The Rhetoric of Solidarity: Nature and Measurement of Social Cohesion in the Self-representation of Civil Society Organizations

Scholars have called to study how social cohesion is discursively negotiated and produced in communication behavior. However, empirical evidence remains scarce. In this study, we investigate to what extent and how civil society organizations (CSOs), part of the backbone of social integration in modern democracies, make references to social cohesion in their public self-portrayals. We develop a standardized measure for content analyzing the manifestation of social cohesion along three theoretical dimensions: social relations, connectedness, and orientation towards the common good. We apply our innovative content measure to the external communication of an original sample of nearly 800 CSOs in Germany, using their websites. Subsequently, we use data from an accompanying organizational survey of these institutions to investigate whether and how certain organizational features help explain variance in social cohesion rhetoric. Findings suggest that CSOs’ external communications employ themes from all key dimensions of social cohesion, revealing a fair amount of variation on all three subdimensions and a summary index of the overall strength social cohesion rhetoric. These different emphases are contingent upon various organizational characteristics, namely the spheres in which CSOs are primarily active, their locations, and their target groups. Whereas culturally and media-oriented organizations as well as sports clubs are largely reluctant to make references to social cohesion, politically active CSOs and those addressing socially disadvantaged communities tend to push more in this direction. The latter tend to operate in more professionalized structures, indicating that referencing social cohesion legitimizes these groups’ political and social purposes in the public sphere.


Introduction
Social cohesion is essential to social integration in modern societies (Lockwood, 1999).This becomes particularly evident in times of persisting crises when societies tend to polarize via ideological sorting and further divide based on identities and beliefs, often catalyzed by digital media (Törnberg, 2022).Social cohesion and related notions like solidarity and unity are thus essential preconditions of communal life.The concept of social cohesion has gained tremendous scholarly attention in recent years (see Schiefer & Van der Noll, 2017, for a systematic review of the literature).A growing number of empirical studies have examined perceptions of social cohesion and its subdimensions in various population segments (e.g., Dickes & Valentova, 2013;Dragolov et al., 2016).More recently, scholars suggested to venture beyond social cohesion as a psychological construct and study how social cohesion is discursively produced and negotiated (Forst, 2020).The importance of studying how the public makes references to social cohesion can be illustrated when we look at the COVID-19 pandemic.The measures to contain the spread of the virus cut deeply into personal freedoms.To motivate compliance, reassure trust in the normative basis of democracy and counter divisions and political cynicism (Stöcker, 2021;Unzicker, 2022), containment policies were framed, for instance, as acts of "solidarity" and "cooperation" in the media and by supporters (e.g., Deutsche Welle, 2020;Falk, 2021).
Although it is a vital question for political and organizational communication scholars and practitioners how social cohesion is expressed, referenced, and mediated, existing research falls short of a focus on communicative processes and contents.With one notable exception (Leupold et al., 2018), there is serious lack of empirical insights into how cohesion is discursively produced and how it can be measured.For this study, we set out to explore how social cohesion is communicatively negotiated and constructed.There are several arenas where this communication may take place: For example, in the legacy media, we can observe political discourse on social cohesion (e.g., Engel & Middell, 2020).Using social media, civil society actors can directly engage and mobilize publics (e.g., Häussler, 2021).However, social cohesion is also a vital part of the debate among civil society actors.We therefore focus our analysis on the external communication of civil society organizations (CSOs), more specifically their self-portrayals, such as how they articulate their identity, visions, values, and relations to society.This presents a promising starting point for a communication-oriented study of social cohesion because CSOs are the backbone of civil society (Grande, 2021;Seligman & Csikszentmihalyi, 2000).
For this article, we ask to what extent do civil society organizations make references to social cohesion in their public self-portrayals (RQ1)?Given the heterogeneity of civil society organizations in terms of institutional structure, location, and scope, among others, we then dissect how the references to social cohesion differ between various types of civil society organizations (RQ2).Since empirical research on the topic is scarce, a major effort and contribution of this study is the systematic development of a first measures of social cohesion that translates existing conceptual work and empirical indicators to the study of text and language.Towards this aim, we first discuss the key dimensions of the concept of social cohesion, thereby focusing on the discursive configuration of those aspects.Afterwards, we translate the elements of social cohesion into concrete indicators that allow us to assess how CSOs communicate social cohesion in their external self-presentation.Using a large-scale quantitative content analysis, we describe the rhetoric of social cohesion on the websites of nearly 800 German civil society organizations, thereby also demonstrating the utility of our measure.In a final step, we combine our content-analytical data with original data from a survey of these organizations to explain how key characteristics of organized civil society may influence the rhetoric around social cohesion.

Social Cohesion and Civil Society Organizations
Social cohesion is a concept with a long tradition in the social sciences, most notably in sociology (Durkheim, 1992; see also Novy et al., 2012).The classic distinction by Georg Simmel between community (Gemeinschaft) and society (Gesellschaft) encapsulates that the social fabric of society is quite different from the simple aggregation of individuals (Schimmang, 2020;Tönnies, 2012).In political science, social cohesion typically relates to the concept of social capital (Putnam, 1995a; see also Jenson, 2010aJenson, , p. 1403) and the idea that the gathering of citizens in voluntary organization is a backbone of the political community and, eventually, democracy (Van Deth, 2008).As such, the concept comes with a strong normative orientation (Dragolov et al., 2016;Forst, 2020).
The political science literature particularly stresses the importance of civil society organizations as pillars of social cohesion.By their very nature, CSOs legitimize themselves by stressing their mission to benefit and improve society and to promote human growth (Gable & Haidt, 2005).Those organizations are what scholars refer to as "positive institutions," i.e., entities that can act as change agents to promote ideals that are instrumental for an equitable and well-functioning society (Seligman & Csikszentmihalyi, 2000), including individual values, civic engagement, social learning, and a sense of community (Seligman et al., 2009; see also Rutledge, 2020).
Given this normative orientation, it is not too far-fetched to assume that CSOs make use of referencing social cohesion when they present themselves and promote their visions and values to their members, relevant communities, and the public at large.Although the portfolio of direct communication tools of organizations has greatly diversified with the rise of social and digital media, professional websites continue to be central venues for building and maintaining relationships with members and relevant publics (Jun, 2011;Taylor et al., 2001), mobilizing volunteers (Boulianne & Steen-Johnsen, 2023;Emrich & Pierdzioch, 2016), and serving as news sources (Callison, 2003;Reber & Kim, 2006).This is partly due to the fact that websites tend to be easy to access and navigate for members, lay publics, and journalists alike (Esrock & Leichty, 2000).Studying original messages in these venues of external communication promises to give direct access to the ways in which CSOs refer to social cohesion-largely unaffected by media gatekeeping or interactions between user communities on social media (Himelboim & McCreery, 2012).

The Three Dimensions of Social Cohesion
Despite -or due to -the long tradition of scholarship, social cohesion remains a rather fuzzy concept for which manifold concept specifications and operationalizations exist (Chan et al., 2006).However, in recent years, the seminal work by Dragolov and colleagues (2016) have received a fair amount of scholarly attention in both conceptual reviews (for an authoritative analysis of the literature, see Schiefer & Van der Noll, 2017) and empirical studies (e.g., Brand et al., 2020).They understand social cohesion as a multi-layered and multidimensional construct that refers to three core aspects of community: (1) social relations between individuals, (2) a sense of connectedness, and (3) a focus on the common good.We decided to use their concept specification as a framework to develop our measure given the level of scholarly consensus around it and because a first operationalization in the wider context of mediated communication (Leupold et al., 2018) resembles their approach closely.Below, we describe the dimensions in detail and argue why and how they could be adapted to the study of language and communication.

Social Relations
For Dragolov and colleagues (2016), the first criterion for social cohesion is the existence of networks of relationships between individuals and social groups.In a cohesive society, these links are typically shaped by mutual trust and facilitated by an environment that accepts people with different value systems and lifestyles as equal members of society.By demonstrating these ideas discursively, the construction of social cohesion is also an inherently communicative process.This builds on earlier theorizing of the media's role in (ostensibly) decreasing social capital (Putnam, 1995b; see also Norris, 1996) and, more generally, the value of information as inherent to social relations (Coleman, 1990).Scholars introduced the idea of communicative social capital, arguing that societal integration and engagement result from communication-related factors (e.g., news consumption, interpersonal discussions) vis-à-vis the social ties that individuals have with their communities (Rojas et al., 2011).
While this scholarship helps to explain individual behaviors and organizational dynamics, less is known about how social relations are portrayed by CSOs themselves or in media coverage of civil society activity.For example, scholarship based on media content analysis demonstrates that local media highlight how social clubs and associations, including faith-based ones, help foster and maintain community and social cohesion in urban contexts (Leupold et al., 2018).These people-oriented aspects of social interaction make up the horizontal dimension of social cohesion.

Connectedness
The second criterion, according to Dragolov and colleagues (2016), is that individuals and groups are tied to institutions in which they are embedded.Socially cohesive environments are those in which people identify with a superordinate entity (such as a nation), have trust in its social and political institutions, and cherish a sense of fairness regarding societal conditions (Dragolov et al., 2016).This comprises social and communicative processes that help "instill in individuals the sense of belonging to the same community and the feeling that they are recognized as members of that community" (Jenson, 2010b, p. 6).Institutions thus help citizens imagine their place in that community -vis-à-vis other community members.Nationalism is commonly referred to as a way of imagining and thus creating community, most notably in the work of Benedict Anderson (2006), for whom newspapers played a key role in generating and reproducing a sense of cohesion.Not only did they focus readers on common topics relevant to the nation-an identity construct they shared -they also constituted "a ritual demonstration of a kind of belonging" (Calhoun, 2016, p. 12).Although Anderson primarily focused on national communities as imagined communities, he accentuates the underlying conditions, dynamics, and institutions that contribute to these sociocultural identification and attachment processes.His work also points to the important role of CSOs in better understanding the linkages between real and imagined communities.
The idea of institutional support in bringing together different social groups, for example, through structured programs, is one of the critical conditions under which intergroup contact can minimize prejudice and foster prosocial outcomes (Allport, 1954;Pettigrew & Tropp, 2006).CSOs provide institutionalized opportunities for involvement, thus enabling social exchange and building social capital (Van Deth et al., 2016; see also Foley & Edwards, 1998;Putnam, 1993).When it comes to mediatized discourses, institutional or political trust are common themes in local communication.It shows the myriad ways in which people interact with (public) institutions, ranging from open days to visit the local fire brigade or police station to the malfunctioning of the local government (Leupold et al., 2018).Connectedness with institutions becomes also apparent in discourses around themes of equity and fairness (Augoustinos & Callaghan, 2019) as well as in identity-shaping and community-building events, rituals, symbols, and memories (Okamoto & Ebert, 2016;Wodak, 2009).The rather institution-oriented (vertical) dimension of social cohesion thus complements the more people-oriented (horizontal) dimension mentioned above.

Common Good Orientation
Dragolov and colleagues' (2016) third component of social cohesion is a focus on the common good.This is typically described as showing solidarity, being active citizens in their communities and showing sufficient respect for established social rules and conventions that govern human interaction (Dragolov et al., 2016).Communicative actions and the rhetoric around those notions have gained some attention from scholars.At the core of this literature are typically crisis and disaster narratives of suffering, hardship, and vulnerabilities that are meant to evoke expressions and acts of solidarity (e.g., Adger et al., 2017;Chouliaraki & Zaborowski, 2017).
Like political actors seeking legitimacy for policy interventions, civil society actors are guided by moral reasoning to justify their actions (Graham et al., 2011).While solidarity is tied to vulnerability perceptions, respect for social rules aligns with a system-based normative position (Dow et al., 2006).This view focuses on authorities (e.g., the state) and the trust in them to act in the public interest (Adger et al., 2017).A socially cohesive community is also constituted by active citizens who contribute to the common good through participating in social and political life, for example, via public debates, elections, and civic activism (Dragolov et al., 2016;Leupold et al., 2018).Such narratives typically revolve around political action and social movement activity as components of civil society discourse (Fowler & Biekart, 2011; see also Hickey, 2009).Likewise, civil society provides civic opportunities for people to encounter and experience collective life, political action, and the common good (De Vries, Kim, & Han, 2022).These activities lend themselves to be part of a broader civil society discourse that seeks to foster or, in some cases, challenge social cohesion.

Method
Our study employed two data sources: an organizational survey of CSOs in Germany conducted in 2020 (Hutter et al., 2021(Hutter et al., , 2023) ) and a systematic content analysis of the website profiles of (the majority of) these CSOs.These data enabled us to describe how CSOs make references to social cohesion in their self-portrayals and how this type of rhetoric around social cohesion is conditioned by their organizational characteristics.

Sample
Our original organizational survey of CSOs was conducted in 2020.It combines a broad range of institutions ranging from rural to urban places and covers various social, political, and cultural domains, including sports, culture, education, and the environment (Hutter et al., 2021).To ensure geographical representation, CSOs from 55 places were selected.These places included 13 smaller towns, 13 mid-sized cities, and 13 larger cities in each state (Bundesland) in Germany, and in addition, CSOs from all 13 state capitals as well as the three city states Berlin, Hamburg, and Bremen (for details regarding sampling and response rates, see Hutter et al., 2021Hutter et al., , 2023)).For each location, a list of CSOs was generated that builds on a comprehensive search for formally registered clubs and associations (eingetragene Vereine) in Germany's commercial register (Handelsregister) (e.g., charities, foundations, neighborhood associations).Moreover, to grasp the more informal, yet ever more important part of civil society, a systematic online search of less formalized citizen initiatives complemented the search in the registry.A total of 1,066 CSOs participated in the organizational survey, accounting for a response rate of around 28%.For our content analysis, we collected a total of 789 public profiles of organizations that provided a clearly identifiable name of their organization in their survey response and had a functional website in German in early 2022, i.e., at the time the coding took place.

Coding Procedure
We content-analyzed the texts on the websites of the CSOs as this type of information provides a synopsis of how the organization views itself and its relation to the wider community.Specifically, we collected all text on the landing page and the 'About Us' section (or a comparable section and relevant subsections) because organizations most likely use these more prominent sections of the website to describe their goals, ambitions, stakeholders, and involved communities.Moreover, people interested in an organization are most likely to first access these sections of the website.At the same time, information on these pages is usually more condensed than, for example, mission statements, and more likely to be encountered and processed than, for example, an organization's charter.
The self-portrayals of the CSO were scrutinized to identify the messages and cues that lend themselves to convey notions of social cohesion.Coder training took place in early 2022 and involved three coders practicing with non-study content that was similar to the study content in structure and complexity, i.e., websites of CSOs that are not part of the final sample used for the content analysis (as recommended by Lacy et al., 2015).Altogether, the training spanned a total of twelve hours and was completed when the protocol and coders produced reliable data, and the intercoder reliability was satisfactory for exploratory research.The coding took place from February to June 2022, with each coder independently coding content.To facilitate coding, PolDem was used, a web-based tool for manual coding and annotating of texts (for more information, see Hunger et al., 2021).1 .

Measures of Social Cohesion Rhetoric
We developed several indicators to measure each of the three dimensions of the multidimensional construct of social cohesion in the CSOs' external communications.This process of operationalization was based on a combination of deductive and inductive approaches, which is considered suitable for exploratory work (Neuendorf, 2002, p. 12).Specifically, we proceeded in a systematic manner by (a) reviewing the extant literature and prior research for relevant indicators-often rooted in survey research and individual-level data (e.g., Berger-Schmitt, 2002;Dickes & Valentova, 2013;Dragolov et al., 2016; see also Schiefer & Van der Noll, 2017) -that are translatable to text analysis, and (b) a close reading of a small set of organizational websites to identify emergent themes.We defined each indicator and provided specific examples in a codebook that the author team reviewed and refined in iterative rounds.All individual indicators were coded as binary variables (0 = no reference, 1 = reference exists) and later aggregated as described below.A final intercoder reliability check was done during the coding, demonstrating high consistency of each index measure.Chance-corrected reliability coefficients, using Krippendorff's (2004) alpha, are reported below (Freelon, 2013).
Social relations.The social relations (horizontal) dimension of social cohesion comprises five items that reference markers of social identity and inclusiveness (K-α = 0.77).The items include references to the following markers: (1) origin (including culture, language, race, ethnicity, religion, nationality, citizenship status), (2) age/generation, (3) gender (including sexual orientation and use of gender-neutral language), (4) (dis)ability, and (5) social class.If a website explicitly mentions all those inclusivity aspects, it received a maximum score of 1 on the social relations variable.To illustrate, if a CSO's website explicitly articulates that the organization provides resources to migrant women and children from poverty-stricken countries it would score on four of the five markers (excluding (dis)ability).
Connectedness.The connectedness (vertical) dimension of social cohesion was recorded with four items (K-α = 0.79): A dichotomous item that gauges (1) whether the CSO views itself as part of a larger civil society network-if, for example, there were hyperlinks, textual references, visuals, or audio-visual elements that link to other organizations, citizen initiatives, political parties, or institutions with similar goals.Since connectedness also involves the "importance of feeling attached to or identify with the social entity" (Schiefer & van der Noll, 2017, p. 588), we also included explicit references to being attached to (2) a social group/identity and (3) a geographical entity (local/regional, national, or trans-/international).We also assessed (4) if the organization related to two or more geographical entities (e.g., the region and the country or a transnational entity such as the European Union).If a website explicitly references all four aspects, it received a maximum score of 1 on the connectedness variable.To illustrate, if a CSO explicitly addresses artists locally, nationally, and transnationally, and highlights its connections to other art societies and associations, it would yield a maximum score of 1.0.
Orientation toward the common good.The common good orientation was recorded on four binary items (K-α = 0.70): We marked (1) whether the CSO explicitly calls on someone's or some group's solidarity (including related notions like helpfulness and goodwill); (2) whether acts of solidarity are explicitly mentioned (e.g., helping neighbors, promoting intercultural encounters, social gatherings); (3) whether forms of civic participation are mentioned (e.g., calls to protest, voting, or approach political decisionmakers); and (4) whether social rules and norms are explicitly called upon (e.g., citing laws or legal frameworks, international or national agreements, norms ensuring human dignity or freedom of assembly).If a website meets all these criteria, it received a maximum score of 1.0 on this index.For example, if a CSO website explicitly articulates that it provides financial and legal support to refugees, calls for people's solidarity, and organizes political protest and petitions while invoking international treaties like the United Nations Convention on the Rights of the Child, it would obtain a maximum score of 1.0 on this dimension.
Strength of social cohesion rhetoric.As we consider all three dimensions of equal importance and none a sufficient condition for social cohesion, we additionally computed an additive index to measure the overall strength of social cohesion in a CSO's rhetoric (K-α = 0.84).Thus, the three index scores were summed up and averaged so that an overall score for the strengths of solidarity cues could be assigned to each organization.

Measures of Organizational Features
We retrieved several organizational features of each CSO from the organizational survey.Table 1 gives an overview of descriptive sample statistics of those CSOs whose public profiles (i.e., their self-portrayals, the unit of analysis for this study) were content analyzed.
Primary sphere of activity.The domain of the primary activity of the CSO was recorded using the categorization by Priemer and colleagues (2017).Respondents were able to select multiple choices among the 14 available options, including civil protection, public health, and education (see Appendix Table 4 for more details and examples) before choosing their organization's major domain of activity.Those were then collapsed into five categories: culture/ media, sports/leisure, education/research, social/religious services, and the environment.
Geographical location.We included information about the location (i.e., state and city) to compare the solidarity markers of CSOs between (1) former East and West Germany, (2) the capital Berlin and the rest of the country, as well as between (3) rural areas (i.e., villages and small towns) and urban spaces (i.e., mid-sized and larger cities).
Organizational type.The CSOs were differentiated into (a) registered associations with formalized structures (e.g., official membership, led by a chair or council) and (b) initiatives with more informal structures (e.g., rather loose associations of individuals).
Professionalization degree.The CSOs indicated whether they represent (a) an all-volunteer organization or (b) a better-resourced and professionalized organization that flags out paid positions.
Political activity.We recorded whether the organization views itself as a political actor (alongside other options).
Addressees.In the survey, organizations were asked about the societal groups that they address.A list of eight possible target groups (i.e., addressees) were provided: (1) people with a migration background, (2) people with shared cultural roots and/or religious beliefs, (3) particularly women or men, (4) socially disadvantaged people, (5) those in need of care, (6) students and school children, (7) families, (8) specific organizations or institutions, and others (using an open-text field).From this answer, we computed a binary measure (subsequently labeled scope of social cohesion) that distinguishes between internal and external scope.If an organization did not indicate more than one addressee (from the list), we defined this as internal scope, i.e., the organization reaches out to members only within that group or organization.For example, if a local sports club refers to solidarity concerning its immediate club members only, a narrower understanding of belonging is expressed.If an organization, however, indicated two or more addressees, we defined this as external scope, which describes references beyond the immediate ingroup, i.e., the organization reaches out to external communities and other groups within society.For instance, if a local sports club speaks of engaging in friendly matches with other clubs to raise money and support refugees from Ukraine, these references were recorded as an external form of cohesion.More than onethird of CSOs in our sample (35%, n = 270) relates to two or more groups (thus representing external cohesion) which shows a considerable level of social embeddedness, while roughly one in six (17%, n = 132) indicated one particular addressee (thus featuring internal cohesion).

Results
In the following, we first answer RQ1 by describing the prevalence of cues to social cohesion in the communication of the CSOs.We start by presenting the results for the subdimensions before combining them in our summary index that provides information on the overall strength of social cohesion.Subsequently, we answer RQ2 by studying whether and how organizational characteristics of the CSOs are associated with a more or less stark rhetorical orientation toward social cohesion.Towards this aim, we ran multivariate analyses in which we subsequently regressed our dimensions of social cohesion and the overall strength on the aforementioned organizational characteristics.

The Three Dimensions and the Overall Amount of Social Cohesion
Concerning the rhetoric regarding social relations (Mdn = 0.20, M = 0.24, SD = 0.25), roughly one-third of CSOs in our sample (34.3%, N = 271) made two or more explicit references to inclusivity.More than a quarter (28.5%, N = 225) made at least one reference (see left-hand side of Fig. 1).In comparison, slightly more than one-third of the CSOs in our sample (37%, N = 293) did not mention this dimension at all.Higher values on this index were largely driven by addressing people's age or generation, their gender or sexual orientation, their origins (including but not limited to their cultural, religious, and ethnic origins), and to a smaller degree, their social class and ability/ disability status.
A more diverse picture emerged when examining those statements that focus on identifying with and belonging to the larger polity and its institutions in which people and CSOs are embedded.The histogram in the center of Fig. 1 shows that most of the CSOs in our sample cluster around the mid-point of the index (Mdn = 0.50, M = 0.47, SD = 0.26).Higher values on this scale were mostly driven by those organizations that reference social cohesion on at least one geographical level (e.g., local or national or international) and their embeddedness in a network of other civil society actors.To a smaller extent, these scores were driven by those CSOs that identify with two or more levels of social cohesion (e.g., emphasizing cohesion at both the local and the national level) and address their attachment to a particular social group.The Bell-shaped curve suggests a relatively homogenous type of rhetoric.However, roughly one in seven CSOs either did not (8.7%,N = 69) or did fully (5.4%,N = 43) engage in this type of rhetoric.
Most variation exists regarding the third dimension, which focuses on the rhetoric that links the individual with the common good (Mdn = 0.25, M = 0.38, SD = 0.28; see righthand side of Fig. 1).While approximately one in five CSOs (21.2%, N = 167) did not make any explicit references in that regard, higher values on this index were largely driven by organizations relating to people's solidarity, helpfulness, and specific acts of solidarity, and, to a lesser degree, by invoking social rules and people's participation in politics and civic life.The distribution of values pertaining to the overall strength of social cohesion as indicated by our summary index is displayed in Fig. 2. Note that this builds equally on the three partial indexes discussed in the previous paragraphs.As the histogram shows, there was a fair amount of variance in the study sample (Mdn = 0.33, M = 0.36, SD = 0.19), which calls for taking a closer look at the underlying structural (i.e., organizational) factors to identify patterns that help explain this variation.

Differences by Organizational Characteristics
Table 2 shows the extent to which various organizational characteristics are associated with the variation in each of the three partial indexes and the summary index.With regard to activity spheres, cultural and media-oriented organizations were significantly less likely than CSOs focused on social and religious services (the reference category) to employ language of social cohesion.They were least inclined to emphasize diversity, inclusiveness, and the importance of social relations, nor did they relate to the common good as a normative framework, which also affected the overall rhetoric on social cohesion.Sports clubs were similar in that regard.In contrast, we found that environmental organizations and climate change groups bring up their institutional networks in their communication significantly more often than others.Additionally, a more fine-grained analysis revealed that organizations oriented toward transnational solidarity (e.g., humanitarian aid, fair trade, human rights), social services (e.g., food banks, shelters), and education-oriented institutions were most likely to engage in a strong social cohesion rhetoric (for a complete analysis of all 14 domains, see Table 4 in the appendix).
Under multivariate conditions, geographical location matters only to the extent that CSOs located in rural areas of Germany were less likely than their urban counterparts to engage in rhetoric around the common good, emphasizing aspects like solidarity and civic engagement. 2 There were no statistically significant differences between East and West Germany as well as between Berlin and the rest of the country. 2 Given the relatively small number of CSOs based in rural areas in our sample, further affected by listwise deletion of missing data used for the OLS regressions, separate analyses reveal statistically significant negative associations between rural location and a CSO's performance on the connectedness index (B = -0.106,SE = 0.035, p < .01), the common good index (B = -0.120,SE = 0.038, p < .01),and the overall social cohesion index (B = -0.092,SE = 0.027, p < .001).Entries shown are unstandardized coefficients from ordinary least squares (OLS) regressions.The independent (explanatory) variables (in the rows) stem from the organizational survey while the dependent variables (in the columns) represent the different indexes based on the quantitative content analysis.1-7: The reference categories are (1) social and religious services, (2) urban areas, (3) West Germany, (4) all federal states except Berlin, (5) citizen initiatives, (6) no full-time employees, (7) does not view or only partly views itself as a political actor.( 8) Since an organization can have multiple addressees (recorded via a multiple-answer question), each group was treated as a dummy variable # p < 0.10; *p < 0.05; **p < 0.01; ***p < 0.001 The organizational type, the degree of professionalization and political activity play a more important role.Registered associations (versus citizen initiatives) and more professional organizations (versus volunteer-run associations) scored higher on the social relations index and the connectedness index, but not on the common good index.CSOs that view themselves as a political actor (e.g., those providing political education, fighting against human rights abuses or engaging in environmental activism) tended to score significantly higher than those that did not consider themselves a player in the political arena on the connectedness and common good index, but not the social relations index.When it comes to the overall rhetoric around social cohesion, CSOs with full-time employees and a political mission scored significantly higher than all-volunteer and non-political organizations.
Finally, we found important differences regarding the scope and addressees of the different CSOs.Those organizations which are devoted to socially disadvantaged individuals and groups (e.g., senior citizens affected by poverty, people with disabilities) showed significantly higher scores than others on social relations, common good, and overall social cohesion.Interestingly, they did not stand out on the connectedness measure.CSOs with a migrant community focus were more pronounced regarding social relations, inclusiveness, and diversity.In contrast, CSOs specifically addressing other organizations (e.g., professional organizations, alliances of associations) tended to score lower than others on the social relations index.
The number of groups addressed in the solidarity cues tells us about the social embeddedness of a CSO and how strongly they relate to their civic environment.If messages are designed in a way that they address multiple diverse groups, one can expect a stronger performance on our index.We therefore distinguished between an internal and external scope of cohesion.Relating these two with the CSOs' index performance described in the previous section allows us to develop a typology as a first application of the index and a way to better understand differences in how CSOs communicate about social cohesion with external publics.Results are presented in Table 3.
Findings demonstrate that three in four CSOs (74.6%, n = 576) performed below the scale midpoint of the summary index (0.50), representing a more limited use of language around social cohesion; and only one in four (25.4%, n = 196) performed above that score.The pattern that emerged from the data shows that CSOs with external cohesion (i.e., targeting multiple groups) also engaged more strongly in solidarity communication.Specifically, while there were still twice as many externally oriented CSOs with weaker social cohesion rhetoric (23.3%, n = 180) than those with stronger cohesion messages (11.7%, n = 90), that pattern was slightly more pronounced for internally oriented CSOs with only one defined group (12.2% vs. 4.9%), and particularly more pronounced for those without defined groups of addressees (39.1% vs. 8.8%).

Discussion
Our study took up the claim that social cohesion is discursively negotiated in civil society (Forst, 2020).We set out to test whether we can identify elements of social cohesion in the rhetoric of external communications of CSOs in Germany.To be able to make these statements, we developed a standardized empirical measure to content analyze the manifestation of social cohesion in organizational texts.Our study makes two important contributions: First, it allows us to make statements about the nature and conditions of the rhetoric of social cohesion in civil society.Second, in doing so, we provide methodological innovation that may be extended beyond the case presented in this study.
To start, as part of their self-portrayals, the vast majority of German civil society organizations in our sample articulated a variety of elements linked to the overall concept of social cohesion.However, we also found a fair amount of variation on all three subindexes and the summary index when investigating indications of social cohesion.In a second step, we identified the extent to which different CSOs made varying use of references to social cohesion.Organizational characteristics indeed helped to explain differences in the cohesion rhetoric of German CSOs.Whereas culturally and mediaoriented organizations and sports clubs seemed somewhat more reluctant to engage in social cohesion rhetoric, politically active CSOs and those addressing socially disadvantaged communities tended to be more eager to push into this direction.These organizations also indicated that they operated in more professionalized structures, which may include paid staff responsible for membership management, outreach coordination and public relations, which in turn are crucial to conveying an ideal self-portrayal to the wider public.These groups may use the references to social cohesion to legitimize their political and social purposes publicly.
A separate analysis of the three measures revealed additional patterns that point to the relevance of location, sphere of activity, and specific addressees.For instance, groups that work for migrants stressed the social relations aspect in their public communication.Also, CSOs in the countryside in Germany seemed less likely to engage in solidarity rhetoric than CSOs in urban areas.Finally, if a CSO has more than one defined addressee -and about one-third have a strong external scope -their rhetoric around social cohesion tended to be richer compared to those groups that emphasize an internal scope or are completely focused on their own business.
For our analysis, we developed an innovative tool to measure references to social cohesion in text via content analysis.As a standardized measure that is rooted in a widely acknowledged concept specification used by Dragolov and colleagues (2016) and others (e.g., Leupold et al., 2018; see also Schiefer & Van der Noll, 2017, for a similar argument), our indicators may be applied to a variety of other communication contexts where calls for solidarity and cohesion The categories in the rows are based on data from the organizational survey while the categories in the columns are based on the content analysis, specifically the summary index of social cohesion rhetoric."None" means that a CSO did not specify a particular target group/addressee in the survey; "internal" signifies that an organization indicated one specific target group; "external" means an organization indicated more than one target group.As for the strength of overall social cohesion rhetoric, the scale's midpoint (0.5) was used to separate weaker from stronger forms of that type of rhetoric.The association between scope and strength of rhetoric is statistically significant, χ 2 ( 2 are typically found, including other organizational entities (e.g., social movements), political elites, and citizens.Such an index is not limited to organizational communication per se but lends itself to be employed to study other types of communication (e.g., social media content) and may be of interest to scholars conducting comparative research.Moreover, the measure may not only be useful in academic research, but may also be employed by communication practitioners as a self-assessment and strategic communication tool.Working with the clearly specified indicators will allow organizations to better reflect upon and adjust its messaging aimed at specific groups and society at large.For instance, organizations that engage in message-driven stakeholder engagement may check whether they use references to social cohesion (e.g., "solidarity" or "civic engagement") merely as buzzwords or whether they combine them with the mentioning of specific actions.
Whereas the measure is standardized, it allows for a certain level of flexibility in application.In research that focuses on antecedents or consequences of referencing social cohesion, the measures can be combined to an overall index of the strength of social cohesion in a message/text to allow for more parsimonious modeling.Importantly, this index will be more reliable and valid than using single-item measures.However, as we have seen, detecting variance on each of the partial indexes (social relations, connectedness, and orientation towards the common good) provides a fruitful opportunity to better understand different emphases.
Our article leaves ample opportunity for future studies into the rhetoric of social cohesion.For instance, researchers could investigate other venues of external communication (e.g., social media profiles).In doing so, we invite scholars to refine our measurement tool in a way so that they may use bipolar scales that also take into account negative values, for example, if social media debates also introduce salient elements of exclusion.In a next step, scholars could further try to link the supply of civil society organizations back to individual attitudes and behaviors.Does a stronger rhetoric of solidarity in civil society relate to more robust prosocial behaviors at the individual level, or might it instead signal the opposite?Future research on social cohesion rhetoric may also employ more qualitative approaches, including (critical) discourse and rhetorical analysis, to detect more nuanced ways in which social cohesion is negotiated in civil society.Lastly, while civil society provides a highly plausible study context where language around social cohesion is most likely to be detected, it is not the only venue for examining such rhetoric.Other research contexts may include party manifestos, political elites' speeches, and journalistic content.

Appendix
See Table 4.

Fig. 1
Fig. 1 Distribution of index scores of the three dimensions of social cohesion rhetoric across all CSOs in the sample, in % (N = 789)

Fig. 2
Fig. 2 Distribution of summary index values of the strength of social cohesion rhetoric (using 0.10 increments; min = 0; max = 1) across all civil society organizations in the sample, in % (N = 789)

Table 2
Multivariate analysis of the strength of social cohesion rhetoric

Table 3
Strength of social cohesion rhetoric by scope of social cohesion ) = 19.401,p < 0.001