Dealing with risk in stakeholder dialog: identification of risk indicators in a public service media organization's conversation and discourse with citizens

Abstract

Purpose

This study operationalizes risks in stakeholder dialog (SD). It conceptualizes SD as co-produced organizational discourse and examines the capacities of organizers' and stakeholders' practices to create a shared understanding of an organization’s risks to their mutual benefit. The meetings and online forum of a German public service media (PSM) organization were used as a case study.

Design/methodology/approach

The authors applied corpus-driven linguistic discourse analysis (topic modeling) to analyze citizens' (n = 2,452) forum posts (n = 14,744). Conversation analysis was used to examine video-recorded online meetings.

Findings

Organizers suspended actors' reciprocity in meetings. In the forums, topics emerged autonomously. Citizens' articulation of their identities was more diverse than the categories the organizer provided, and organizers did not respond to the autonomous emergence of contextualizations of citizens' perceptions of PSM performance in relation to their identities. The results suggest that risks arise from interactionally achieved occasions that prevent reasoned agreement and from actors' practices, which constituted autonomous discursive formations of topics and identities in the forums.

Originality/value

This study disentangles actors' practices, mutuality orientation and risk enactment during SD. It advances the methodological knowledge of strategic communication research on SD, utilizing social constructivist research methods to examine the contingencies of organization-stakeholder interaction in SD.

Citation

Schwägerl, C., Stücheli-Herlach, P., Dreesen, P. and Krasselt, J. (2024), "Dealing with risk in stakeholder dialog: identification of risk indicators in a public service media organization's conversation and discourse with citizens", Journal of Communication Management, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/JCOM-01-2023-0002

Introduction

Recent approaches to strategic communication have drawn upon conversation relevance. Zerfass et al. (2018, p. 493), for example, defined strategic communication as “the purposeful use of communication by an organization or other entity to engage in conversations of strategic significance to its goals.” Organizations' engagement in conversation with stakeholders is motivated by various strategic management goals, such as adjusting activities to meet stakeholder expectations (e.g., Freeman, 1984), maintaining legitimacy (DiMaggio and Powell, 1983), or consulting stakeholders about specific issues (e.g., Johansen and Nielsen, 2011; O’Riordan and Fairbrass, 2008).

A specific type of organization-stakeholder conversation is stakeholder dialog (SD), which can take several forms, such as public meetings (Lane, 2018; McComas, 2003; Russmann and Lane, 2020; Van Burgsteden et al., 2022; Weder, 2022) and mediated online dialog (Elving and Postma, 2017; Hetze et al., 2019). SD is used by corporations and NGOs to deliberate on organizational policies and activities (Brand et al., 2020), as well as by governments mandating organization-stakeholder dialog to obtain legitimacy of decision-making or when statutory regulation demands it (Lane, 2018, p. 658).

As a communication praxis of strategic management (Rüegg-Stürm and Grand, 2015), SD engages stakeholders in conversation to reveal different perspectives and conflicting interests, and to attain reasoned agreement (Ferraro and Beunza, 2018; Lane, 2018; Russmann and Lane, 2020). SD inevitably leads to controversies (Rüegg-Stürm and Grand, 2015); it aims to establish legitimacy and relationships resulting from such controversies through “organizational common sense” (Shotter and Cunliffe, 2003, p. 23) or “metaconversation” (Robichaud et al., 2004). SD planning and implementation entail decisions about practices that initiate and facilitate dialog with citizens and, thus, the co-production of organizational discourse (Cooren, 2015). These decisions should consider the implementation of specific norms of democratic and rational deliberation (Ferraro and Beunza, 2018; Gilbert et al., 2023).

This study focuses on an SD where citizens were invited to participate in interconnected meetings and forums. The study uses the case of a German public service media (PSM) organization to interrogate organization-citizen interaction corresponding with or damaging SD’s aspiration to create mutuality orientation. It analyzes these interactions and the resultant discourse.

Germany has two main PSMs. One used SD for the first time since its founding in 1950. Public debates on PSMs concern funding, impartiality, digital transformation and PSM performance, including audience perceptions of media offerings, journalistic role expectations and quality relative to the media’s remit in democratic societies (Sehl, 2020; Sehl et al., 2022). The German PSM engaged in dialog with citizens about their expectations of its present and future performance in an endeavor to scale SD up to public discourse on its performance and funding in the country (Buhrow, 2021). The SD specifically derived implications for short-to mid-term programming changes, which the PSM could implement within its mandate, and summarized suggestions for media policy regarding its remit, structure and funding (Arbeitsgemeinschaft der öffentlich-rechtlichen Rundfunkanstalten der Bundesrepublik Deutschland, 2021).

Deliberation exposes stakeholders' concerns and issues in situated dialog (Gilbert et al., 2023), the dynamics of which elude interactional control. From the organization’s perspective, it wishes to mitigate this risk. SD’s open-endedness conflicts with PR’s position in a managerial paradigm that seeks to control and predict developments (Kent and Taylor, 2002). The SD’s capacities unfold while risks are enacted. However, mitigating risks through interactional control will be to the organization’s own detriment (Kent and Taylor, 2002). Moreover, controlling interaction, based on the idea that “the process determines the result” (Kaptein and van Tulder, 2003, p. 221), risks losses for all involved. Studies reveal such attempts (e.g., Lane, 2018; Russmann and Lane, 2020). However, risk indicators suggesting a potential dialog breakdown while SD occurs remain nonoperationalized. Understanding the forms and emergence of SD’s risks allows strategic communication to validate the consequentiality of the decision to engage in SD as a delicate conversation genre.

This study encompasses two interrelated objectives:

1. First, it identifies risk indicators in SD’s discursive processes, that is, the emergence of topics and identities indicating the risk that actors (organizers, citizens) retreat to their positions, compromising SD’s deliberative quality.

2. Second, it illustrates method development, mobilizing conversation analysis, corpus linguistic discourse analysis and their triangulation to advance the methodological knowledge on SD analysis in diverse interconnected formats and determine how risks emerge in interaction across a single format.

Organizers' openness to risky interaction (from their own perspective, see Kent and Taylor, 2002) in SD can be verified through their practices. SD is a facilitated and managed co-production of organizational discourse, aimed at enacting an organization’s risks to develop strategies and renew legitimization (Weick, 2001). Organizational discourse, as the linguistic expression of facilitated interaction between organizational members and stakeholders, emerges through face-to-face interactions and written text. Consequently, it can be viewed as having a dual meaning: an ongoing structured process of turn production (sequential verbal activities by participants, see Sacks et al., 1974) and an evolving system of typical form and content (Larsen-Freemann and Cameron, 2008) or patterns of language use (Bubenhofer, 2009) that can be identified during SD and establish connections between the texts and contexts of a specific organization.

Literature review

Characteristics and normative foundations of stakeholder dialog

SD enacts organization-stakeholder deliberation according to stakeholder theory (Richter and Dow, 2017) in both verbal interactions (meetings) and mediated written interactions (forums) encompassing temporal and local scaling. Stakeholders are individuals and groups who are affected by an organization’s activities and possess the potential to influence those activities (Freeman, 1984). They hold legitimate expectations of organizational activities, critically evaluate these activities and can exert pressure on the organization (O’Riordan and Fairbrass, 2008). Citizens are PSMs' primary stakeholders (Campos-Rueda and Goyanes, 2022). Studies have not explicated citizens' stakes in PSMs, but, instead, refer to the regulatory frameworks according to which PSMs must provide unbiased and impartial information. This objectivity requires PSMs to conceive of citizens as stakeholders with democratic needs instead of media consumers (Spigelman, 2014).

Stakeholder theory strives to consider stakeholders' legitimate interests in strategic decisions, without prioritizing organizations' self-interests (Freeman, 1984). Along these lines, SD promotes mutually beneficial outcomes for actors (i.e., organizations, stakeholders) and should engage actors in communicative rationality to be deemed valid (Richter and Dow, 2017): that is, achieving consensus through rational argumentation, with participants accepting universal validity claims and accomplishing a shared definition of the situation (Burkart, 2007), or dissent, if actors' views and interests are incommensurable, but actors are willing to accommodate their views on how to handle the issue at stake (Brand et al., 2020; Weder, 2022). Actors' alignment regarding their orientation to SD’s presupposition of mutuality is verified through their communicative practices (see Van Burgsteden et al., 2022).

Public relations research proposes deliberative dialog concepts to examine SD’s validity (e.g., Lane, 2018). These concepts aim to positively impact organization-public relationships (Chen et al., 2020; Dhanesh, 2017; Kent and Taylor, 2002) and are applicable to SD implementation (Lane, 2018). The dialog principles suggested by Kent and Taylor (2002, pp. 24–29) reflect such deliberative qualities: 1. Mutuality: actors' orientation toward an inclusive and collaborative exchange; 2. Propinquity: publics being consulted in matters affecting them and being willing and able to express their expectations; 3. Empathy: mutual support and acknowledgment; 4. Risks: consequences emerging from interaction that were not anticipated by actors yet are regarded as opportunities for improved organization-public relationships; 5. Commitment: privileging interactions that are mutually beneficial to conversation that aims to fulfill self-interests.

Nevertheless, these principles have seldom been implemented (Lane, 2018; Russmann and Lane, 2020). SD is often associated with application scenarios that fulfill organizers' self-interests (Elving and Postma, 2017; Hetze et al., 2019; Johansen and Nielsen, 2011). Dialog is two-way communication in these cases (e.g., see Elving and Postma’s (2017) discussion of SD by corporations on social media), which is, to some extent, discursively scripted by the roles that stakeholders bring to the table (e.g., as clients) and by discursive conventions determined by these roles (Johansen and Nielsen, 2011). However, Ferraro and Beunza (2018) argue that a prerequisite for dialog is the questioning of such roles, instead of actors' persistence in predefined and alleged roles. This understanding of SD exists in the political sphere and company–NGO SDs on social responsibility, with research investigating whether organizers (governments and mandated personnel that conduct and moderate dialog, or companies and mandated personnel) avoid social constraints and invite dissenting views (e.g., Dhanesh, 2017; Weder, 2022).

Risk in stakeholder dialog

Organizations face a threefold general uncertainty in SD: unforeseeable environmental conditions, stakeholders' changing expectations, and organizational performance criteria (Rüegg-Stürm and Grand, 2015). SD addresses these uncertainties neither as threats nor problems but as opportunities for legitimization, change and renewal. Hence, SD is communication praxis of strategic management (Rüegg-Stürm and Grand, 2015). It focuses on strategic uncertainties and risks by deliberately questioning organizational routines and managerial engagement (Grand, 2016) while enacting the “dialogic risks” (Kent and Taylor, 2002, p. 29), which result from actors' unconditional exchanges in dialog’s situation dynamics.

In a corporate context, SD considers a corporation’s social responsibilities in policies and activities (Brand et al., 2020), thus necessitating a balancing of responsibilities with self-interests. PSMs, however, act in the public interest. Organizers are mandated to foster shared understanding of PSM performance and legitimacy as societal issues in a political context that normatively favors public interests over self-interests. Controlling dialog to entrench self-interests would be to the PSM’s own detriment. Thus, SD risks are actors' losses (Kent and Taylor, 2002; Raupp, 2015) if they result from actor practices preventing reasoned agreement and mutuality orientation. The PSM loses the opportunity to ground its strategic decision-making and social license, and citizens lose their impact on the PSM’s activities. Risk enactment, therefore, parallels risk communication: it exposes decision-making, restricts self-interests and strengthens the organization’s long-term strategic ability (Raupp, 2015) by engaging actors in dialog, the context of which is shaped by the tension of self and public interest (Palenchar et al., 2017).

Hence, SD risks stem from the collision of conflicting interests, actors' diverging motives to engage in dialog, and the micro-level performance of dialog (e.g., McComas, 2003). Organizers may clarify actors' mutual perceptions of motives and risks (Umansky and Fuhrberg, 2018) and conflicting positions (Brand et al., 2020) to avert a breakdown in dialog, which requires them to allow for agonism (Davidson, 2016; Weder, 2022). Biased perceptions can lead to mutual distrust and conflicts. A need exists to establish mutual understanding and common ground (Umansky and Fuhrberg, 2018).

However, organizers' perception that controversies and conflicts equate to risk guides them in SD planning and implementation: Lane (2018) and Russmann and Lane (2020) show that organizers perceive opposing opinions as risks, prioritizing organizational goals, controlling interaction and avoiding controversies and interaction with stakeholders with dissenting views, and striving to accomplish predefined outcomes. Organizers fear that dialog may raise the bar to a degree where citizens expect to influence the organization’s agenda, and that situational dynamics may lead to disclosing sensitive information (Lane, 2018).

Deliberative standards are enacted while SD is happening (Gilbert et al., 2023). The extent to which organizers take perceived risks is evident in qualitative interviews with organizers (Lane, 2018; Russmann and Lane, 2020; Umansky and Fuhrberg, 2018) and surveys with citizens (McComas, 2003). Except for Van Burgsteden et al. (2022), the extant literature has not advanced methods to examine the micro level of public meetings or the interactional logics of different SD formats, which the operationalization of SD risks requires. Speakers in conversations are guided by mutual interpretations (Vasilyeva et al., 2020), hence analysis of such interpretations facilitates the specification of the exact points in interaction at which situated dialog is at risk of breaking up.

Dissent affecting mutuality orientation

Deliberative SD, in contrast with organization-centric risk perception and a narrow non-communicative concept of strategy, predicts that risks do not equal dissent as the plurality of opinions is a prerequisite to the constructive process of actors understanding each other’s interests in and views of a common issue. Organizers, for example, display altered understandings as a result of citizens' concerns when conveying the consequentiality this altered understanding has for the SD process, such as when passing concerns to experts for clarification (Van Burgsteden et al., 2022).

Dissent fosters transformative relationships among actors if mutuality orientation toward communicative rationality is established, wherein actors accommodate and apprehend their perspectives (Brand et al., 2020; Weder, 2022). In dialog’s sequential logic, Van Burgsteden et al. (2022) identified linguistic indicators that display actors' orientation to building transformative relationships as a result of organizers' practices to validate citizens' concerns, prompt citizens' explications of concerns and display the concern’s significance. For example, in Van Burgsteden et al.’s study, one citizen suggested a municipality alderman the grammatically informal Dutch pronoun you after the alderman explained how this citizen’s concern could be managed. The alderman’s benevolent approval indicated a shift “from a state of possible or incipient conflict to one of agreement, alignment, and even familiarity” (p. 75).

The above example shows dissent strengthening mutuality orientation as actors shift to a dyadic relationship concerning the specific issue being addressed. Weder (2022) emphasizes organizers' role in actively stimulating dissent through problematization practices such as questioning taken-for-granted definitions of situations and issues, soliciting different perspectives and facilitating the contextual variables and conversational practices that may guide actors' orientation toward enlightening each other.

Using SD to facilitate practices of problematizing relates to its conception as a venue of agonism (Davidson, 2016 in Weder, 2022) through actors' accomplishing views of an issue different from their initial views, hence transforming their would-be robust belief systems (Lorino, 2021). However, agonism relies on actors' willingness to accommodate each other’s views, which organizers must clarify. If mutuality orientation is absent or organizer-stakeholder relations are adversarial, organizers' unveiling dissent can clarify conflicting positions. In cases where these positions remain unclear, actors might retreat to their positions (e.g., Brand et al., 2020). Thus, organizers should clarify views in the interest of mutuality orientation.

Public service media remit reflected in citizens' expectations of performance and stakeholder dialog

PSMs in democratic societies enable public deliberation through trusted information (Campos-Rueda and Goyanes, 2022). This remit necessitates recognizing citizens' democratic needs when providing services and interacting with them (Spigelman, 2014).

PSM legitimacy relates to the public interest. Public interest can evolve due to citizens' changing expectations of, for example, PSM access and content in the digital transformation context (Sehl, 2020). Changes in citizens' expectations may be gradual. However, establishing a general definition of public interest is difficult. This study presumes that public interest is derived from PSMs' remit: citizens' expectations to be educated, informed and entertained (Campos-Rueda and Goyanes, 2022; Glässgen, 2015; Spigelman, 2014).

Their remit suggests engaging with citizens in dialog about expectations. Effective dialog practices, which constitute relationships, fall within an organization’s social responsibility (O’Riordan and Fairbrass, 2008, 2014), especially for PSMs responsible for delivering the program mandate. From this perspective, SD should mutually benefit organizers and citizens. We assume that mutual benefits arise when shared expectations for SD are met: citizens unconditionally voice and deliberate on their expectations of PSM performance, and PSMs enhance trust by gathering a nuanced assessment of their performance to inform strategic planning (O’Riordan and Fairbrass, 2014).

Effective dialog requires citizens and PSMs to agree that SD practices support the notion that “an informed citizen is the cornerstone of democracy, and the ultimate reason public broadcasting exists” (Campos-Rueda and Goyanes, 2022, p. 2498). SD becomes a civic engagement venue if actors deliberate on PSM performance within citizens' diverse social contexts (see Sehl et al., 2022). SD theoretically scales up to the institutionalization of values in society (Richter and Dow, 2017). Thus, organizers' practices should both mirror the PSM’s remit when engaging citizens in dialog and expect SD outcomes to flow into public discourse. These outcomes represent the meaning systems that emerge from networked conversations in between organizations and the public.

SD research points to actors' interests in and perceptions of the societal issue to be dealt with (Brand et al., 2020), which affect their orientation toward dissent and conflict. PSM legitimacy rests on fulfilling the remit and meeting citizens' expectations (Campos-Rueda and Goyanes, 2022). Media perception corresponds with expectations: what is expected and defined as desirable and what is considered essential ultimately determines the evaluation of things, Campos-Rueda and Goyanes (2022) explain. Organizers, therefore, must consider citizens' perceptions of PSM performance in SD planning and implementation: Sehl et al. (2022) report that Germans had high trust in the country’s PSMs (at a time before the SD had occurred), yet a significant proportion perceived that established media inadequately represent their immediate social environment or take seriously the topics that were critical to them. Since PSMs act in the public interest, citizens' perceived adequate representation of their lifeworlds and topics are critical to their evaluation of PSM performance and, thus, legitimization for society. Accordingly, the PSM’s SD is an opportunity for citizens to clarify their concerns about PSM performance. Citizens are assumed to presuppose organizers' practices to reflect the PSM’s remit. The findings shared by Sehl et al. (2022), thus, can sensitize organizers to enact practices attuned to exposing citizens' perceptions and issues, and, hence, create dissent oriented toward transformative relationships (Van Burgsteden et al., 2022).

Empirical research

Research questions

The benefits of SD arise from actors' taking and enacting risks. SD’s discursive processes, however, are pervaded by practices aimed at managing dialog’s contingencies, as empirical research suggests. Rather, strategic communication within SD is a discursive practice aimed at facilitating organization-stakeholder interaction. Organizers shape these interactions (Van Burgsteden et al., 2022). Therefore, they are responsible for purposefully enacting practices inclined to accomplishing SD’s complex task of establishing and maintaining actors' mutual orientation through direct and mediated conversation. Consequently, this study explores the entanglement of actors' practices, mutuality orientation and the enactment of risks while SD takes place.

Analysis must exceed the local level of conversation if SD encompasses verbal and mediated communication, with conversation analysis and corpus linguistic discourse analysis tailored to the different interactional logics of public meetings and written forums. From the language use, the two methods determine how meaning and actor images (Dreesen et al., 2023) are created in the communication process: hence, they are appropriate for interrogating “constitutive processes” between an organization and stakeholders (Van Ruler, 2018, pp. 374–375). The triangulation of these types of data and methods aims to capture the complex occurrence of risk from the respective perspectives of the data and method (Flick, 2004).

The present study’s main research question asks which risk indicators of the organization-citizen relationship can be identified in the linguistic data from conversations and written discourse in the German PSM’s SD. We derive two subquestions:

RQ1.

How can the enactment of SD’s risks be operationalized using conversation analysis and corpus linguistic discourse analysis?

RQ2.

How can both methods be triangulated to investigate SD from the perspective of strategic communication in its “discursive” sense?

Data

Data emerged from three stages of the SD: (1) an online meeting in May 2021 with 139 randomly selected citizens, (2) a public online forum where organizers invited commentary on its performance from June 27 to July 31, 2021, and (3) a second online meeting with 91 citizens in November 2021.

Online meetings

All video-recorded material from the May and November 2021 online meetings (approximately six hours) was analyzed. The May 2021 meeting started with plenary contributions from the PSM’s then managing director, its then communication director and an agency representative (additional moderator), followed by three moderated breakout sessions in which citizens discussed their current perceptions of the PSM and their expectations of its future performance. Breakout session results were reported to the plenary. In the November 2021 meeting, the PSM reported the results of the online forum and took a stance on its future priorities based on the May 2021 meeting and forum. Citizens discussed the PSM’s articulated priorities and their perceptions of the SD process in two additional breakout sessions.

Online forum

The online forum contained 14,775 posts (807,226 words), structured into seven thematic subforums on which citizens could post. The subforums' themes were derived from the citizens' statements in the public online meeting. In addition to citizens (n = 2,452), PSM representatives used the forum as topic sponsors (n = 16), forum moderators (n = 19), and PSM team roles (n = 4). While most posts originated from citizens, some subforums had a higher proportion of moderator and topic sponsor posts (see Table 1).

Methods

Conversation analysis

Actors' verbal practices in the meetings were examined using conversation analysis to clarify the social situation dynamics in the here-and-now of talk. According to the principle that “overwhelmingly, one party talks at a time” (Sacks et al., 1974, p. 700), speakers coordinate their utterances (turns) so that these turns refer to each other and are sequentially ordered. The method’s sequential perspective reconstructs actors' recursivity (Krippendorff, 1994), exhibiting practices through which they raise and negotiate the significance of topics and orient to each other’s identities (Pomerantz and Fehr, 1997; Schegloff, 1992).

Levels of analysis include how turns are distributed among organizers and citizens, and ways in which actors explicate and respond to topics and identities, hence constitute mutuality orientation through their conversation. The method’s capacity to reconstruct the emergence of activities and the consequentiality of these activities in subsequent turns (Vasilyeva et al., 2020) allows examination of the qualities of organizers' and actors' enacting risks through their conversation. The transcription was based on GAT 2 transcription system (Selting et al., 2009).

The analysis adopted Pomerantz and Fehr’s (1997) approach, describing interactional characteristics of sequences. A sequential analysis of the material aimed to investigate organizers' and citizens' activities was performed first. Then one researcher and two annotators independently searched for sequences corresponding to this study’s analysis levels. This study focuses on practices typically discernible on the interaction’s surface, facilitating case identification and analysis. The analyses were discussed within the research team to understand jointly how cases for detailed analyses were identified and the results were obtained.

Corpus linguistic discourse analysis

Corpus linguistic discourse analysis identifies patterns of language use in large text corpora and quantitatively analyzes the linguistic surface of statements (Bubenhofer, 2009). In the first analysis, we applied latent Dirichlet allocation, a probabilistic machine learning approach that identifies unknown thematic structures in large text collections (Blei et al., 2003). This approach belongs to the topic modeling family of algorithms and operates on the distribution of words in documents (in this study, forum posts). It produces ordered lists of frequently co-occurring words characterized by thematic coherence (topics). For example, a topic calculated for the forum contains deutsch (“German”), film (“film”), sprache (“language”), verstehen (“to understand”), and englisch (“English”). These “top words” indicate that the posts discuss the audio language used for films. The word lists were interpreted text-linguistically to ensure “aboutness” topics were included in our analysis (Van Dijk, 1980), that is, thematic references constituted in content words which have a semantic meaning (e.g., German, film). Analysis excluded topics with vocabulary typical for interaction in forums (e.g., greetings and thank you phrases).

The topics ascertained from the top words show alignment or divergence of citizens' topical structures with the organizers' predefined topical structure. We calculated 25 topics with the R toolkit mallet (Magnusson and Mimno, 2013), resulting in the distribution of topics over individual posts; this allowed us to aggregate the topic modeling results over the available metadata (specifically, topic distribution per subforum and user group). The examination of topics by user groups can indicate, for example, whether organizers' topics have disconnected from citizens' topics, or organizers introduce topics unrelated to citizens' posts, hence mutuality orientation at the topic level is limited.

Second, all posts were queried for the construction of type X als Y (“X as Y”), with X representing a personal/indefinite pronoun and Y representing a noun (e.g., ich als Zuschauer [ “me as a viewer” ], ich als Mama [ “me as a mother” ]. Unlike topic modeling, this approach is deductive as the corpus was queried for a specific German construction used to explicitly express citizens' identities (understood as roles in which speakers see themselves, Hyland, 2008). X as Y constructs were coded to identify verbalized concepts of social groups by asking wh-questions (e. g. who is speaking? What role is being demonstrated?). These concepts were then sorted (axial coding) to obtain coherent categories, e.g. “profession” (Glaser and Strauss, 1967).

The statistics package R was used for the illustrations (R Core Team, 2022).

Results

Conversation analysis results

We examined how speakers coordinate actions (turn-taking), which constitutes dialog and mutuality. Further, analysis queried the practices through which they raised topics and contextualized their identities, and demonstrated mutual orientation to these topics and identities. As facilitated interaction, the course of the meetings limited the recursivity of speaker turns (i.e., actors' utterances) through moderators. The moderators allocate turns to individuals by role, yet are responsible to ensure the meeting’s deliberative quality when facilitating interaction.

Organizers applied categories (identities) to citizens (such as shareholders, see Figure 1, line 9) and clients (see Figure 2, lines 7–8), thereby limiting citizens' self-defined identities in the dialog. Additionally, organizers did not invite citizens to negotiate the identities citizens ascribed to the PSM (such as fourth estate; see Figure 3, line 4) or the categories citizens used for themselves (such as client, see Figure 4, line 5). Citizens' varying stakes in the PSM’s activities were inherent in the identities used by the actors. The stakes these categories imply had no procedural consequences during the interaction: neither organizers nor citizens discussed these categories' consequentialities for citizens' stakes in subsequent turns.

Plenary session moderators ratified citizens' critiques and comments but did not directly respond to or elaborate on their content. The course of interaction did not revolve around actors' contributions as actors' reciprocity was suspended. Figure 5 shows the moderator responding to a citizen who reported results from a breakout session to the plenary. The report includes criticism of the PSM for not having included enough young people in the sample due to using phone recruitment. Rather than prompting an explication of the citizen’s concern, the moderator routinely moved forward on the agenda and recognized the next citizen.

Corpus linguistic discourse analysis results

We first examined actors' mutuality orientation during the five weeks the forum was in operation by analyzing the mean topic proportion per subforum. While a series of topics had a clear preference for a specific subforum (e.g., topics 15: content and platforms, 16: mediathek, i.e., the PSM’s media library, correlating with subforum 6: media and audio library), thus corresponding with the predefined structure, topics 21 (gender-neutral language) and 12 (climate change) were linked to multiple subforums (Figure 6). Gender-neutral language most frequently occurred in subforum 2 (people and opinion), presumably due to the subforum’s thematic openness. Gender-neutral language also occurred frequently in subforum 4 (region and lifestyle), subforum 7 (I am also concerned about) and subforum 1 (generation future). Similarly, climate change was linked to subforums 1 (generation future), 3 (knowledge and context), and 5 (program ideas). Thus, both topics eluded the given thematic structure.

Second, we investigated the mean topic proportion per user group to examine risk indicators at the level of mutuality orientation in and between different user groups (Figure 7). This inquiry indicated which topics were raised by the organizer's user groups (PSM representatives, moderators) and which were raised by the citizens. Topics linked to a specific user group included, for example, topic 20 (greetings and thank-you phrases), mainly used by the PSM team and topic sponsors, and topic 8 (moderation, contribution), which emerged through moderators' posts in response to other users' questions or posts that violated the forum rules.

The results indicated two contrasting topic types. Citizens primarily addressed gender-neutral language and climate change, unlike topic 23 (the PSM’s societal mission) and topic 15 (platforms and content), which were introduced by the PSM’s team and topic sponsors, to which citizens responded infrequently. The topics gender-neutral language and climate change were at odds with the predefined thematic structure.

Finally, we examined the discursive formation of identities. Such formations may indicate, for example, actors' mutual acknowledgments of their identities in the aggregation. We identified 152 X als Y (“X as Y”) constructions. Consider examples (1)–(4):

1. […] Kann da nur als blinder Nutzer für den entsprechenden Personenkreis sprechen […] (“I can only speak as a blind user for the corresponding group of people”) (2932_0)

2. […] aber ich werde als Zuschauer nicht persönlich begrüsst […] (“but I am not greeted personally as a viewer”) (1218_2459)

3. Ich als Mann führe unseren Familien-Haushalt zu etwa 75–80% […] (“I, as a man, run our family household about 75–80% of the time”) (3746_10596)

4. Man kommt sich als Zuschauer behandelt wie ein geistig Behinderter vor (“As a viewer, you feel treated like a mentally handicapped person”) (3340_9455)

Two research team annotators independently manually coded these 152 instances to identify underlying concepts of identity (see the quantification of categories in Figure 8). Consider examples (2) and (4), coded with the category “audience,” (1) coded with “minority,” and (3) with “gender.” Since this was a qualitative coding process, categories were not always clear-cut.

The most frequently occurring category was “audience,” which subsumed general and specific terms for viewers (e.g., Zuschauer “viewer,” Tagesschau-Seherin “viewer of tagesschau“ [the name of a German television news service]), readers (e.g., Videotextleser “Teletext reader”), listeners (e.g., Zuhörer “listener,” Rundfunkteilnehmer “Broadcast subscriber”), and (paying) customers (e.g., Kunde “customer,” zahlender Kunde “paying customer,” Beitragszahlerin “contributor”). The categories were connected with expressions of perceived exclusion, that is, citizens did not feel acknowledged as viewers, readers or listeners, and felt excluded from program decisions (ex. [2] and [4]). The second most frequent category subsumed terms with political semantics: users referred to themselves as Bürger (“citizens”), Gesellschaft (“society”) and members of a nation (Deutsche “German,” Europäer “European”); located themselves within the political spectrum (Demokrat “Democrat,” rechts konservativer Mensch “right conservative person”); and referred to the former division of Germany (West-Kind “child of the West,” ehemaliger DDR-Bürger [Deutsche Demokratische Republik] “former GDR citizen” [German Democratic Republic]). This was followed by the category “profession” (e.g., gelernter Mediengestalter “trained media designer,” Journalist “journalist”, Lehrer “teacher”), and specific academic backgrounds (e.g., Anwalt “lawyer,” Sprachwissenschaftlerin “linguist”).

Discussion

Risk indicators in conversation analysis data

Conversation analysis revealed two risk indicators: (1) organizers' lack of employment of deliberative dialog principles when suspending reciprocity and not prompting elaboration on citizens' critiques in the breakout session results (Figure 5), and (2) actors' inaccurate acknowledgment of citizens' legitimate stakes in the PSM’s activities conveyed through the actors' articulations of identities, for example, the organizer addressing citizens as shareholders (Figure 1) and clients (Figure 2).

Risk indicators (1) and (2) stemmed from their potential to assume relevance for the course of interaction. (1) may increase the risk of citizens sticking to their positions and of leaving conflict unclarified. (2) harbors the risk of actors' false consensus (Umansky and Fuhrberg, 2018), that is, mutual agreement on the stakes based on actors' inaccurate understanding of these stakes, as the categories and their implied meaning for citizens' stakes had no consequentiality during the interaction. False consensus can lead to actors' joint agreement on an action program that, in retrospect, turns out not to meet citizens' legitimate interests, hence questioning the validity of the SD.

The two indicators are intrinsically linked to each other: actors' accurate acknowledgment of their stakes and the PSM’s accountability toward citizens is a requirement to create a shared view on PSM performance. Consider the conceptual differences, for example, between shareholders and clients. Identities can be ratified or rejected in subsequent turns. Identities determine mutuality orientation in dialog, in other words, organizers' acknowledgment and clarification of citizens' stakes in its development and dialog guided by the deliberation standard that acknowledges these stakes.Table 2

The organizers' categories (Figures 1 and 2) and not inviting citizens' elaboration on their categories (Figures 3 and 4) display the organizers' lack of orientation toward “transformative engagement” (Van Burgsteden et al., 2022, pp. 65–67), that is, dialogical situations that facilitate the actors' mutual enlightenment with their views, and “communal orientation” (Lane, 2018, p. 660), according to which organizers display their understanding of citizens as active agents in the PSM’s development. Dialogic risk emerging from the contingencies of the conversation (Kent and Taylor, 2002), with citizens critiquing organizational activities or the organizers' SD practice, is illustrated in the example of a citizen’s comment regarding the organizers' sample (Figure 5). The comment addressed the deliberation standard of inclusion (Gilbert et al., 2023), yet was not brought to a close by the moderator, which would have allowed the actors to jointly clarify the concern and its consequentiality for the SD. Actors' practices of bringing forward and dealing with critical topics corresponded to the risk of the breakdown of dialog on this particular topic.

Risk indicators in the discourse data

The analysis identified three risk indicators relating to organizer-citizen alignment:

1. Autonomous discursive formations, specifically topics emerging counter to predefined subforums, indicated a shift in organizer-citizen mutuality orientation, as these formations defied the subforums' order structure and were additionally driven by dynamics according to the forum’s interactional logic (Figure 6). Organizers can, of course, only partially preset and control the topical structure due to the platform’s interactive dynamics, which allow for autonomous discursive formations, yet the design of forums determines the deliberative quality therein (Gilbert et al., 2023). Further, though the topical structure was defined according to the results of the first meeting, discursive formations suggest citizens' partial deviations from this structure. Topics 21, gender-neutral language and 12, climate change, may compromise deliberative quality if they polarize opinion and citizens renounce rationality, that is, citizens do not support their claims with information sources and evidence (Gilbert et al., 2023). The data showed that long threads materialized on topics 21 and 12 solely from citizens' posts, which indicates the risk toward the emergence of discursive enclaves disconnected from the organizers' topics.

2. Topics critical for the PSM’s legitimacy (societal role) were deliberately introduced by the organizers but hardly responded to by citizens (Figure 7). Moderators' enhanced activities contradict organizers' expectations of deliberation on legitimacy-related topics, which limits the PSM’s outcomes.

3. The discursive formation of identities and organizers' failure to prompt explication compromised organizer-citizen alignment. X as Y expressions contextualized citizens' perceptions of PSM performance in relation to their biographies, which conveyed their more diverse understanding of their identities than, for example, the organizers' shareholder and client categories in the meetings (Figure 8). Media-biographical information, however, complicates organizer-citizen alignment if citizens rationalize their positions with these contextualizations. Prompting explications could have clarified how citizens related these contextualizations to their perceived interests in the PSM’s performance. We could not detect PSM responses to these contextualizations in the data.

The risk potential of autonomous discursive formations is further determined by these topics' connection to public discourse on PSMs, instead of being discursive enclaves limited to the forum. Citizens' perceptions that the PSM does not take their issues seriously (Sehl et al., 2022) may be reinforced if they experience that the organizers avoid addressing controversial topics, instead of using these topics to resolve conflicts and interests.

Method development

Conversation analysis and corpus linguistic discourse analysis are social constructivist research methods; their different roots and characteristics suggest focusing on comparable specifics of face-to-face and mediated dialog. While conversation analysis elicits the social meaning of linguistic data, corpus linguistic discourse analysis is suitable for uncovering recursive patterns of language use and their correlations without presuppositions and sequential interpretation. A conversation and an online forum could be analyzed using either method. Thus, an oral conversation could be searched for patterns of language use of potentially risk-indicating identity construction, and a forum could be searched for corresponding interactions.

Crucial to our study, however, is how the different methodological premises are used to promote further research. In our case, triangulation was used to understand the emergence of risks and their enactment along the different interactional logics of meetings and the forum. A combination of the two data types and methods allows the identification of occurrences where the practical achievement of reasoned agreement is at risk in the fleeting qualities of conversation (meetings) and mediated conversation (forum), while actors address issues critical to the organization. Conversation analysis reconstructs these risks in the interactive dynamics of conversation in turn-by-turn sequential analyses of actors' verbal practices in meetings, including actors' displays of their understanding and identities, as well as the emergence of topics that unfold over “sequences of actions” (Van Burgsteden et al., 2022, p. 68). The method exhibits actors' attempts to assume control throughout the interaction and actors' responses to such practices in subsequent turns. It discloses the emergence of the locally situated meanings of abstract concepts such as client (Vasilyeva et al., 2020, p. 44).

Corpus linguistic discourse analysis inspects the unfolding of the discursive formation of topics and identity-defining frames in asynchronous, mediated interaction toward a wider public and overhearing audience in forums. Rather than following the organizer-defined order structure, the method determines this order structure based on patterns of language use. The combination of conversation analysis and corpus linguistic discourse analysis enabled the contrasting of the PSM’s categories in meetings with the discursive formation of citizens' self-descriptions in the forum and the elicitation of the nuance of these self-descriptions (Table 2). Both methods can identify verbal and written actor practices that demonstrate the mutuality orientation or the absence thereof as actors' guiding activities.

Implications

Theoretical implications

SD engages stakeholders to deliberately question organizational activities and policies for the organization’s strategic benefit. PSM remits, in particular, place deliberative standards on their SD. The enactment of these standards determines the SD’s perceived validity among actors and the wider public. Enactment to reflect dialog principles (Kent and Taylor, 2002) can be examined in actors' practices.

Kent and Taylor (2002), thus, assist in specifying the constitutive practices of deliberation and organization-stakeholder relationships in SD (Ferraro and Beunza, 2018). Further, Kent and Taylor (2002) assume that dialog is oriented toward actors' understanding of each other’s views. Their dialogic principles operationalize deliberation’s ethical function (Gilbert et al., 2023) through, for example, the authors' claims for actors' orientation toward unconditional verbal exchanges free from social coercion and power asymmetries. Dialog is constituted in microlevel practices, the performance of which transcends actors' self-interests in SD and reflects their willingness to accommodate their views if adversarial (Brand et al., 2020).

To reiterate, organizers shape SD interaction (Van Burgsteden et al., 2022) and are, thus, required to continuously monitor the accomplishment of mutuality orientation as a prerequisite for realizing the capabilities of SD. These capabilities unfold in interaction, which places high demands on organizers' sensitivity to situation dynamics while deliberative dialog principles are implemented. Organizers' adherence to pre-planned methods to facilitate dialog, rather than adjusting their methods as SD progresses, harbors the risk of compromising deliberative quality. Further, the PSM’s remit suggests that the organizers should involve themselves in citizens' exchanges in the forum to facilitate deliberation on topics relevant to the SD’s objectives. Organizers may consider practices such as prompting the explication of controversial issues and problematizations, therefore complementing SD’s focus on reasoned agreement with agonism (Brand et al., 2020; Davidson, 2016; Weder, 2022). The initial level of trust the PSM enjoyed in the population suggested conceptualizing SD to achieve reasoned agreement. Nevertheless, this goal, as well as the practices to accomplish it, must be evaluated during SD, if, for example, issues arise in the discussions or topical enclaves emerge in forums that turn out to be disconnected from the public discourse (Gilbert et al., 2023).

Thus, strategic communication practice becomes part of deliberative processes in SD (Weder, 2022). A conceptual task involved in this role is determining how an organization’s social and political context, and the public discourse on its performance within this context, translates to the facilitation of practices aimed at accomplishing SD goals. SD planning should adapt practices to the organization’s situatedness in public discourse to overcome introspective, and hence potentially biased, perceptions of stakeholders' topics (Dreesen and Krasselt, 2021).

Therefore, practices impacting actors' relationships assume a prominent role in organizers' efforts to accomplish the ideal SD. Ferraro and Beunza (2018) propose that deliberative practices are guided by actors “reinterpreting” (p. 1188) their relationship as a dialog rather than as an adversarial exchange; actors' mutual trust, which must be established and maintained during the SD; and common ground, in the sense of actors' shared understanding of the presupposition to SD.

With the premise that deliberative practices are constituted in communication and enacted with specific means and purpose, we return to the concept of strategic communication as “the purposeful use of communication (…) to engage in conversations of strategic significance to its goals” (Zerfass et al., 2018, p. 493). Our results suggest that engaging in SD should be understood as practices that create interactions that aim to produce shared understandings of the reciprocal effects of value creation processes and organizational policies on society, relative to the organization’s specific sociopolitical public value context. This concept suggests theorizing the role of strategic communication in strategy building and rebuilding (Van Ruler, 2018, p. 379) as a totality of inducements for co-creative communication within an organization’s “social plenum” (Schatzki, 2016, pp. 32–35). Scholarship might discuss how SD’s specificity relates to the discipline’s diverse paradigms.

Practical implications

Based on the theoretical considerations and the findings from the present study, several practical implications emerge. First, organizers' need to attune their practices to stakeholders' formulations of issues and concerns and prompt explications and problematizations accordingly. Therefore, organizers must develop sensitivity to submerged formulations of stakeholders “gesturing” toward a problem instead of explicating it (Van Burgsteden et al., 2022, p. 78). Organizers, in turn, can exploit dialog’s potential to achieve transformative relations at these points in the interaction since such stakeholders' formulations indicate that the risk of dialog breakup has not yet materialized with actors having retreated to their positions (Van Burgsteden et al., 2022).

Second, actors' engaging in dialog renders them vulnerable because actors share information, concerns and opinions, giving up the power that lies within such information (Kent and Taylor, 2002, p. 28). Vulnerability normatively imposes organizers the obligation to promote same-level relationships (see Kent and Taylor, 2002, pp. 25–26). By contrast, organizers' passing over stakeholders' issues and concerns would constitute asymmetrical organizer-stakeholder relationships with stakeholders being unaware and overhearing actors using this information for their own benefit. Vulnerability, however, is also an opportunity to create trust: if organizers bring about vulnerability through their prompts, which results in actors' exposing a “delicate” issue, mutual trust emerges (Van Burgsteden et al., 2022, p. 78). Moreover, trust emerges from stakeholders' perception of SD as valid, which in turn relies on organizers' practices to constitute a “spirit of mutual equality” (Kent and Taylor, 2002, p. 25) to be brought about by such prompts. The procedural specifics of the SD examined, thus, would have suggested inviting explication that organizers deemed critical for accomplishing actors' transformative relations as a requirement for each other’s learnings (Kent and Taylor, 2002, p. 28), especially in light of the study results on PSM performance (Sehl et al., 2022).

Third, actors' common interest in the societal issue at stake predicts that they will reflect validity while risk is enacted; organizers' avoidance of exposing shortcomings in SD may deteriorate actors' trust in SD efficacy. Organizers, thus, should consider that the clarification of concerns on SD validity can be critical to maintaining actors' trustful relations. However, in this study, the suspended reciprocity in direct conversation and time constraints did not allow for the expansion of such issues (Figure 5).

Fourth, the clarification of actors' perceived stakes articulated through citizens' identity-topic contextualizations safeguards accurate mutual acknowledgments (Kent and Taylor, 2002, p. 26). Actors' inaccurate knowledge of each other’s stakes, for example, may hinder the accomplishment of a shared view on PSM legitimacy and performance and increase the risk of false agreement (Umansky and Fuhrberg, 2018).

Fifth, organizers should be knowledgeable about the public discourse on the organization’s activities and policies since clarification of stakeholders' perceptions can avoid misunderstandings, conflicts, or false consensus (Umansky and Fuhrberg, 2018). Organizers, thus, can use the results of such analysis to develop sensitivity when preparing themselves for the issues stakeholders may bring forward (see Dreesen and Krasselt, 2021). Moreover, organizers may develop practices to target false conflicts from such results, for example, by making stakeholders aware of existing conflicts, prompting their perspectives and elaboration on conflicts, and creating a shared understanding of the process of conflict resolution. Online platforms may additionally assist to elicit discourse with regard to stakeholders' discourse on the SD itself and its perceived validity.

Conclusion

This study contributes to the limited research on SD’s capabilities to emerge from organization-stakeholder interaction. The idea that SD outcomes are practical achievements of communicative interaction (Ferraro and Beunza, 2018) guided us to interrogate the ability of SD practices to enact an organization’s risks so that its remit remains citizen-driven.

Our study confirms extant studies' observations that organizers attempt to mitigate dialogical risks (e.g., Lane, 2018). Organizers suspending reciprocity in meetings and employing moderators with superior interventionist rights to act on their behalf can be seen as devices that aim to control interaction, hence mitigating risk according to organizers' understanding of it, driven by organizational self-interests.

In terms of our main research question, which asked what risk indicators regarding the organization-citizen relationship could be identified in the linguistic data on conversations and written discourse in the PSM’s SD, risk indicators arise from actor practices that display a lack of mutual orientation toward topics and identities. The conversation analysis identified these displays through sequential analysis. The corpus linguistic discourse analysis identified autonomous formations in the forum that emerged counter to the organizers' predefined topical structure from an aggregate perspective.

Regarding the first subquestion, “How can the enactment of SD’s risks be operationalized using conversation analysis and corpus linguistic discourse analysis?”, the conversation analysis results indicate the enactment of risks as interactionally achieved occasions that prevent a reasoned agreement and mutuality orientation in meetings. The discourse analysis results indicate risks as actor practices that constitute autonomous discursive formations of topics and identities in the forum. In terms of the second subquestion, “How can both methods be triangulated to investigate SD from the perspective of strategic communication in its ‘discursive’ sense?”, these methods must be triangulated to elicit mutuality orientation at the level of actors’ topics and identities with respect to the formats' interactional specifics.

One limitation of our study is that the results refer to the local level of interaction. Future research on public meetings and forums should examine actors' practices and guiding activities in comparable contexts (see Table 2) to allow for generalization. Another limitation results from the aggregate view of corpus linguistic discourse analysis, which the examination of individual posts could compensate for. Future research should also consider how organizers' perceptions of risk indicators influence their interpretation of SD outcomes or how these interpretations affect the organization’s strategy. Analyses of strategic documents could reveal the consistency of organizers' SD planning, implementation, evaluation and sensemaking.

Figures

Figure 1

Excerpt from transcript public meeting May 2021

Figure 2.

Excerpt from transcript public meeting November 2021

Figure 3.

Excerpt from transcript public meeting November 2021

Figure 4.

Excerpt from transcript breakout session May 2021

Figure 5.

Excerpt from transcript public meeting November 2021

Figure 6.

Mean document topic proportion (θ) per subforum

Figure 7.

Mean document topic proportion (θ) per user group

Figure 8.

Concepts of identities articulated by “X as Y” constructions

Table 1

Thematic subforums, user roles and postings in the PSM’s online forum

Thematic subforums	Topic sponsors' posts	Moderators' posts	Citizens' posts	Posts from PSM team	Total
1: Generation Future	55 (5.6%)	33 (3.4%)	878 (90%)	14 (1.4%)	980
2: People and Opinions	152 (4.1%)	179 (4.8%)	3381 (91%)	13 (0.4%)	3,725
3: Knowledge and context	28 (2.7%)	26 (2.5%)	972 (94%)	13 (1.3%)	1,039
4: Region and lifestyle	11 (1.4%)	18 (2.3%)	750 (95%)	9 (1.1%)	788
5: Program ideas	61(1.4%)	75 (1.7%)	4166 (95%)	62 (1.4%)	4,364
6: Media and audio library	312 (21.9%)	9 (0.6%)	1092 (77%)	10 (0.7%)	1,423
7: I am also concerned about	44 (1.8%)	103 (4.2%)	2272 (93%)	37 (1.5%)	2,456
Total	663	443	13,511	158	14,775

Source(s): Authors own creation

Table 2

Key concepts and data types of conversation analysis and corpus linguistic discourse analysis to analytically approach the enactment of risk in SD

	Analytical dimensions	Method	Data types
Conversation analysis	• Sequential organization of conversation (e. g., Sacks et al., 1974) and actors' reciprocity (Vasilyeva et al., 2020) • Context is constituted by the interplay of various situational and extrasituational parameters, e.g., coordinating activities, actors' demonstrated orientation to the social situation, and setting (e.g., Goodwin and Duranti, 1992; Schegloff, 1992) • Actors' guiding activities, e.g., their orientation to speakers' demonstrated identities and understanding of concepts (Vasilyeva et al., 2020)	Microlevel analysis of speaker turns	Transcribed conversation data
Corpus linguistic discourse analysis	• Patterns of language use to elicit actors' roles, identities and topics (Foucault, 1979; Dreesen and Krasselt, 2021) • Discursive formations and interconnections of actors' roles, positions, networks, identities and topics dynamically emerge in discourse (Dreesen et al., 2023)	Data-driven (topic modeling), data based (syntactical patterns)	Machine annotated corpus data

Source(s): Authors own creation