the fear of isolation and social identity theory¹⁰ for community identity. We conducted a scenario-based study with three scenarios where participants could use click-speech against their own opinion to support their community. In total, we gathered 164 replies from the study participants. To analyze the data, we used structural equation modeling with the partial least squares approach.¹⁶

Theoretical background

Click-speech on social media: Liking and sharing

Click-speech is a form of opinion expression that uses liking and sharing buttons.¹⁷ Reasons for click-speech are multi-faceted. Individuals use click-speech when they deem the content worthy of being visible to their peers¹⁸ or to show which content they find entertaining or gratifying.⁶ Furthermore, they seek discounts on products or like Facebook pages to get updates on new content. In this regard, liking or sharing does not necessarily mean that users like certain content⁴ but use the feature as a bookmark to gain monetary advantages. Click-speech is used to satisfy the need to seek status from community networks.^19,20 The presented content drives individuals and carefully considers the public evaluation of their online self and affiliations.^21,22 Therefore, individuals use²⁰ click-speech for compliance purposes, such as feeling the urge to show their acceptance toward an individual rather than the shared content.²² This fact is furthermore supported by the fact that social media users report using click-speech out of habit without considering the topic.⁸ These instances indicate that social media users use click-speech in a context heavily characterized by community relationships and how they identify within such communities.

Based on the provided research, click-speech has a significant social aspect. Research showed that community identity strongly influences behavior, especially in social contexts.²³ Research on how community identity affects click-speech is scarce in this regard. In this research, we focus on the community identity of individuals on social media and how specifically this identity influences user click-speech behavior.

Identity on the individual and community level

Identity is the internalized shared meanings and expectations concerning individuals’ behaviors. Such identity tries to answer the question, “Who am I?.”^24,25 Based on their identity, individuals derive who they are and how they should act concerning others, for example, in a role (e.g., a fiancé), a person (e.g., a good person), or within social groups (e.g., member of a social media community).²⁶ Research broadly identified two levels on which identity can emerge: identity on the individual and community levels.²⁷

Identity at the individual level refers to values that individuals hold independently of others. It represents traits or characteristics that make an individual different from others, such as having one’s name, a favorite movie, being an introvert, or being the older sister. Research has examined how individual-level identity influences behavior when embedded in a social network.^23,28 Individuals have a concept of what it means to fulfill a specific role (e.g., being knowledgeable about a topic as a professor) or try to behave in ways that conform to traits and values associated with personal identities, such as being a good person (e.g., trying to help people in need). In social media, a role might be a community moderator who tries to enforce specific community rules (e.g., sanctions for discriminating against community members). In social media, a person knowledgeable about nutrition and always helps fellow members of a health community could be associated with an individual-level identity. Both identities distinguish individuals from their social media peers.

Identity on the community level refers to identifying an individual through a group.²³ In this research, indicated by the research question, we focus on community identity as one major influence on click-speech. We do so, since communities on social media inherit a unique social context that is more encapsulated than the public space of social media. Community identity shares traits or characteristics that make people similar to others. This can be seen in communities with similar interests, such as political affiliation, ethnicity, or gender. Individuals develop many identities as they interact with their environment.²⁹ This identity manifests in becoming like other group members.³⁰ Individuals engage in depersonalization, which means that individual values or beliefs are suppressed to satisfy the group’s welfare.^11,12 In political communities with explicit values and narratives, individuals may identify more strongly with community values and reproduce certain narratives (e.g., the narrative of a stolen election). The process of depersonalization supports the notion that in community identity, differences between individuals are minimized while similarities are reinforced.

Spiral of silence theory

Because community identity is an integral part of the self, being isolated from the community (i.e., one no longer fits in) is a real threat to community members. The spiral of silence theory explicitly taps into this fear of isolation as it describes the human nature to fear separation and isolation from our fellow human beings (in this case the social media community).¹³ Therefore, it relies on fear of isolation, which incorporates this fear. This fear arises from having an opinion on an issue that is incongruent with what is perceived as public opinion. It is said that every person uses a quasi-statistical calculus to determine what the mainstream view on a given topic is. We argue that this quasi-statistical calculus also occurs within social media communities. The spiral of silence theory depends heavily on the social environment in which opinions are embedded. Therefore, we provide further context and explore its applicability in social media communities.

Due to the fear of isolation, individuals who do not follow public opinion censor themselves.^31–33 As a result, public opinion remains relevant in public discourse, while alternative views are silenced. This dynamic reinforces the preexisting public opinion and cements its existence, further inducing fear of isolation in individuals with incongruent opinions. The mechanism of fear of isolation is an emotional state that arises only in specific situations, is not stable over time, and disappears after a particular situation.^31,33

The spiral of silence theory found application in social media contexts and could predict the disclosure of opinions.^31,32,34 More specifically, the theory predicts click-speech.^17,31 Fear of isolation discouraged individuals from using likes to disclose their opinion.³¹ However, this research was conducted in a setting not associated with a community where a post was embedded. In this regard, the research provides a broader context for applying the spiral of silence theory.

By integrating the spiral of silence theory, we account for the adverse effects of community identity. While community identity is a necessary part of the self and individuals want to be part of a community, the spiral of silence theory suggests that being part of a community creates pressure to stay in the community. This pressure manifests as a fear of isolation since isolation from the community results in a loss of community identity. Only when individuals are part of a community can this fear of being isolated from that community exist.

Research model

The research model as depicted in Figure 1 is based on identity theory (particular community identity) and the spiral of silence theory. We argue that community identity poses as a positive characteristic of an individual. This characteristic however does not postulate under which mechanism behavior may occur. In order to make sense of the ramification of the community identity and the mechanism that it may triggers, we contextualize the fear of isolation from the spiral of silence theory.

Figure 1. Research model with community identity and fear of isolation influencing click-speech.

Usually, users form community identity due to a common interest and homophily with their members.^35–39

As a result, there is a solid attachment to the community, and the community becomes a concept that members strongly identify with. This can lead to situations where one’s opinions are either changed or not seen fitting into the group.^24,30 Members feel the urge to fit into the group as a defining characteristic of their identity. This leads to an affirmation of a prevailing community opinion. To demonstrate how they fit in, 7click language is one way to show that they like and share the content of others. It is not so important what the content is. However, the mere act of liking and sharing the content of peer community members is sufficient to fit into the group and satisfy the community identity. Because the importance of the content is diminished, community members also use click language on content that only sometimes confirms their opinions. Furthermore, the number of shares and likes indicates community cohesion, incentivizing click-speech. Research indicates that engaging with content fosters community cohesion.⁴⁰ Furthermore, the notion to share content against the own opinion can also be found in the research on non-sharing content. Social media users consider the ramifications of sharing content, as for example damaging the own reputation. It may be expected to express certain opinions, as staying silent may also be a form of opinion expression (e.g., in terms of great social injustices, other social media users may demand providing a standpoint to signal solidarity). Consequently research points to users to not share such content.⁴¹ Therefore, we postulate that:

H1:

Community members with high community identity will use click-speech against their opinion.

Individuals are less likely to express their opinion when they feel it opposes public opinion.¹³ They remain silent, and the opinion is not challenged and remains a public opinion. Because social media involves communities in which opinions are shared, we argue that these very mechanisms apply. For individuals to experience fear of isolation on social media, they must be part of an online community. Community identity can explain this affiliation, which considers social ties and the community as part of the self. Due to community identity being a vital part of their self,^10,42 losing such a part of the sole anticipation of losing such a part may induce fear. Individuals without a community identity feel that they no longer belong to a community and therefore remain alone on social media without social connections. In other words, they become isolated. In particular, this isolation can manifest in online community members refusing further interaction or banning members from the community. Therefore, we postulate:

H2:

Community members with high community identity will experience high fear of isolation.

Nowadays, social media platforms afford different forms of opinion expression, such as commenting, liking, or sharing content.^43–45 As a result, many-to-many communication has emerged, with users forming diverse interest groups ranging from politics to entertainment. As like-minded people engage with each other, they strengthen their bonds and solidify their opinions.³⁹ Likewise, they fear they will not be able to participate in their respective communities because they will be isolated if they do not conform to their community (H2). As a result, community members naturally seek ways to reduce the fear of isolation. Since click-speech is an effortless behavior visible to other community members, it provides an easy mechanism to reduce the fear of isolation. Even though the posted content may not be in line with their opinion, we argue that the benefits of click-speech (i.e., showing presence and strengthening social ties) outweigh the potential cognitive dissonance⁴⁶ (i.e., the negative feeling that occurs when cognition and behavior do not align) the liking or sharing of a post would induce. Therefore, we postulate that fear of isolation leads to self-censorship and increases easy-to-use mechanisms like click-speech. Thus:

H3:

Community members who fear isolation use click-speech against their own opinion.

Research method

Scenario-based study

To test our hypotheses, we conducted a scenario-based study in early 2021 consisting of three topics: immigration (ongoing topic), the 2020 presidential election (the most recent topic at the time of the study), and COVID-19 (the current topic at the time of the study). The wording was inspired by previous research, as scenario-based studies provided reliable data in the context of the spiral of silence theory^47,48 and embedded the three topics in social media. The topics chosen were derived largely from mainstream news at the time. This way, we could ensure that participants had an (1) understanding of the topics and (2) induced a more realistic outlook, since the topics were discussed on social media very thoroughly. We asked participants in our study to consider themselves part of a social media community. They disclosed the platform (e.g., Facebook, Reddit, Instagram) and the topics of interest to their community (e.g., politics, movies, books, sports, music). We tailored the scenario to focus on whether the content was broadly accepted within the community. The scenario should have focused on who posted the content to reduce the potential influence characteristics of the source could have on the reader. The scenarios ended with whether the participants would share their opinion. Each participant was able to view all three scenarios. The detailed scenario descriptions can be found in the appendix.

Study procedure

At the beginning of the study, we asked participants which community they primarily belonged to, and we kept that particular community in mind throughout the rest of the study as seen in Figure 2. First, we assessed participants’ community identity relative to their pre-identified community before introducing them to the different scenarios. In the final part of the study, participants provided demographic information. Throughout the study, participants were required to complete attention checks (e.g., reporting the scenario topic or marking a specific response). If the participants did not pass the different attention checks, their responses were omitted, as we could not make sure they actually read the study instructions.

Figure 2. Study procedure with three scenarios.

After each scenario, we asked participants whether they would share or like the outlined post. We also asked about the fear of isolation. The survey was hosted on surveymonkey.com. The sample consisted of US American citizens on Amazon Mechanical Turk (mTurk). To ensure quality but not focus on professional mTurkers, we set the minimum past approval rating of tasks to 97% while already needing 500 tasks approved. We paid each participant over minimum wage. Table 1 lays out sample characteristics.

Table 1. Sample characteristics with 164 used replies.

Full Sample	Finished Studies	Attention check passed	Male	Female	Non-Binary
296	250	164	50.00%	49.39%	0.61%

The participants were between 22 and 71 years old, with 37 participants between 22–30 years, 63 between 31–40 years, 40 between 41–60 years, and 24 between 61–71 years. All participants, except 2, reported to have a high school graduate, diploma or the equivalent concerning their educational background.

Measurement and operationalization

Click-speech: Click-speech was measured in two dimensions: sharing and liking. We adapted the items from Kim and Dennis,⁴⁹ asking “The likelihood you would share the post” and “The likelihood you would like the post .“Items were measured on a seven-point Likert scale ranging from “very unlikely” to “very likely.”

Community identity: We measure community identity with the collective self-esteem scale: identity from Luhtanen and Crocker.⁴² An example item is “The online community I belong to is an important reflection of who I am.” Items were measured on a seven-point Likert scale ranging from “strongly disagree” to “strongly agree.”

Fear of isolation: We developed five items based on past research utilizing fear of isolation.^33,47,50 The items were topic focused and suggested that individuals may be isolated from the community if they express their opinion that is not in line with the community opinion. Items were measured on a seven-point Likert scale ranging from “strongly disagree” to “strongly agree.” Four scholars were involved in a card-sorting procedure to test the validity of our newly developed five items. The goal of card sorting was to distinguish the construct of the fear of isolation from similar concepts. The procedure incorporated the five developed fear of isolation items, three items for the construct anxiety,⁵¹ and five items for the construct fear of negative evaluation.⁵² In the procedure, we altered all items to fit the social media context so that the participants could not distinguish the items solely based on the presented context. The results show that the items for fear of isolation were mapped correctly (85% correct classification). Every item was mapped one time falsely. Therefore, according to past research, card-sorting suffices interrater reliability.^53,54 In conclusion, the items for fear of isolation differ from similar concepts and describe the construct.

Results

We evaluated the data using structural equation modeling with the partial least squares method using SmartPLS 3.3.9. We align with several rules to elaborate on whether the sample size is sufficient. First, our sample size of n = 164 suffices the rule of ten, which would require 20 finished surveys.¹⁶ Second, we evaluated the sample size using the minimum explanatory power of the model’s R². Past spiral of silence studies reported R² ranging from 10% to 35%.^33,47 We used a conservative approach and took an R² value of 10% for this calculation. Paired with the maximum arrows pointing at a construct (two in this case on click-speech), the minimum R² method would require a sample size of 110 completed surveys.¹⁶ Since the sample size exceeds the presented values, we present various quality criteria: internal consistency reliability, convergent validity, and discriminant validity.

We used Cronbach’s alpha and composite reliability (CR) as values for the internal consistency reliability (see Table 2). Values above 0.70 are desired for both and exceeded for our sample, as seen in the table below.

Table 2. Internal consistency, AVE, and Fornell-Larcker Criterion suffice all quality criteria.

	Cronbach’s Alpha	Composite Reliability	Average Variance Extracted (AVE)	Fornell-Larcker-Criterion
				Click-Speech	CI	FoI
Scenario 1
Click-Speech	0.901	0.953	0.910	0.954
CI	0.872	0.906	0.709	0.313	0.842
FoI	0.960	0.969	0.863	0.216	0.251	0.929
Scenario 2
Click-Speech	0.945	0.973	0.948	0.974
CI	0.872	0.904	0.706	0.267	0.840
FoI	0.972	0.978	0.899	0.277	0.263	0.948
Scenario 3
Click-Speech	0.894	0.950	0.904	0.951
CI	0.872	0.907	0.712	0.294	0.844
FoI	0.972	0.978	0.898	0.318	0.195	0.948

CI = Community Identity; FoI = Fear of Isolation.

We test the average variance extracted (AVE) and factor loadings to assess convergent validity. Values above 0.5 are desired for the AVE, which was exceeded in this sample.¹⁶ Furthermore, all factor loadings exceed the threshold of 0.708¹⁶ and are significant at the 0.001 level. Therefore, we can conclude convergent validity based on these two criteria.

Lastly, we apply three criteria to determine discriminant validity: cross-loadings, Fornell-Larcker-Criterion, and heterotrait-monotrait ratio (HTMT). For the cross-loadings,⁵⁵ we found that the associated items correlate most with their constructs. The Fornell-Larcker-Criterion also suggests discriminant validity since the squared AVE for their respective construct is highest. The HTMT values above 0.90 do not show discriminant validity.¹⁶ In Table 3, we present the HTMT values for each scenario. All values are below the threshold of 0.90.

Table 3. HTMT values < 0.9 suggest discriminant validity.

	Click-Speech	CI	FoI
Scenario 1
Click-Speech
CI	0.317
FoI	0.232	0.249
Scenario 2
Click-Speech
CI	0.252
FoI	0.289	0.263
Scenario 3
Click-Speech
CI	0.303
FoI	0.341	0.193

CI = Community Identity; FoI = Fear of Isolation.

Since the applied quality criteria were sufficient, we tested the hypotheses using structural equation modeling (partial least squares, see Table 4). For H1, we find that on every occasion, community identity influenced click-speech (Scenario 1: β = 0.276, p < .001, Scenario 2: β = 0.208, p < .01, Scenario 3: β = 0.242, p < .01). Therefore, we find support for H1. For H2, we find that community identity significantly influences fear of isolation throughout the scenarios (Scenario 1: β = 0.251, p < .01, Scenario 2: β = 0.263, p = .01, Scenario 3: β = 0.195, p < .05). Lastly, for H3, we find significant influences of fear of isolation on click-speech for scenarios 2 and 3. In scenario 1, we see fear of isolation was insignificant, while on the other occasions, it remains significant (Scenario 1: β = 0.147, n.s., Scenario 2: β = 0.222, p < .01, Scenario 3: β = 0.271, p < .01). Therefore, we find partial support for H3, which will be elaborated on in the discussion section. Concerning the explanatory power, the R² values can be found in the table below.

Table 4. Path-coefficients and explanatory power for click-speech.

	Scenario 1	Scenario 2	Scenario 3
Influence on Click-Speech	Click-Speech	Click-Speech	Click-Speech
Community Identity (H1)	0.276***	0.208**	0.242**
Fear of Isolation (H3)	0.147n.s.	0.222**	0.271**
Influence on Fear of Isolation	Fear of Isolation	Fear of Isolation	Fear of Isolation
Community identity (H2)	0.251**	0.263**	0.195*
Dependent Variable R^2	11.8%	11.7%	15.8%
Fear of Isolation (state) R^2	6.3%	6.9%	3.8%

Significance key: ***p < .001; **p < .01; *p < 05.

Since fear of isolation did not influence click-speech in scenario one, and we gathered data about liking and sharing, we investigated the influences in more detail in a post-hoc analysis (see Table 5). Here, we find that for scenario 1, fear of isolation significantly influences sharing (Scenario 1: β = 0.163, p < .05) but not liking (Scenario 1: β = 0.147, n.s.). Likewise, the effect size toward liking indicates no effect. We also compared effect sizes (f²) of the remaining influences and found that all range within small effect sizes (0.02 < f² < 0.15).⁵⁶

Table 5. Path-coefficients and explanatory power for liking and sharing.

	Scenario 1		Scenario 2		Scenario 3
Influence on Click-Speech	Liking	Sharing	Liking	Sharing	Liking	Sharing
Community Identity	0.288***	0.239**	0.211**	0.200**	0.267***	0.190**
Fear of Isolation	0.118n.s.	0.163*	0.214*	0.216*	0.246**	0.271**
Effect Size f^2
Community identity	0.088	0.060	0.047	0.042	0.081	0.040
Fear of Isolation	0.015	0.028	0.048	0.049	0.069	0.081
Dependent Variable R^2	11.4%	10.4%	11.5%	10.9%	15.7%	13.0%

Significance key: ***p < .001; **p < .01; *p < .05.

Discussion

This research aims to answer why members of social media communities like or share (in short: use click-speech) even when others post messages that are against their own opinion. To explain this, we use the concept of community identity with the spiral of silence theory and collect data from 164 individuals in three different scenarios. Our results show that individuals engage in click-speech even when disagreeing with the posted opinions due to their strong identification with their respective communities.

Our results suggest that community identity influences click-speech in each scenario, indicating a unique context (support for H1). Because the scenarios were set up so that participants always held an incongruent opinion toward the community, we see how strong community identity drives click-speech. These findings are consistent with existing knowledge, suggesting that liking and sharing, mainly, are the primary means of maintaining personal connectedness. We show that social media provides an environment where users constantly share their opinions in various ways. In our post hoc analysis, we find no differences in H1 between liking and sharing as click-speech, as all effect sizes are small. We show that social media mechanisms such as liking and sharing incentivize opinion disclosure and create the illusion of community cohesion, which may strengthen community identity.

Furthermore, our results indicate that community identity influences fear of isolation in all scenarios (support for H2). This finding is consistent with our hypothesis because the more substantial the ties to a social group, the greater the perceived loss when isolated. As identity theory suggests,²³ community identity is as much a part of an individual as individual identity. Individuals may fear separation from the social media community that they consider an integral part of their identity.

Finally, we find partial support for fear of isolation influencing click-speech (partial support for H3). In one circumstance (scenario one), we did not find a significant influence of fear of isolation on click-speech. The reason for this may lie in the characteristics of the scenario topics. Scenario one provided an ongoing issue (immigration) in the United States that has been part of the political discussion for many years. In contrast, scenarios two and three were very present and widely visible at the time in various news media. Therefore, we assume that community members are likelier to use click-speech against their opinions on hot topics. In the post hoc analysis, only sharing is significant in this case. It seems that sharing was sufficient for the topic of immigration, which was not that present at the time of the study. One explanation could be that community members associate liking with some form of endorsement while sharing simply passes on information about a topic. Therefore, simple sharing may be sufficient to deal with the fear of isolation, in contrast to the hot topics of the presidential election and COVID-19.

We argue that liking and sharing are simple strategies for reducing fear of isolation in most cases. We can explain this by the relatively low cognitive effort required to engage in such behavior while reaping the benefits of reduced fear. This finding has broad implications for evaluating the number of likes and shares on social media. According to the results, liking and sharing are primarily driven by community identity rather than personal beliefs in social media communities. Therefore, we show that the phenomenon of confirmation bias^57,58 (i.e., liking or sharing posts that reaffirm currently held beliefs) is diminished in this specific context.

Theoretical contributions

These insights let us contribute to theory and literature in the following aspects. Overall, we contribute to the research streams on click-speech and identity.

Related to click-speech, we contribute in the following ways. An existing discussion investigated click-speech on social media.^17,31 So far, research has shown that click-speech is used to form social ties on social media.^19,20 Despite the importance of those findings, research has remained quiet about the influence of affiliation to a specific community, which has been core since the emergence of social media communities, especially spaces of like-minded individuals.⁵⁹ Previous research on click-speech has observed behavior at the platform level. Social media communities provide a distinct space for individuals with more in common than the public space of social media. This opens up the possibility of developing opinions that are very different from what would be considered public opinion. Because such communities can establish strong ties to take real-world action, focusing on social media communities is particularly important regarding radicalization and polarization. We add to the existing literature by overcoming this shortcoming and suggest that the benefits of click-speech at the community level for strengthening or building social ties are enhanced.

Furthermore, focusing on click-speech against our opinion, we expand on research that vastly investigated how individuals interact with content that reaffirms their opinion.^49,60 With this novel view, we provide more context to click-speech since social interaction within social media communities can overcome mechanisms such as confirmation bias. Our insights provide strong arguments for why the environment in which information is embedded is crucial in understanding how individuals interact with it. Specifically, for recent and hot topics in social media communities, distinct opinions are disregarded, and posts will be liked and shared to increase affiliation with the community. In this regard, we show that the spiral of silence mechanism of fear of isolation reaches beyond self-censorship, as originally deposited,¹³ and influences click-speech against the own opinion.

Related to community identity, we contribute as follows. We discuss community identity as an antecedent of fear of isolation. Therefore, we contextualize⁶¹ the spiral of silence theory within social identity theory. The mechanism of the spiral of silence theory, fear of isolation, has been operationalized as an emotional state that arises in situations where individuals consider disclosing an opinion that is incongruent with the community. We provide deeper insights into the roots of this fear in the context of social media communities: the loss of community identity. Community identity is constantly being reinforced as social media provides various opportunities for individuals to engage within communities. The potential loss of this identity drives community members to behave in ways that are contrary to their beliefs. This contextualization helps us understand that community identity has two sides for an individual. On the one hand, community identity is an integral part of the self. This can be seen as the positive side of interacting with the community. Members behave in ways that fit in because they want to fit in. Their community identity is a positive side that they want to develop. On the other hand, with the possession of community identity comes the potential for loss. With the fear of isolation, we account for the negative ramifications associated with community identity. We furthermore provide an underlying mechanism that manifests cognitively from the community identity. We show that both sides must be addressed to understand click language behavior in social media communities.

Practical contributions

For practice, we further understand how echo chambers can potentially emerge. Community identity also reinforces dominant opinions. The mechanism of liking and sharing provides an easy-to-process metric of how popular an opinion may be. The research shows that the opinion may not necessarily be popular, but community members want to fit in. Social media users need to understand that liking does not always imply approval and does not necessarily reflect the popularity of an opinion within their community. Knowing this, social media users should be more cautious about what they like or share to counteract the potential push of an opinion that may not even represent the community, just the loudest.

If social media users understand the ramifications of their click-speech behavior, they can foster a more open community where diverse opinions are tolerated. The existence of dominant opinions silences further discourse and can ultimately alienate individuals further from society for simply wanting to be part of the community. In extreme cases, this would be the emergence of echo chambers that could radicalize their members.

Limitations

This research has several limitations. First, the participants were offered a hypothetical situation. Therefore, there could be social desirability bias because they knew they were being tracked.⁶² Second, the proposed sample was gathered in the US. We could assume that individuals feel different toward a collective due to cultural differences. Communication may be affected by concepts such as collectivism and individualism.⁶³ Since a US sample may be arguably more individualistic, we could argue that the proposed influences are even more vital in cultural settings that are more collectivistic. Lastly, all participants saw all scenarios. Due to constant exposure to an environment that argues against their convictions, participants might start to cope and fall in line with the community’s opinion.

Since we conducted a scenario-based study, it calls into question how realistic such study scenarios are and whether the gathered data is actually a mirror of real-world behavior. This criticism is to some extend valid, however, since we adhered to past research that showed validity with scenario-based studies within this specific context,^47,48 we feel confident that also these insights have merit within real-life situations.

Our post-hoc analysis shows that only the influence on liking is insignificant while sharing becomes significant. We did not find differences in liking and sharing for the other scenarios. Future research should investigate differences between liking and sharing to determine under which circumstances individuals would like or share. Furthermore, we only looked at circumstances where individuals use click-speech against their opinion. Future research may investigate under which circumstances this behavior is dismissed. Indications for such behavior could be found in a social calculus,⁶⁴ in which individuals try to balance their behavior to achieve a particular outcome. Since we focused on specific communities, it is also worthwhile to understand if this aspect of click-speech against the own opinion also happens in a more public space, as is the case of Twitter, where everyone can see Tweets. Even though the platform does not provide distinct options to create communities, such as Facebook groups, users still feel affiliated with a specific group and act accordingly.

Conclusion

This research showed that community identity, as an extension of the self, significantly influences the expression of light opinion through click speed. We showed that social media communities provide a social environment where opinions are reinforced, even though community members may disagree with the content. This is due to the strong identification with the community as part of identity and the fear of potential loss of identity. The spiral of silence provides a context for the emotional state in which individuals find themselves when confronted with an incongruent opinion within their community. The fear of isolation has been shown to influence click-speech. This behavior, in turn, promotes community opinion and silences alternative views.

We build on past research, which investigated the influence of fear of isolation on click-speech in an anonymous environment,³¹ and in uncivil social media environments.¹⁷ We provide an additional context, the social media communities, to be considered an application for the spiral of silence theory. We furthermore contextualized the theory within social identity theory to determine the community identity as an antecedent of fear of isolation in that specific context.

Ethics approval

Due to the non-interventional nature of our scenario-based study we did not require ethical approval in line with the guidelines of our ethical committee GEHBa (https://www.gehba.de/home/).

Disclosure statement

No potential conflict of interest was reported by the authors.

Appendix

Scenarios

Immigration

One day, you see a post from a group member addressing issues concerning immigration. The group member’s opinion is not in line with yours; you feel the opposite of what is stated in the post. Since the post was shared yesterday evening, several comments from the community already support the presented opinion. Now, you are thinking about contributing to the discussion.

Presidential election

You see a post from a group member addressing the topic of election fraud in the context of the 2020 presidential election. The group member elaborates on the topic and comes to a completely different conclusion than you have on the matter. The post is already a few hours old, and several comments from the community supporting the group member’s conclusion have been posted. Now, you are considering whether you should engage in the conversation.

COVID-19

In your community, a group member posted information about whether wearing masks help reduce the spread of COVID-19. The group member articulates a stance on the efficacy of masks that you do not share. However, several comments from the whole community have accumulated over time that supports the group member’s stance. You are also thinking about posting your opinion on the matter.