Why so quiet? Exploring inhibition in digital public spaces

Social network sites have been considered as important arenas for public debate, but as a large proportion of users do not actively participate, there is a need to further our understanding of a phenomenon as hidden, unnoticed and invisible as non-participation. We argue that inhibition is a valuable socio-psychological lens to study non-participation, usefully extending the conceptual framework of political communication regarding non-participation, and offering a more precise way of theorising the intent behind this apparent passivity. Using representative survey data collected in Norway in 2017 (N = 2064), we first sensitise the multi-layered concept of inhibition through combining different dominant approaches: the spiral-of-silence theory, the harsh debate climate, political efficacy, and specific properties of social network sites related to identity and impression management. Second, we show that inhibition functions as an in-between concept balancing participation and non-participation in social network sites. Through factor analysis principal component factor (PCF), we integrate established theories that allow us to define overarching dimensions of inhibition, demonstrating that it is a complex phenomenon not easily understood through one specific theoretical perspective.


Introduction
The majority of Internet users do not actively or visibly participate in social network sites (Kushner, 2016;Malinen, 2015;Nonnecke and Preece, 2000;Sun et al., 2014;Van Dijck and Nieborg, 2009), and thus far we know little about these users. In this study, we ask what inhibits people from actively participating or contributing to public debates on social network sites (SNS), in particular when they actually wish to voice their opinion. Hayes et al. (2005) describe inhibition as a balancing act between wanting to express something and worrying about potential risks. In line with this definition, we propose inhibition as a valuable socio-psychological lens to study online public non-participation for two reasons. First, it provides insight into what stops people from voicing their opinion or being active, visible participants in SNS. Second, by studying inhibition, we also indirectly study the intent to participate, because only those who want to take part may have experienced inhibition. This study analyses the multidimensionality of inhibition as a way of advancing our understanding of online public non-participation.
The term public can be said to encompass 'any issues affecting how we live together that require common solutions' (Couldry et al., 2007) and 'connotes ideas of citizenship, communality and things not private' (Papacharissi, 2002). Public participation, similarly to political participation (Verba et al., 1995), can therefore generally be described as activities that engage with such issues. Non-participation could, in line with the definition of participation, be described as not being involved or not participating, which in this study refers to users not commenting on news stories or not voicing their opinion in a debate or otherwise engaging in what may be considered public SNS-activities.
Even though it concerns most users, there are few studies of non-participation in SNS and those that do exist mostly attempt to reformulate why non-participation should be considered a valuable activity (Crawford, 2009;Sun et al., 2014). We propose that studying inhibition can provide valuable insights into the intention behind non-participation in SNS. However, to advance our understanding of non-participation in SNS, we need to look beyond single concepts that exist when it comes to inhibition. We have identified four dominant explanatory frames that highlight different aspects of inhibition: spiral-ofsilence theory, the harsh debate climate, political efficacy and specific properties of SNS related to identity and impression management. In this study, statements representing these different aspects of inhibition are integrated into one survey to allow us to explore the multi quality of what it is that inhibits users from voicing their opinion and taking part in debates on SNS. Norway makes a compelling case for this study. Most Norwegians are users of SNS (Statistics Norway, 2019); it is a society that is characterised by equal rights and freedom of speech (Freedom-House, 2018;Reporters Without Borders, 2019), and yet, Norwegians are not particularly active in online debates (Moe et al., 2017). We ask whether Norwegian users feel inhibited from taking part from such public SNSactivities and why.
This study focuses on the experience of inhibition when wanting to post or actively participate in a public debate on SNS. 1 It offers important insights into reasons why people who are active in the public sphere nevertheless feel inhibited on SNS. Our aim is to advance our understanding of non-participation by studying inhibition by combining four different theoretical approaches that rarely communicate across. We employ exploratory factor analysis to find the overarching dimensions of inhibition in SNS. The intention is not to pit one theoretical angle against another, but rather to study them together as dimensions of inhibition that may be separate or intertwined and overlapping. We consider the exploration of the multidimensionality of inhibition as a prerequisite for bridging different concepts and theoretical frames, in which inhibition plays a crucial role but which are rarely considered in their mutual interaction. Next, we use ordinary least squares (OLS) regression models to explore whether there are certain users who experience more inhibition than others, to what degree inhibition makes users refrain from taking part in public SNS-activity, and whether certain types of inhibition affect active participation more or less. The analysis and variables used are described in detail in the methods section.

Literature review
Active and political participation is generally seen as beneficial for the legitimacy of political decision making. Accordingly, active participation is usually framed as a distinctly beneficial phenomenon (Jenkins, 2006;Putnam, 2000) and, conversely, non-participation as negative -as defined not by what it is but by what it is not: the absence of the more desirable active citizen participation.
SNS represent digital forums similar to places where citizens can take part in everyday political debates (Neuman et al., 2011, p. 11), but there are divergent understandings of the implications these spaces have for participation. Proponents of mobilisation theory claim that the Internet facilitates democratic participation because all users have equal opportunities (Neuman et al., 2011;Rojas and Puig-i-Abril, 2009;Storsul, 2014). In contrast, reinforcement theory proponents claim that the Internet might deepen the traditional divides, as users with high socio-economic status will benefit from higher quality resources available also online (Norris, 2001). Again others argue that online political participation would mostly reinforce already established forms of engagement, leaving the state of affairs principally unaltered (Calenda and Meijer, 2009;Norris, 2001). Such theories describe the connection between online and offline participation, yet the common theme in this debate is, again, the normative assumption that active political participation 2 is intrinsically good (Lutz and Hoffmann, 2017) and that non-participation is less good. Consequently, participation tends to be discussed, and non-participation tends to be ignored.
While online participation has predominantly been understood by researchers as content creation or actions that can be observed and counted, non-participation has remained under-theorised so far despite the established fact that most users do not contribute content themselves (Kushner, 2016;Malinen, 2015;Nonnecke and Preece, 2000;Sun et al., 2014;Van Dijck and Nieborg, 2009). Even though non-participation might involve considerable cognitive and emotional effort (Ewing, 2008), scholars tend to overlook this majority.
Non-participation can be understood as activity as opposed to a passive non-behaviour, and some researchers argue that actively logging in and paying attention online also contributes to the community by providing a gathered audience for others; in other words, listening can be seen as a form of participation (Crawford, 2009;Sun et al., 2014).
Users can even choose to refrain from activity to further a cause considered socially undesirable, which may follow from collective or individual political choices (Casemajor et al., 2015). From these perspectives, deliberately choosing non-participation will also qualify as active, and therefore these scholars suggest to looking for intent rather than observable results. Conversely, Morozov (2011) puts the self-evident nature of participation in social media into perspective by presenting it as 'slacktivism': merely a form of self-staging, which fails to translate into offline participation or political change. How intention matters in questions of participation or abstention is therefore suggested as an important research angle.
What we find is research dominated by a normative understanding of political participation as something countable, active and beneficial, while we know little of the intent behind non-participation online. Along these lines, we argue that the socio-psychological lens of inhibition advances our understanding of non-participation as a way of theorising such intent.

Existing theories relating to inhibition
Inhibition plays a part in different theoretical perspectives, yet even if these separate theoretical frames are closely related and might overlap, they tend to be applied in isolation. The main theoretical perspectives found in the research literature point to different aspects of inhibition. First, the so-called spiral-of-silence (SOS) theory frames inhibition as self-censorship. This theory argues that in a polarised opinion climate people are more inclined to keep their opinions to themselves if they seem not to have the support of the majority (Hayes et al., 2006(Hayes et al., , 2013Kwon et al., 2015;Noelle-Neumann, 1974). Research on students using Facebook concluded that the desire for social approval does influence the extent of opinion exchange on SNS (Kwon et al., 2015). SOS theory was operationalised by asking about worry about others disagreeing, being potentially wrong or being misunderstood. In addition, a statement about starting to write but choosing to self-censor is included.
Second, inhibition can also be related to incivility (Papacharissi, 2004) and social contagion in online debates (Hermida, 2014: 41-42), which may cause people to worry about being harassed or attacked. The online disinhibition effect (Suler, 2005) assumes that being online somehow detaches people from responsibilities and inhibitions they have in 'real life', thus allowing them to say whatever they want without facing the consequences. Another explanation for the more aggressive climate on SNS is offered by the SIDE theory (Social Identity model of Deindividuation Effects). This theory describes how visual anonymity and lack of cues (Joinson, 1998;Walther, 2011) lead to deindividuation, which, in turn, prompts an in-group or out-group dynamic and in that way allows for negative attitudes towards others (Bennett, 2012). Such theories mostly focus on incivility in an aggressive online debate climate and do not relate directly to inhibition. We argue, however, that the tone of the debate is relevant, not only because people may want to shield themselves from potential hostile responses (Stroud et al., 2016), but also because people who perceive the online debate to be of a low standard may be less inclined to take part themselves (Springer et al., 2015). Inhibition due to fear or avoidance of incivility was operationalised by asking about the participants' feelings regarding criticism or harassment.
Third, inhibition may also be related to properties that are unique to computer-mediated settings when compared with unmediated ones, such as an unknown audience, potential context collapse, and blurring the lines between private and public (boyd and Ellison, 2007). Moreover, SNS-specific affordances such as persistence, searchability, replicability and scalability (boyd and Ellison, 2007) may induce the feeling of being less in control of what is posted. People feel inhibited, in this case, not because of their perception of the opinion climate, but rather because of their understanding of the SNS platform itself. Still, the judgemental and sometimes even derogatory tone in online debates can form a backdrop for the readers' feeling of having less control and ability to shield oneself from potential attacks (Bazarova and Choi, 2014;boyd, 2008;Litt, 2012;Marwick and boyd, 2011;Meyrowitz, 1985;Storsul, 2014). Inhibition pertaining to SNS-settings was operationalised by asking about worry about the amount of response, being misused or taken out of context, finding it difficult to express disagreement on SNS or preferring face-to-face settings.
Finally, we included in our analysis the perception of SNS as arenas for public debate and the feeling that one ought to, or can, voice one's opinion. This feeling is associated with the idea of internal political efficacy (Campbell et al., 1954), as well as with efficacy linked to posting on SNS. Internal political efficacy refers to the users' perception of their ability and competence to understand and participate effectively in politics (Balch, 1974). A low sense of internal political efficacy can then translate into inhibition based on lack of self-confidence that erodes one's motivation to express oneself. Social media political efficacy bears a resemblance to external political efficacy (Campbell et al., 1954), but relates to the user's evaluation of SNS specifically as a place for public or political participation (Velasquez and LaRose, 2015). Inhibition, along these lines, would derive from considering SNS not to be the correct venues for this type of activity. Inhibition relating to political efficacy is operationalised by asking if it feels natural to take part, or if one feels obliged to respond, if one finds self-expression difficult, if one feels one has nothing to say, or that it would not make a difference.
We expect that different dimensions of inhibition related to these theoretical frames are intertwined, and we, therefore, discuss all of these dimensions in this study.

Data and method
Our data originate from an online panel collected in fall/winter of 2017 by the MeCInproject, 3 which resulted in a dataset composed of 2064 Norwegian participants, 18 years and older. The sample intended to be statistically representative of the overall Norwegian population based on age, gender and level of education; however, it ended up a little skewed towards the older age groups, and higher education levels (Statistisk Sentralbyrå, 2017) and analyses are weighted to rectify this. Only weekly or more frequent users of SNS are included in the analyses, which leaves us with 83% of the total sample (N = 1720). We chose to include the questions about inhibition in the MeCIn-survey because something as untraceable as non-participation or inhibition will only be uncovered through interrogation, and not through non-reactive content analysis (Ruiz et al., 2011).
The analyses consist of exploratory factor analysis (PCF) to find the overarching dimensions of inhibition in SNS, followed by two hierarchical (OLS) regression models that investigate the inhibition factors as predictors of inhibition, and use demographic variables and inhibition as predictors of public SNS-activity.
The exploratory factor analysis contains 15 measures of inhibition. We started the project investigating the existing theories that relate to inhibition and then operationalised these into 15 statements to cover the whole range within one survey (described in the text earlier and in Table 2). The intention was to bridge these differently framed and constructed concepts of inhibition and combine them into one multi-layered concept of inhibition as a prerequisite for using inhibition to study the intentions that underlie nonparticipation in SNS. Even if the statements are based on different theoretical approaches, they overlap at times and are thereby not limited to a single theoretical perspective. For example, it is debatable whether worry about criticism (Q61f) relates to SOS theory (i.e. criticism of opinion) or the harsh debate climate (i.e. criticism of person) -perhaps both. In the same way, self-censoring, as described in SOS theory, is likely to be enhanced by the asynchrony in SNS debates. In fact, as described earlier, we expected these theoretical frameworks to be intertwined.
In the regression models (OLS), the control variables, demographic features, and public SNS-activity (alpha.89) are used in connection with the inhibition measures. The demographic features -age-group, education and income levels, and gender -are relevant to include, as they are known to be related to efficacy (Beaumont, 2011; Table 1), general public participation (Morrell, 2003) and SNS participation (Mossberger et al., 2007). Public SNS-activity is an index (0-5) (See Table 1 for descriptive statistics), whereby each participant has a number corresponding to the activities he or she reported. The five types of SNS-activities included in this index are (a) write posts about society or politics, (b) start debate/discussion themes, (c) participate in a debate, (d) post links to news about society or politics, (e) comment on news about society or politics. Public SNS-activity is crucial to include because it represents the visible action in SNS that inhibition, in theory, would reduce, allowing us to explore to what degree inhibition makes users refrain from participating and whether or not there are certain types of inhibition that affect such participation more or less than others. Two regression models are included. The first shows if there are particular groups of users more prone to experience inhibition, and the second shows to what degree inhibition and demographic differences impact public SNS-activity.

Results
Almost 90% of the participants reported that they experienced one or more types of inhibition related to participating actively in SNS, leaving only 11% reporting that they do not feel inhibited from actively participating. Furthermore, 84% say they never voice their opinion, share or comment on posts related to society, news or debates on SNS (public SNS-activity). Although the questions about inhibition on SNS do not concern one specific SNS platform, it is reasonable to think that the participants mostly answered in terms of Facebook, as this is the platform predominantly used (93% use Facebook). The first factor (Factor 1, Table 2) has to do with worry connected mostly to others' behaviour or reactions, but also with worry about one's own ability to voice an opinion. This factor seems to unite the different ideas of SNS as 'hard and dangerous' places that are frightening or difficult to take part in. This factor is found to be a dimension across variables from all the included theoretical angles. It encompasses worries about others' reactions, such as being criticised or that others might disagree with what one posts, but also not being able to control the outcome when posting, such as being misunderstood, being wrong or not getting enough response from others. Furthermore, it includes feeling unable to express oneself well or having nothing to contribute. We refer to this first factor as social discomfort.
The second factor (Factor 2, Table 2) captures what we summarise as scepticism of SNS as appropriate spaces for debates. It consists of three statements: worries about having what one says misused or taken out of context, worry about being harassed or bothered, and the preference for discussing face-to-face rather than on SNS. Since these statements are within the context of SNS, and harassment or misuse appears as issues often associated with online debate (Hermida, 2014;Papacharissi, 2002), this can be understood as an evaluation of the platforms. We interpret this as a factor that indicates a dismissal of public SNS participation. We label this second factor dismissal of SNS as a venue for debate.
The third factor (Factor 3, Table 2) seems to capture a feeling of not being obliged to participate in SNS debates. It includes survey statements concerning the opinion that it is not natural or that one ought not to take part, and preferring to express disagreement face-to-face instead of via SNS. Although they do not feel compelled to take part, they use last-minute self-censorship, which involves initially starting to write a response or post, but then choosing not to post it. We refer to the third factor as preferring presence over participation.
The fourth factor (Factor 4, Table 2) consists only of one variable, which is not believing that sharing/posting will make a difference. This factor was retained as Horn's parallel analysis (PA) supported four factors. However, it consists of one variable, which makes it difficult to interpret and is thus omitted from the following analysis and discussion.
The first three factors discovered in this analysis do not fit squarely into one theoretical frame or the other. This does not mean, however, that there is no evidence to corroborate these theoretical approaches; rather, they do appear as intertwined and overlapping dimensions, whereby it is not easy to separate one theoretical approach from the other. Next, we turn to the question of whether there are particular groups of users more prone to experience inhibition (Table 3). Social discomfort (Model 1) and Dismissal of SNS as a venue for debate (Model 2) are found to not have a significant relationship with any of the demographic characteristics included. One might expect Social discomfort to be gendered (Coffé and Bolzendahl, 2010), but in our analysis, the relationship with gender is not significant. Preferring presence over participation is found to have a statistically significant relationship with being older (0.057) and having a lower income level (−0.119) (Model 3). Preferring presence over participation is the only factor that has a statistically significant relationship with any of the demographic characteristics, but where essentially very little of the variation (2%) can be explained by demographic differences.
To establish to what degree inhibition and demographic features impact public SNSactivity, we performed another regression analysis (Table 4). Being male is the only demographic feature with a significant positive relationship with public SNS-activity in all the models (0.074-78), which supports earlier research that has found that men use SNS to express their opinions more often than women (Rollero et al., 2019). Surprisingly, Social discomfort, the factor that explained most variance (Table 2), is found not to significantly impact public SNS-activity (Table 4, Model 2). That means that this may be a type of inhibition that is most experienced, but also that this is the type of inhibition that is least likely to stop people from taking part. The other two dimensions of inhibition -Dismissal of SNS as a venue for public debate and Preferring presence over participation (Model 3 & 4) -both have a significant negative relationship with public SNS-activity (−0.077 & −0.068), which means that experiencing this type of inhibition is associated with less public SNS-activity.

Discussion, conclusion and limitations
This study sought to advance our understanding of non-participation in SNS by studying inhibition by combining different theoretical frameworks that tend to be applied separately. Our analysis demonstrates the complexity of inhibition, with overarching dimensions that appear intertwined across these different theoretical frameworks. What is more, we found that most users do experience some type of inhibition vis-à-vis SNS and that this seems to be unrelated to demographic characteristics. Besides, we found that not all dimensions of inhibition seem to impact public SNS-activity equally.

Inhibition shows intent
Our results provide an argument against the claim that the large number of users who do not actively take part in commenting, creating or sharing content on SNS are simply not interested in doing so. Most users report to experience inhibition (almost 90%), which means that they paid attention to what is going on in the domain of SNS, and, arguably, that they have also felt an urge to express something. Building on this notion, studying inhibition, by proxy, becomes a study of paying attention and listening in (Crawford, 2009;Dreher, 2009). We argue that the feeling of inhibition is a meaningful distinction between those in the 'silent masses' on SNS who care, want to take part and are listening in, and those who do not care. As such, it can be understood as an in-between concept balancing participation and non-participation in SNS that provides us with an indication of intent. Particularly, the factor Preferring presence over participation (Table 2, Factor 3) is relevant for discussing inhibition as an indication of intent. This dimension is about making a conscious choice to not have a voice or be visible, but about being present without feeling that it is natural to become more actively involved. To feel inhibited, one must pay attention. Preferring presence over participation is arguably about listening in and paying attention, and, at the same time, feeling inhibited about using SNS as places for debate. Moreover, as demonstrated in Table 4, this type of inhibition is found to have a significant negative impact on public SNS-activity, meaning that those who experience this type of inhibition are less likely to participate actively, also compared with those who experience the type of inhibition we labelled Social discomfort.

Worry as a central aspect of inhibition
The first factor we identified is Social discomfort ( Table 2, Factor 1), a dimension uniting the different ideas of SNS as 'hard and dangerous' places where it is frightening or difficult to voice one's opinion, thus combining different types of worry. We, therefore consider the concept of worry to be central to the dimension of Social discomfort. In terms of psychology, worry can be defined as negative thoughts or emotions that come from a proactive cognitive risk analysis, done to avoid or solve anticipated possible threats and their potential consequences (Schacter et al., 2011). As such, worry is understood as a natural response to anticipated future problems. We argue that since worry is conscious and involves mental labour (Freeston et al., 1994), it should be understood as an activity. In this study, worry is a dimension of inhibition which appears to be an amalgamation of variables that initially stem from all four included theoretical angles: efficacy, harsh tone of the debate, the affordances of SNS and SOS theory. Worry can therefore be understood as an overarching dimension of SNS inhibition. Furthermore, Social discomfort does not correlate with views on politics, efficacy and political participation outside of the context of SNS, suggesting that this is a psychological dimension connected to the concept of worry rather than to political dispositions. Still, Social discomfort is found to be the type of inhibition that is less likely to stop people from taking part in SNS, meaning that both those who are active in public SNS-activity as well as those who are not.
Social discomfort overlaps with SOS theory, a theory suggesting that worry about others disagreeing stops people from voicing their opinion (Noelle-Neumann, 1974). Worry about disagreement is, in this case, found to be closely related to other types of worry, and Social discomfort is the most influential factor in the analysis ( Table 2). The underlying concept of worry about disagreement, according to SOS theory, is fear of social isolation (Noelle-Neumann, 1993). Critics of SOS theory have suggested that fear of isolation may be one of many factors (Salmon and Kline, 1983), or not a factor at all (Salmon and Neuwirth, 1990). Even though this study does not provide insight into what, exactly, underlies worry about disagreement, worry about others disagreeing is part of the same dimension as worry about a range of different things, making the feeling of worry itself central, rather than the specific reason for worrying. This finding implies that worry about disagreement, rather than appearing in isolation in the context of SNS, is part of a more layered phenomenon of worry.

Inhibition seems intertwined with identity management on SNS
The experience of inhibition in SNS seems intertwined with self-presentation (Brown, 1998;Goffman, 1978). Dimensions found in this study, such as Social discomfort and Dismissal of SNS as a venue for debate, can be said to indicate different aspects of inhibition that also concern identity management in SNS.
Dismissal of SNS as a venue for debate (Table 2, Factor 2) consists of statements that, when combined, summarises scepticism of SNS as appropriate venues for debate. This factor could be understood as not just scepticism about individual debates, but a more general dismissal of SNS as appropriate public spaces in which to stage debates. Moreover, this suggests a particular view of the people who do participate; they harass or bother others, or they misuse what people say or take it out of the intended context. Non-participation can thus function as a sign of distinction (Bourdieu, 1984). These 'non-participators' have an idea of how a debate should be conducted and where the proper debates are carried out (f2f), and also believe that people debating in SNS do not possess this knowledge. Such inhibition can, therefore, be understood as giving rise to an elitist practice. This type of inhibition whereby people consciously choose to not take part in SNS out of what they consider to be good taste can be interpreted as a part of a personal identity project (Giddens, 2008), which is in line with earlier findings that suggest that online political participation is strongly influenced by social desirability (Vitak et al., 2011).
However, this personal narrative that can be created through public SNS-activity must be affirmed by others (Goffman, 1978), since performing identity is never a solo-project but a collaboration with those around us. Consequently, a public display of not articulating oneself well or being misunderstood, wrong (and potentially corrected) or criticised in the domain of SNS might feel like a threat to one's personal narrative. The dimension Social discomfort, therefore, can be seen to share a similar connection with identity management. If we incorporate this view into the setting of SNS, we find that the affordances specific to SNS (boyd, 2010) will make whatever gets posted more visible as each post is permanent, searchable and potentially distributed to more people than intended. Moreover, this visibility not only applies to the post itself but also to other people's reactions to it (Hermida, 2014). Not surprisingly, then, it was found earlier that users of SNS actively use strategies for suppression (Strano and Wattai Queen, 2012) to maintain their identity. Alternatively, as in this case, they feel inhibited from posting anything in the first place. This type of inhibition may be understood as anticipating others' reactions and, not being willing to risk exposing oneself to them, as clashing with one's narrative of self.
To choose to abstain from taking part in debates on SNS may be meaningful for people's self-presentation strategy, yet for other users of SNS, this appears as non-participation. It is not surprising that self-presentation becomes a topic when discussing the use of SNS, as many previous studies have emphasised the importance of identity management here (see for example Bargh et al., 2002;DeAndrea and Walther, 2011;Zhao et al., 2008). However, we also find that the private/public distinction seems blurred; we ask about inhibition related to news, society or politics, but self-presentation appears central. Such findings are supported by earlier research on SNS (Burkell et al., 2014).

Limitations
While using an online panel survey enables us to get a broad understanding of inhibition as a phenomenon, it also has its limitations. First, there is the potential problem of the participants providing an ex-post rationalisation regarding why they do not engage in online discussions in order to reduce cognitive dissonance regarding why they are not engaging even though society might expect them to do so. If this is the case, the answers may lean towards folk theories (Palmer et al., 2020) or stereotypes of what they think is going on in SNS. Second, the items that are included in the survey to measure inhibition do not function as an SNS-inhibition scale 4 and creating such a scale would require further data and testing. This is particularly evident from Factor 4, which only consists of one variable and therefore is hard to fully explain. Moreover, not all initial theoretical frames are equally represented or represented in such a way we can be sure they are mutually exclusive. Still, our study demonstrates the importance of finding ways of studying intent behind apparent silence in SNS, while also providing a fruitful starting point for developing an SNS-inhibition scale. Third, this study falls short of exploring the contextual side concerning inhibition on SNS, in particular with respect to Facebook, where the perceived audience and the perceived nature of the communication might be contextually different, and further studies are needed to gain further insights into this. Finally, contrary to studies of SNS participation (Mossberger et al., 2007), we found little association between inhibition and demographic characteristics. However, we did not include ethnicity, which might be a relevant background variable in terms of inhibition.

Conclusion
Inhibition is a complex multidimensional phenomenon, particularly in relation to digital platforms. There are many reasons why users of SNS feel inhibited from taking part, some of which are similar to the feeling of inhibition in offline situations, and some are specifically related to SNS. In this study, we argue that inhibition can advance our understanding of online non-participation since it functions as an in-between concept to study the intent behind apparent passivity in the world of SNS. In other words, we claim that inhibition should be understood as a conscious activity and as a way of theorising such intent. Furthermore, we demonstrate that the majority of users of SNS experience inhibition and that certain dimensions of inhibition are significantly related to users engaging less in public SNS-activity. Therefore, we consider the study of inhibition to indeed function as a study of the silent majority, often overlooked in the research of SNS. This study also demonstrates that we cannot easily understand inhibition through one specific theoretical angle. It explores the floating boundaries that exist between private and public in the world of SNS, where we find that inhibition from voicing one's opinion and public non-participation is mostly related to insecurity associated with the platform and other users' reactions, as well as with people's efforts at self-presentation, rather than with their civic role. Focusing on inhibition allows us to sidestep the current debates; instead of determining what should and should not count as participation, we focus on why people chose to refrain from participation, allowing us insight into the intent that lies behind these choices. Future research should refine and develop an SNS-inhibition scale, and