Voicing the organization on social media: towards a nuanced understanding of coworker voice and sources of control

Purpose – This article aims to (1) increase the knowledge of how coworkers experience voicing the organization on external social media and (2) deepen and nuance the knowledge of the sources of voice control involved in such communication processes. The study helps understand coworker voicing on social media as situated identity expressions through which coworkers negotiate and contest the organizational identity, thereby co-constituting a polyphonic organization. Design/methodology/approach – The study draws upon a constitutive perspective on communication and a communication-centered perspective on identity and organizational identification to investigate the voicing of organizational members of the Swedish Police Force on social media. The article is based on a qualitative study where interviews with police officers and communication professionals at the Swedish Police Authority constitutethemainempiricalmaterial.Acontentanalysisofselectedsocialmediaaccountsprovidedimportant backgroundinformationtotheinterviewsandenrichedtheunderstandingofcoworkervoice. Findings – This analysis shows that coworkers voice the organization differently. Furthermore, the study of howcoworkersexperiencethisvoicingindicatesthatthese variationsinhowcoworkersvoicethe organization dependonhowstronglytheyidentifyordisidentifywithorganizationalidentityandimageexpressionsvoicedbysignificantothers.Basedontheanalysis,thisstudypresentsfourvoicepositionshighlightingcoworkers ’ varying degrees of identification/disidentification when voicing their organization on social media and reflecting upon their experiences of voicing. Furthermore, the analysis also demonstrates four sources of voice control: (1) management, (2) colleagues, (3) significant non-members and (4) the status and position of the coworker ’ s voice. These four sources of voice control influence coworkers ’ voices on social media. Practical implications – This study also contributes with practical implications, for example that the traditional idea of monophonic organizations must be revised and also embrace a polyphonic, bottom-up approachtostrengtheninginternaltrustandorganizationalidentity.Thiscomesnaturallywiththepriceofless controlandpredictabilitybymanagementbutwiththebenefitsofincreasedcoworkerengagementandpride. Originality/value – This study contributes new knowledge and a nuanced understanding of coworker voice on social media and the sources of control that influence coworkers ’ voices.

The introduction and implementation of social media in organizational contexts have increased organizations' possibilities to communicate and interact with various stakeholders (Madsen and Verhoeven, 2019;Verhoeven and Madsen, 2022).While the vital role of coworkers as communicators on social media is widely recognized, there is still a significant need to gain more knowledge about the actual voicing practices and how coworkers understand this voicing (see Cassinger and Thelander, 2020).
In recent years, there has been a growing trend in theory and practice to embrace coworkers as communicators, while strategic communication management more often is framed as a collective responsibility that all organizational members must author, enact and control to create value for the organization (Andersson and Rademacher, 2021).The intensified focus on non-communication professionals voicing is at least partly a result of the increasing managerial attention given to the organizational brand, image and reputation in organizing and management (Kornberger, 2010;Mumby, 2016;Waeraas and Dahle, 2020).Managers' increased attention to the importance of coworkers' voices, especially in service delivery and other forms of customer interaction, is nothing new (e.g.Carlzon, 1989).However, the increasing focus on managing all communication "strategically" has in recent decades contributed to an intensification in efforts to manage coworkers' voices (Christensen et al., 2008;Torp, 2015).From a management point of view, considering coworkers' voices in the strategic management of communication makes sense, for example, as external stakeholders perceive coworkers' voices about the organization and its products and services as more trustworthy and authentic than traditional PR or marketing campaigns that emanate from the official company voice (Henderson and Bowley, 2010;van Zoonen et al., 2014).In other words, several rational and efficient reasons exist for the increasing interest in coworkers as communicators.
Even if coworkers are ascribed an active and vital role in organizations' communication, most previous research has a management-centric perspective on how coworkers' voices can contribute to creating a consistent and positive image and reputation (Waeraas and Dahle, 2020).For example, it is often taken for granted that coworkers act, should act or want to act as ambassadors of an organization (Andersson, 2019(Andersson, , 2020;;Heide and Simonsson, 2021;Mazzei and Quaratino, 2018).The emphasis on employee advocacy, defined as when coworkers speak positively about the organization and defend it from criticism (Thelen, 2020(Thelen, , 2021)), is another example.The management-centric perspective that dominates research in this area emphasizes rationality, instrumentality, order, control, predictability and measurement (Simonsson and Heide, 2021).Clearly, this perspective has certain disadvantages and constraints.The perspective privileges those in power and exaggerates the importance of managers for organizational success, and still neglects to embrace the power of coworkers and their needs and understandings.Even more problematic is that researchers tend to have a too simplistic and un-reflected view of the importance of communication in the enactment of organizations.
Clearly, there is a need to take a profound communicative perspective to better understand the complex organizational reality that goes beyond the more simplistic understandings of organizations as easily manageable containers (cf.van Ruler, 2018;Putnam, 1983).When it comes to epistemology, we follow the social constructionist approach, where organizations are understood as phenomena in being and continuous processes of organizing rather than as stable entities observable "out there" (Tsoukas and Chia, 2002).Following the social constructionist approach means that the concept of "coworker" is more relevant to apply than "employee".While employee implicitly means that organizational members are passive and willing to be managed and controlled, coworker emphasizes that individuals are active co-authors of organizational reality.Further, organizations are compositions of coworking individuals with and without management functions that communicate and co-construct solutions.With this article, we contribute to the emerging communication-centered stream of research that stresses the performativity of coworker voice and investigate coworkers' voicing practices (e.g.Cassinger and Thelander, 2020;Christensen, 2023;Christensen and Christensen, 2022).The vital communicative role of coworkers is also reflected in the concept of communicative organizations, where different voices are valued and actively listened to in order for the organization to develop, adapt to change and lead progress (Heide et al., 2019).
This article aims to (1) increase the knowledge of how coworkers understand voicing the organization on external social media and (2) deepen and nuance the knowledge of the sources of control involved in such communication processes.Our theoretical framework is based on a communication-centered perspective on organization and includes theories on organizational identity and organizational identification (Brown, 2015;see Brown, 2017).Based on our theoretical framework, we understand coworker voicing on social media as situated identity expressions through which coworkers express, negotiate and contest the organizational identity, thereby co-constituting a polyphonic organization together with significant others.
The article builds on an empirical study conducted in the Swedish Police Authority.The Swedish Police Authority entered the social media scene early and today attracts millions of followers on Facebook, Instagram and Twitter.While the use of social media started as a grassroots initiative with police officers using their spare time to communicate on the authority's official social media accounts, many officers now see social media as a natural and vital part of the police job.Given this, the Swedish Police Authority's social media activity stands out compared to the activity of the average organization, and the organization thus presents us with an extreme case, which is more suitable as the conditions to produce rich empirical material are better (Flyvbjerg, 2006).

Theorizing coworker voice
Employee voice can be traced back to the industrial revolution and Karl Marx and Adam Smith, who took the laborer's perspective (Kaufman, 2020;Sumanth and Lebel, 2016).Modern research in employee voice started in the late 1980s, and it has since been a growing research interest in coworker voice which has given rise to different suggestions on conceptualizing voice (Morrison, 2011(Morrison, , 2014;;Van Dyne et al., 2003).Morrison (2011) argues that voice was initially defined as any behavior intended to improve work conditions and that early research focused on removing personal dissatisfaction.However, later research has expanded the concept of employee voice to be less self-focused and more other-and organization-focused (Morrison, 2011).While some definitions focus on communication targeted towards colleagues and managers (see, e.g.Morrison, 2011), voice has also been defined as communication with individuals and groups outside the organization (Bashshur and Oc, 2015).
Another development has been to treat voice as a multi-dimensional construct comprising parallel types of voiceacquiescent, defensive and prosocial (e.g.Van Dyne et al., 2003).This multi-dimensional understanding of voice highlights that voice not only entails speaking up about problems and issues but may also mean expressing positive information and ideas about the organization.The development towards a broader understanding of voice is also evident in Tourish and Robson's (2006) understanding of upward communication as both supportive and dissenting voice in the form of feedback from non-managerial staff to managerial staff.Tourish and Robson point out that managers often suppress coworkers' dissenting voices.Similarly, Morrison (2011) emphasizes that voice entails a risk for coworkers since managers and other coworkers may perceive organizational changes and improvements as threatening since they may challenge the status quo.Thus, as Tourish and Robson (2006) argue, speaking up in the organization may be followed by criticism and sanctions.Moreover, Martin et al. (2015) show that coworkers' perceptions of personal control and autonomy are significant for their willingness to express their voice and engage in constructive dialogue with colleagues and managers.Although it has not yet received the same research attention as internal voicing, speaking up externallypositively or negatively may also evoke adverse reactions and criticism (not least on social media).Finally, it should be noted that coworkers' communication is often described as an extra-role behavior or organizational citizenship behaviorthat is as something voluntary and not part of the employer's explicit expectations or the coworker's formal job description (Morrison, 2014).However, Madsen and Verhoeven (2019) suggest that coworker communication is increasingly considered as an in-role behavior.
In recent years, there has been a growing research interest in coworker voice in the research fields of public relations and strategic communication.For example, public relations researchers have investigated how internal communication drives coworker voice (Thelen and Men, 2020) and how coworker voice is linked to commitment and organizational engagement (Kim and Leach, 2020;Ruck et al., 2017).Although these ways of understanding coworker voice are important, they tend to build on a view of communication as something taking place in an already pre-existing organization, reproducing the understanding of organizations as objective, stable entities or containers (see Putnam, 1983).However, as Ravazzani and Mazzei (2018) highlighted in their article on coworker voice in the form of anonymous online dissent, social media and other online outlets have meant that it increasingly transcends traditional organizational "boundaries" and reaches wider audiences than previously.This increasing possibility for coworkers to actively negotiate what the organization is online challenges the conventional notion of organizations as pre-existing and static entities.It also highlights the potential for researchers interested in coworker voice to embrace the idea that organizations are communicatively constituted (i.e.CCO paradigm) and negotiated (e.g.Schoeneborn et al., 2018;Taylor and Cooren, 1997;Taylor et al., 1996).
One example of research on coworker voice that embraces such a view is Cassinger and Thelander (2020).They draw upon a performative perspective on communication (e.g.Austin, 1962;Butler, 1993) and understand coworker voice as a socially situated performance through which coworkers individually and collectively negotiate the meaning of work with colleagues and an imagined audience.Another example is Christensen and Christensen (2022), who critique the dominant organization-centric perspective on coworker voice and convincingly argue that researchers should consider that organizational members' voices are also informed and disciplined by societal norms and expectations.Waeraas and Dahle (2020) also focus on how coworker voice is disciplined and highlight the disciplining effects of internal proactive and reactive measures, such as coworker policies and sanctions, that managers apply to ensure that coworker voices are aligned with the organizational voice.Christensen and Christensen (2022) and Waeraas and Dahle (2020) thus point to the relevance of focusing on control when theorizing coworker voicing.The significance of the concept of control is further highlighted by Christensen (2023), who suggests that coworker voicing can be understood as a form of identity performance influenced by three identity sources, namely the identities (1) provided by management, (2) desired by coworkers themselves and (3) those approved by others.Christensen suggests that coworkers draw from these three identity sources to perform a situationally relevant identity.
Summing up, in this article we draw upon a communication-centered perspective on coworker voice.This perspective entails a broader view of coworker voice not only directed internally, but also to external stakeholders and including a variety of expressions.Furthermore, the traditional managerial focus of strategic communication is based on a solid belief that it is possible to "orchestrate" organizational members' voice in a particular direction (cf.Shotter, 2008).Hence, in the orchestrated, monophonic form of strategic communication, "each voice is simply fitted harmoniously or systematically into the whole" (Shotter, 2008, p. 516).From a social constructionist approach, in contrast, the many voices of organizational JCOM members enact and re-enact the organizational reality.This is related to Bakhtin's (1984) notion of the polyphonic form of organization where independent voices, in communication, are being related to one another contrapuntally.Organizations as polyphony comprise manifold contrasting voices simultaneously and autonomously expressing themselves (Hazen, 1993).By rejecting the traditional understanding of strategic communication as a means to produce one monophonic organizational voice and embracing the idea of polyphonic organizations, we can better understand coworkers' voicing (Christensen et al., 2015).However, the emphasis on polyphony does not mean that there are no boundaries to voicing and that all voices are equally influential, which makes it pivotal to understand the sources of voice control in polyphonic organizations (cf.Schneider and Zerfass, 2018).
Thus, this emerging understanding embraces the notion that organizations are communicatively constituted and negotiated, and highlights the relevancy of being attentive to what sources of control are in play as coworkers voice their organization in different organizational contexts.However, to date, few empirical studies have investigated how coworkers experience voicing the organization on external social media and why they choose to make their voices heard (van Zoonen et al., 2014) or what sources of control are at play in such voicing instances as well as coworkers' experience of them.Examining coworkers' experiences of voicing the organization is critical as it can provide an understanding of what identity ideals, moral norms and other sources of control influence coworkers' voice socialization and their continuous evaluation of their voicing.Next, we will further develop this communication-centered approach and its relation to organizational identity and identification.
Coworker voice, organizational identity and organizational identification: a communication-centered approach As already discussed, a communication-centered perspective within organization studies has directed attention to the fact that organizations come into existence through the many voices that talk them into existence in discourse (Taylor and Cooren, 1997).This implies that an organization does not exist as a singular entity with a voice on its ownas suggested by functionalist communication-as-transmission-oriented perspectives; an organization must always be voiced by agents speaking on their behalf (Benoit-Barn e and Cooren, 2009;Taylor and Cooren, 1997).An agent can be anyone or anything that, given its actantial roles (Taylor and Cooren, 1997; see also Greimas, 1983), has a legitimate claim to speak on behalf of the organization, for example as a "coworker," "customer," "journalist," or "virtual assistant".However, how legitimate the claim of the agent is, and thus how authoritative the agent's voice is, depends on the interlocutors' statuses or positions (Searle, 1976) and other communicatively constituted institutions (e.g."markets") that assign agents different actantial roles (e.g."customer").This highlights the fact that although any agent can speak of the organization and thereby claim to speak on its behalf, the performative force of an agent's voice depends on (1) the status and position of the agent and (2) the institutions made present through the actantial role the agent performs when speaking or writing.The status of coworkers is particularly high since this actantial role involves daily acting as principal for the organization.
Social media has enabled coworkers to make their voices heard to a greater extent than before and has become a central discursive arena on which the identity of the organization, what the organization is (Albert et al., 2000;K€ arreman and Frandsen, 2020), is constituted, negotiated and contested.In our analysis of how coworkers voice the organization on social media and experience this voicing, we draw upon communication-centered approaches to organizational identity.Traditional, functionalist perspectives on organizational identity tend to conceptualize it as something essential, objective, with tangible features (He and Brown, 2013).Constitutive and discursive perspectives instead conceptualize organizational identity as constituted "by the multiple, changing, occasionally consonant, sometimes overlapping, but often competing narratives centered on them, authored by those who participate in them" (Brown et al., 2005).Brown, Humphreys and Gurney point out that this constitutive process is never neutral, given that the many voices that constitute the organization are entangled in asymmetrical power relations that enable and constrain the voices' possibilities to author the organization.Moreover, the communicative constitution is a continuous process given that "meanings are never permanently fixed, and control over discursive space is never total" (Brown et al., 2005).Communication-centered perspective on organizational identity thus enables us to be attentive to the subtle identity struggles (Fleming and Spicer, 2008) as coworkers voice the organization and when they reflect on their experience of voicing.
Moreover, the organizational identity usually gives organizational members a more or less enduring answer to the question "Who are we?" (Chaput et al., 2011).By including the voices of significant non-members in the constitution, negotiation and contestation of organizational identity, we also acknowledge the role of significant non-members' answers to the question "Who are they?"Here, we draw upon Hatch and Schultz's (2002) conceptualization of organizational identity as an ongoing conversation between members' identity expressions ("Who are we?") and significant non-members' image expressions ("Who are they?").As pointed out by Hatch and Schultz, organizational culture ("Who we really are") is also central to the ongoing conversation of organizational identity.However, since organizational culture, to a large extent, is tacit, we limit our focus in the analysis to the more explicit identity expressions on the official social media accounts and coworkers' reflections on their voicing.
Furthermore, to make sense of coworkers' experiences of voicing their organization, we also draw on theories of organizational identification to analyze how coworkers, in their accounts, identify/disidentify with their organization.Communication-centered perspectives on organizational identification understand identification as accomplished either through individuals' self-narratives (Scott et al., 1998) or through communicative processes "through which people develop a common basis for collective action" (Chaput et al., 2011, p. 253).We draw upon both understandings in our analysis since we are interested both in coworkers' actual voicing on social media and their experience.

Method
As we draw upon the epistemology of social constructionism, we will focus on the process of enacting a socially constructed reality through humans' actions, interpretations and communication (Berger and Luckmann, 1966).The study aims to investigate the interpretations of coworkers at the Swedish Police Authority to deepen the understanding of voicing the organization on social media.Consequently, we have conducted a qualitative case study that may produce unique, context-dependent knowledge, which is advantageous when developing new knowledge (Flyvbjerg, 2006).Qualitative research is not some technique for collecting empirical materialit is an approach based on the idea that reality is socially constructed (Morgan and Smircich, 1980).
The studied organization is not an average organization but may instead be described as an extreme case with ideal prerequisites for the successful use of social media.As such, the Police Authority is an intriguing organization to many followers.The core operations concern activities such as law enforcement, general social order and public safety that entail a large portion of excitement and dangerous situations.Furthermore, the police profession is also heavily depicted in pop culture, increasing the interest in police officers and their work.
The empirical material encompasses interviews and a content analysis of eleven social media accounts.We interviewed 13 coworkers, nine police officers and four communication practitioners, selected based on a purposive sampling strategy (Palys, 2008;Patton, 2002).To get as information-rich "cases" as possible, we only interviewed police officers active on social media and communication practitioners involved in the authority's social media work.Another selection criterion was geographic regions, covering urban and rural areas, as that may reflect different kinds of police work.Furthermore, the persons we chose to interview were also active on the social media accounts included in the content analysis.While we mainly focus on the interview material in this article, the content analysis provided an important background material and enhanced our understanding of the voicing on social media.Having an insight into and an understanding of the interviewees' actual communication practices on social media enriched the interviews, as we and the interviewees could relate to concrete examples in our conversations.On average, the interviews lasted one hour and were all recorded and transcribed verbatim.The semistructured interviews focused on motives for communicating on social media, communicative coworkership, preconditions, content, interaction and relationship building.
As mentioned, we also conducted a content analysis of posts on some of the authority's official social media accounts.Based on discussions with our contact person at the police authority, we chose to analyze three Facebook accounts (43 posts), five Instagram accounts (72 posts) and three Twitter accounts (395 posts) for one month.As we wanted informationrich cases, we only chose social media accounts with regular postings, that is that were actively used.We also wanted to cover a variety of voicing and included accounts on three different social media platforms.Furthermore, the chosen accounts were a mix of accounts run by a team of police officers and some accounts run by just one police officer.The latter are known for having many followers and a relatively strong position, and may have an opinionleading role.It should also be noted that the selected accounts were the authority's official social media accounts, and we have consequently excluded coworkers' private accounts (some police officers are active opinion leaders in their private accounts).
The analysis of the interviews is based on abduction, which is an inference technique in which the researcher searches for the most likely explanation of an observation, for example by iteratively going back and forth between the empirical material and relevant theories to produce new insights and understandings (Alvesson and K€ arreman, 2011;Alvesson and Sandberg, 2013).After the interviews were transcribed, we read and re-read them several times, looking for patterns in how the interviewees" positioned themselves vis-a-vis the organization in their account and how they experienced voicing the organization on social media.Our search for patterns in the interviewees" accounts was guided by the analytical generalization method positioning (Halkier, 2011), as this is a suitable analytical technique in qualitative inquiries in which the researcher is interested in understanding the constitutive communication processes in and through which individuals express and negotiate something and how they relate to it.During the readings of the material, we continuously consulted relevant theories on coworker voice, organizational identity and identification both to search for reasonable explanations of the patterns we found in the material and also to challenge our interpretations in line with a reflexive pragmatist approach (Alvesson, 2011).
The aim of the content analysis was to gain a descriptive understanding of what police officers voice in official accounts.Therefore, in comparison to the abductive approach in the analysis of the interviews, we employed a more inductive approach during the content analysis in the sense that the categories we identified, "Crime prevention", "Law enforcement", "Police work in everyday life", "The Police Authority as an organization", "My life as a police officer", "The judicial system" and "Society-related", were derived from the empirical material without much consultation of relevant theories during the analysis.The result of the content analysis is presented in the below table (see Table 1).

Results and analysis
In this section, we will present and discuss the findings of the analysis of the empirical material.Firstly, we will focus on the interviewees' reflections on who they are on social media.Based on how they position themselves vis-a-vis the organization as they voice their organization on social media accounts and reflect upon their experiences of voicing, we present four voice positions highlighting coworkers' varying degrees of identification/ disidentification.Secondly, we will focus on the issue of voice control and the four different sources of voice control we identified in the interviewees' accounts.

Whose voice do the police officers express on social media?
The social media accounts selected for this study were all official authority accounts, that is established and administrated in the name of the Police authority.Thus, the individual police officers communicating on these accounts are expected to do that on behalf of the authority.However, posts are often signed with the police officers' first names and written in a relatively personal tone, which raises the question of how police officers understand who they represent and how to balance being an authority representative and still come across as personal and authentic.
Our analysis shows at least four different voice positions highlighting how coworkers identify/disidentify with the organization as they reflect upon their voicing on social media, namely those positioning their voice as (a) the organization's voice, (b) the organization's voice through self-narratives, (c) the police officer's voice in organizational accounts and (d) the police officer's voice in private accounts.Below we will discuss these four voice positions, starting with the organization's voice.
The Organization's voice.For some coworkers, voicing the Police means speaking with the organization's voice.For example, one interviewee, who is an authority representative on the official Authority account, emphasized an actively refraining of voicing private opinions: "[. ..] you do not post it to have any personal gain, but because the Authority wants to go out with something, some message".Similarly, another interviewee stressed that Police officers should express the Authority's voice and not their own, even though he also acknowledges that it, at times, is hard to make a sharp distinction between the voice of the Authority and the voice of individual Police officers: I was about to say I do not know which is the chicken and which is the egg because for 21 years now, I have been trained in the Authority.Many of the Police's opinions or ways of being have shaped me so that I may have become this correct bureaucrat.Because my opinions are things that we choose to report or on things that are in some way criminal, for example, [and then] it is easy for me to take on that role.[. ..]So, they kind of go hand in hand or merge into each other, the Authority's guidelines These examples thus illustrate the voice position of those police officers, who in their reflections on their voicing, express the strongest degree of identification with the organizational voice.For these officers, their presence and voice on social media is regular police work, which includes tasks such as informing the public and asking for tips to help solve crimes, but in a digital context.Many of the interviewees explain that social media is an excellent complement to "regular" police work as many times they can reach a wider public with their information, as the following example illustrates: Social media posting aims to reach many more than we did before.It passes maybe 100 cars when we have a night car check, and when we talk about it, it may reach 20,000 followers.Our purpose is to educate the public, to show, for example, what people do wrong and the consequences.
Additionally, several interviewees experienced that this increased reach has had positive implications for their work as people perceive them as more present in their local Police district.For example, one officer explained: "Maybe someone thinks we are never in that particular square.But if they follow us on Instagram or Facebook, they see that we are there daily." In short, those interviewees who understand their voice as the organizational voice see themselves as an extension of the police authority; they represent the official standpoint of the authority, and they have a role in developing the trust of the police as an authority.For them, personal opinions do not belong in official Authority accounts as the accounts should be used mainly to perform what they perceive as "regular" Police work, such as informing the public about laws and regulations and asking followers for tips when a crime has been committed.
The Organization's voice through self-narratives.Other interviewees positioned themselves as voicing a more personal version of the organization's voice.For example, in our analysis of how police officers voice the organization in the official accounts, we found that almost 60% of the 510 analyzed posts contained content giving followers an insight into the everyday life of police officers.One interviewee explained that the reason for giving followers an insight into the work and what happens "behind the scenes", is that followers appreciate that type of content and the transparency it provides: Then we try to make it a bit more personal.We might tell a little story about what happened during the work shift or something like that.You get to follow along and see what happens during a work shift.And it can be anything from the fact that we are going to do a traffic control here and here or have done, or now we have just visited that school, to, well, now we're polishing our boots or washing the car.That you get to see a bit more police in everyday life.What we actually do.And I think this has actually been very successful.People find it really, really fun to follow.It gives them a little more insight into the work of the police.
Another interviewee explained that followers are given an insight into everyday work as an alternative image to more negative ones: For me, the primary purpose [of posting] is to build trust, at least in this with social media.It is very important because what you do is very visible, and it is often taken out of context in many ways.And it is easy to stand there and watch an intervention, film it, and scream that we are doing it wrong.And maybe we do that sometimes too! [. ..] we are also human.We can also make mistakes.So it's clear that we're not against being scrutinized and questioned.When I was a teenager, I saw an intervention where I thought the Police acted brutally and unfairly.It colored my image of the Police for a long time to come.And that is what you want to counteract.In fact, the majority of us actually work because we want something good and want to do good and want to help.Thus, in contrast to those interviewees who emphasized that police officers should refrain from expressing personal opinions when voicing the Police on official accounts and who understand themselves as the organization's voice when voicing on social media, the Police officers that post about the everyday work want to contribute with an alternative story of the Police that stands in contrast to stories that flourish on social media or stories presented in the traditional media, often with large portions of criticism.Therefore, these officers perceive social media as an important platform that enables them to voice the organization in their way, thereby providing counter-narratives to the more negative Police Authority narratives circulating.
The police officer's voice in organizational accounts.Other interviewees positioned their voice as distinct from the organization's voice, as the following example illustrates: [. ..] my account is an official authority account, but I am not the Authority.That difference is not always easy to understand [. ..].But I am trying to be clear by saying that this is me as a police officer at NN communicating.An orderly police officer at NN, but of course, I am often drawn into discussions that concern the authority in general.However, then I usually try to be clear by saying that this is not the Authority's opinion, but this is my opinion.Many times, there are things I cannot answer about why the authority does this or that.
As the quote shows, this officer clearly distinguishes between personal and the Authority's opinions.Thus, in contrast to those interviewees that positioned their voice as the organization's voice and who stated that police officers should refrain from expressing personal opinions when voicing the Police in official accounts, this interviewee sees no problem with expressing opinions on different societal matters and on work situations that they encountered.This tendency of some police officers to voice personal opinions was something we could also see in our analysis of the official social media accounts.One of these more opinionated interviewees explained that management even encourages them to express personal opinions on different matters: Given the officer's formal role and social media activity, this police officer has attracted many followers and was even referred to as a "social media star" by one of the other interviewees, who took this officer as an example of more "opinionated" voices in official accounts.While the officer posts as a person, this account is an official Authority account.However, the officer underlined that the posts are personal, not from the Authority even when asked by management to communicate something: If someone is going to make a statement about something that is a bit difficult, they often come to me and say, "Can't you take this?You've kind of got a knack for this".I do not know about that.But sure, I can usually help, but with that comes the price of not being able to tell me what to think and feel [.This voice position thus starkly contrasts the one taken by those interviewees stating that police officers should refrain from expressing personal opinions.Furthermore, interviewees taking this voice position make a more apparent distinction between the voice of the organization ("they"), often associated with the voice of management and their voice ("I").Compared to the two previous voice positions, interviewees taking this voice position thus demonstrate a higher degree of disidentification with what they perceive as a more undesirable organizational identity oftentimes voiced by management.
The police officer's voice in private accounts.The fourth voice position we identified in the analysis was those police officers who post content from their work but in their private accounts.We did not interview any of these officers, but they were frequently mentioned by both the police officers and non-police officers we interviewed.Although some saw it as somewhat problematic and potentially trust-damaging, most interviewees took a relatively neutral stance towards these private accounts, pointing both to the opportunities and risks of posting in your professional role in private accounts.For example, one police officer explained: There are those who have private accounts that are police officers who are not an official government account, so they are freer to write what they want.And there's also freedom of expression, so of course, you can do that unless there is some specific secrecy or something.But I feel that it can sometimes be a bit damaging if you do not think about it.I do not have a specific example now, but the individual police officer who runs this type of account has a great deal of responsibility.At the same time, they are also very good in many ways.But there I can see that sometimes there might be a little disadvantage, you could say.
Similarly, another interviewee, a communication practitioner, also maintained that these accounts are valuable to the Police but that posting as your professional self in your private accounts comes with great responsibility: I would say that most of those with such accounts publish great posts and do great advertising for the Police and that the followers get great insight.But there will always be someone who sometimes posts something that is less well thought out.And then the public can have quite a hard time distinguishing between individual police accounts and the Police Authority.You must remember that too.
Although we did not include private accounts in our content analysis, we scanned some popular private accounts with many followers.We noted that these officers mostly voiced the profession rather than the organization in that most police-related posts were posts providing followers an insight into their work as police officers.While not a systematic analysis, we also noticed that some private accounts were fearless in voicing their critique against management and managerial decisions they disagreed with.Thus, this fourth voice position indicated a stronger identification with the profession than the organization.Furthermore, like those interviewees who positioned themselves as the police officer's voice in official accounts and at times disidentified with management's voice, the police officers posting professional-related content in their private accounts also take an active stance against management from time to time, thereby using their forceful voice (due to a large number of followers) to publicly contest and negotiate "Who we are".

Sources of voice control
A recurring story is that the authority's communication on social media started as a task performed by some true enthusiasts who did a great job during their free time, that is it started as a bottom-up process rather than a management-driven initiative.Today, it seems that most police officers who are active on the authority's official accounts are compensated and allowed to spend time on social media during working hours.Nevertheless, some interviewees reported that they often check the account in their spare time, which is still a task of their free will (rather than being part of their formal job description).While the interviewees stress that they are free to choose a topic and tone in their communication, they often add that this is "freedom within certain limits."In the analysis, we identified that these limits derive from at least four sources that influence, or control, coworker's voicing on social media: (1) management, (2) colleagues, (3) significant non-members and (4) the status and position of the coworker's voice.
The first source of voice control, management, can be exemplified by the handbook for those who use social media in their service.The social media handbook is about 30 pages and provides guidelines about aspects such as purpose, target groups, language use, attitude and use of images.The purpose, presented in the first paragraph, states that: "the Police Authority is to be perceived as a uniform and clear sender on social media."Despite the lengthy guidelines, the police officers do not seem to consider the handbook particularly limiting.The experience is that the handbook provides some rather basic information that is valuable, especially in the beginning.The interviewees talk about these rules as supportive rather than oppressive.Also in a more general sense, the interviewees tend to stress voluntariness, room for maneuvering and being trusted by the managers.One police officer reflected: "It's quite unique that an employer lets one go like this.I have had previous jobs where you would not have that trust in your employees."The trust given to speak on behalf of the organization is obviously experienced as quite rewarding, and the desire to live up to that trust may explain why the police officers seem to find rules and guidelines supportive.
The second source is the colleagues.Several interviewees talk about the importance of the learning process taking place in teams of writers who manage the same account.For example, one police officer describes this process as follows: [. ..] and we also help each other.So, for example, if I feel I might want to post something like this, then I can ask [my colleagues in the team]: what do you think about this?Because then you can also help each other in this group.Uh, but that may not be so good, or yes, that can work great, but you may need to change this wording, for example.
Another police officer says everyone decides what and how to write, but the ground rule is that they show each other images and text before publishing.The same police officer further argues that "it is not perceived as a burden, but it's great that we have the opportunity to discuss what might be good and less good."We conclude that formal rules and guidelines, as well as informal norms developed collectively in teams, enable and constrain coworkers voicing on social media.
The third source influencing coworkers' voice are significant non-members, mostly followers, through their reactions to police officer's voicing on social media.The interviews demonstrated that police officers consider the reactions of social media followers in the continuousand often collectivelearning process of what works and does not work on social media.
We made a post about crashing cars, and then someone wrote that you should not be bossy, and then I apologized.We get criticized pretty quickly if we have a tone, which is not okay.Even though you did not mean it when you wrote it, someone can comment that it was ironic and bumptious to write like that.Thus, in line with Christensen and Christensen's (2022) argument, not only organizational norms but also societal norms and expectations are disciplining coworker voices.The public's expectations of how a trustful police officer should act are thus, to some extent, informing and controlling coworker voice on social media.Even if there are not always direct reactions to the posts, the comprehensive visibility of the voicing has a disciplining effect (see Christensen and Christensen, 2022;Madsen and Verhoeven, 2016;M€ uller, 2017).The combination of a visible speaker and an invisible audience makes the speaker quite vulnerable to criticism and triggers processes of self-discipline (Christensen and Christensen, 2022).
A fourth source we identified in the analysis is the status and position of the coworker's voice.Some of the police officers voicing the Police either in official or private accounts on social media have several thousand followers, and despite sometimes pushing the boundaries, they are recognized and appreciated by their colleagues.The many followers and the near-celebrity status grant some police officers a higher status and a unique position.For example, a communication officer comments: JCOM In a way, if you look at NN on Twitter, there are a lot of personal opinions.But there, I think they have made a small balance between the gains and losses because if he did not have any personal opinions, his account would not be so large, and we would not benefit from it.[. ..].If we were to have a crisis or if something happens that we need to inform about, then we call NN, and he publishes it.So, he is a crucial channel for us.[. . .]We are a little pragmatic; I would call it.Should it be about value issues, that he expressed things like racist opinions, we would never accept that.So, it is within certain limits, and if he were to cross such a limit, you would react.
Thus, whereas all police officers are legitimate agents having an actantial role granting them the status and position to represent and voice the organization, the voices of those coworkers who are well-known and have many followers seem to have a stronger performative force in the sense that their voice is taken into account in the ongoing negotiation to a greater extent than other officers' voices.In turn, this offers more possibilities for them to push established norms of how to act as an organizational representative and voice the organization.

Concluding discussion
In this final part, we will first summarize the main findings and clarify the theoretical contributions of our study.After that, we will discuss some practical implications for managing and working strategically with coworker communication on social media.

Summary of results and contributions to theory
In this article, we investigate (1) how coworkers understand voicing the organization on external social media and ( 2) what sources of control are involved in such communication processes.Previous research on coworkers' roles as communicators has to a great extent, employed a management-centric perspective focusing on how to motivate and manage coworkers as advocators for the organization.Paradoxically enough, coworkers' understanding and actual communication practices have hitherto been scant in research on coworkers as communicators.As a response, this study takes the perspective of coworkers and draws upon a communication-centered perspective on organization, organizational identity and organizational identification.Thereby, we aim to contribute to the emerging stream of research that emphasizes the performativity of coworker voice and investigate the sources influencing coworkers' voicing practices (e.g.Cassinger and Thelander, 2020;Christensen, 2023;Christensen and Christensen, 2022;Waeraas and Dahle, 2020).As our analysis shows, coworkers' voicing on social media is a complex, multifaceted process involving much more than ambassadorship and advocacy, incited and controlled by managers.Coworkers' voicing practices on social media and their reflections on voicing places the ongoing negotiation of "who we are" (i.e. the organizational identity) center stage on social media.The results from the authority being studied here show that coworkers' communication on social media is much more a process of building identity and trust rather than just advocating or branding the organization.
Our analysis contributes to a more nuanced understanding of the complexity of voicing.We identified four voice positions highlighting coworkers' varying degrees of organizational identification/disidentification, when voicing their organization on social media accounts and reflecting upon their experiences of voicing.Consequently, we add to previous literature by emphasizing the performativity of coworker voice and showing the variations in coworker voice and coworkers' rationalizations for voicing.Those interviewees who strongly identified as the organization's voice, emphasized that police officers should refrain from voicing personal opinions on official social media accounts.Instead, they emphasized the importance of presentifying the organization (Cooren, 2009).In other words, making the voice of what they perceive as the official Authority and official rules and regulations present through their Voicing the organization on social media voicing.This voice position thus stands in strong contrast to the other three voice positions as these officers emphasized either voicing their version of the organization, their personal opinions or their discontent with other voices (often management's voice) was important to them.Previous research has shown that members may manage a variety of discursive positions in relation to the organization (e.g.Elsbach and Kramer, 1996;Toyoki and Brown, 2014).We also noticed this in the analysis, as exemplified by how several interviewees switched between referring to a "we" and "they" depending on what organizational identity and image they positioned themselves against when reflecting upon whom they voice on social media.Thus, although we identified four different voice positions highlighting coworkers' varying degrees of identification/disidentification, these positions should be understood as analytical abstractions.In everyday voicing on social media, police officers may take different positions vis-a-vis the organization, depending on the topic or what significant members (e.g.management) or nonmembers (e.g.journalists, followers) have voiced.
We also identified four sources of control that influence coworker's voicing on social media.Thereby, we contribute to previous conceptual and empirical research highlighting the simultaneously enabling and constraining power of societal norms and expectations (Christensen and Christensen, 2022), identity material provided by management, colleagues and the coworker itself (Christensen, 2023) as well as the social media platform conventions and organizational culture (Cassinger and Thelander, 2020).We identified the peer-based influence of colleagues as the most significant source of voice control.Previous research on peer-based concertive control in self-managing teams emphasizes that colleagues' either sanction and discipline or affirm to influence each other's behavior (Barker, 1993).The police officers in our study instead talked about a less forceful form of concertive control exercised through the team reflections on each other's voicing.Thus, our study partly supports previous studies by demonstrating that several of the sources of control identified in previous research are important to understanding coworker voicing on social media.However, we also offer additional insights.For example, we show that the status and position of the coworker's voice, depending on the number of followers the coworker has, also enables, and constrains the performative force of coworkers' voicing on social media.
Summing up, we hope to broaden and advance the research agenda on coworker communication on social media by having demonstrated the relevance of a communicationcentered perspective and of seeing coworkers as active co-constructors of organizational reality.It often appears to be taken for granted that employees' communication only creates value when it adds to a consistent, univocal communication and is supportive of management goals.In this article, we have shown that there is both more complexity and value in coworkers' communication than management-centric research in strategic communication acknowledges.

Practical implications
As previously discussed, there is a growing consensus on coworkers as key actors in organizations' strategic communication, not least on social media.Although not always formally or explicitly defined as part of the responsibility of employees, many organizations still expect their employees to take an active and engaged communication role (Madsen and Verhoeven, 2019).The question then is how to motivate coworkers to do that and how to support their voice without making it lose its authenticity and coworkers' inherent engagement (cf.Heide and Simonsson, 2018).
In too many organizations there is still a monophonic ideal, reflected not least in strategy documents emphasizing the importance of speaking with one voice to be perceived as consistent and trustworthy (Simonsson, 2021).The monophonic ideal tends to nourish topdown control and message alignment, which counteract coworkers as active and creative communicators.One implication from this study is the value of embracing a polyphonic or sometimes even a cacophonic approach to coworker communication.
As indicated in the analysis of our results, there are examples of coworkers' communication giving rise to a somewhat cacophonic organization.However, these dissenting voices simultaneously contribute to a vibrant and inclusive organizational identity negotiation, and it is reasonable to assume that the possibility for coworkers to speak freely is beneficial for intra-organizational trust and the feeling of psychological safety (see Edmondson, 2019).Furthermore, the opportunity to voice the organization is highly appreciated by coworkers, who are sometimes frustrated with how "others" (e.g.journalists, managers, followers, colleagues) speak of their organization and occupation.Making their voice heard and telling their story is crucial for coworkers to feel pride in their organization and their work and increase the public's trust.Thus, rather than speaking with one voice, it can be a strength that the organization speaks with many voices to show the public the plurality of voices, individuals, experiences and opinions that make up the collective "we".The results from this study imply that organizations that embrace the ideal of a polyphonic organization in which coworkers are allowed and even encouraged to make their voice heard must accept a degree of cacophony as this is probably an inevitable "byproduct".However, opposing efforts to shape coworker voice in line with the univocal voice ideal frequently have negative consequences such as frustration, cynicism and even psychological and emotional harm among coworkers (Waeraas and Dahle, 2020).Managerial efforts to "orchestrate" coworker voices may clearly also weaken the organizational "we" and identification, which seems to be one of the strongest preconditions for coworkers' willingness to voice their organization on social media.Thus, if coworkers do not identify with or feel pride in their employer, the first step is to work with that before trying to turn coworkers into brand ambassadors on social media.From external stakeholders' perspective, a more univocal and directed voice will most likely be perceived as less authentic.Authenticity comes from being invited behind the scenes to learn about coworkers' everyday work life and get to know some coworkers.
From a communication management perspective, a polyphonic, bottom-up approach, comes, of course with the price of less control and predictability.However, a bottom-up driven strategy does not mean that there should be no rules or guidelines at all (and in many cases, there are also laws that, of course, need to be followed).As indicated in this study, if coworkers feel trusted, they also want to live up to that trust and, hence, they also appreciate some guidelines on how to be supportive of organizational goals.What can also be done from a management perspective is to initiate and facilitate collective conversations about what it means to voice the organization on social media, but also what tensions and challenges it brings.In this study, we found that peer-based influence seems to be the most important source of control, but sometimes those collective learning and influence processes do not happen automatically but need to be facilitated.Such conversations are also an important opportunity for managers to listen and learn from coworkers and to show their appreciation of their coworkers' communication engagement.
. .]I am not a megaphone for what I feel are stupid decisions.[. ..]I have no desire to be a megaphone for anyone.I believe in this organization, and I want to be able to tell it.What I see, what I think and feel.Then I have to understand that there are risks in being too open-hearted.

Table 1 .
Result of content analysisand opinions and our own [. ..]And if we have an [personal] opinion, we just have to keep quiet about it.For example, I may not think that A-tractors should be allowed to exist, but now the rules are in place, and I cannot, and would never, write that they should not exist [. ..] it would be very strange to express that opinion.