Social Innovation Ignored: Framing Nonprofit Activities in European News Media

News reporting typically has a dual function: it mirrors what is going on in real life, but it also shapes how actors behave. Previous studies suggest that media presence, by way of shaping public and policy perceptions, influence how well nonprofits are able to raise funds and mobilize human resources. Yet, we are lacking insights into how the third sector is actually framed in the media, in particular with regard to innovation, which increasingly complements the more traditional functions of advocacy and service provision. To find out, we performed a longitudinal content analysis and an in-depth framing analysis on national and regional newspapers from nine European countries. The analyses demonstrate that third sector activities, especially those related to social innovation, are largely ignored. We find no systematic evidence that crises increase news attention to nonprofit activities. The third sector is becoming more newsworthy when it co-engages with government and business actors, but can benefit only little from this “positive glow”. We suggest how research on these matters can be taken forward, with a specific focus on the agenda-setting theory of mass media, the strategic management of nonprofit organizations, and collaboration in the context of social innovation.


Introduction
News reporting typically has a dual function: It mirrors what is going on in real life, but it also shapes how actors behave (Hallin and Mancini 2004). This function has received particular attention with regard to the relation between media and politics, but there is increasing interest in how media affect civil society organizations (Andrews and Caren 2010). The link between the media and the third sector is getting more attention, because scholars have found that news media played a key role in determining third sector actors' capacity to bring about broader social change (Andrews and Biggs 2016;Vliegenthart et al. 2005). This can be connected to the more recent hypothesis that media climate frameworks are central moderators of action capacity in view of societal challenges, in particular with regard to the perception of who can bring about social innovation (Krlev et al. 2014). Despite the relevance of the topic, the role of newspapers and other journalistically edited media is an understudied aspect not only of social innovation, but also of nonprofit and voluntary sector studies in general (Helmig et al. 2012).
Instead, existing research on perceptions of the civil society activities draws on citizen surveys (Bekkers and Bowman 2009;Sargeant and Lee 2004) or organizational branding studies (Venable et al. 2005), especially circulating around issues of third sector accountability and The data analysed in the article has previously been published in a working paper, Media Framing of Third Sector Activities in Europe, Deliverable 2.3 (Part 1/2015) of the ITSSOIN-project, financed by the European Commission's 7th Framework Programme, and preliminarily presented at the ISTR conference in Stockholm in 2016. trustworthiness. Consequently, current insights on the social climate frames nonprofits are embedded in are limited to self-reported public opinion. The dominant topic frames are diligence focussed (trust and accountability) rather than effect focussed (innovation). What is more, from the relatively few studies focussing on the nexus between media and the third sector we know that in marked contrast to media content related to government and business, third sector activities do not seem to be treated as equally important by journalists. They tend to act as ''superficial friends'' only (Hale 2007). All taken together suggests that third sector roles might be unduly reduced in media reporting and perceptual frames. This might have unfavourable consequences for nonprofits on the levels of policy, finance, and human resources: First, because policy makers regard newspapers, radio and television as guardians of informed citizenship and therefore derive from it what they should pay attention to and what they may ignore (Schudson 2003). Second, because perceptions shape how people engage in volunteering (de Wit et al. 2017), and third because perceptions also determine who citizens donate (see, for example, Michel and Rieunier 2012 on brand image and donations). Due to the severity of the potential consequences, the relative lack of media studies on the third sector more generally, and the timeliness of the innovation issue more specifically, we explore the overarching research question: How does media reporting frame third sector organizations, in particular as regards their capacity to produce social innovation?
We define social innovation as addressing neglected social needs by means of new approaches (Nicholls and Murdock 2012). Prominent examples of social innovation include microfinance (Phills et al. 2008), cooperative forms of organization (Mulgan 2008), or the reverted image of disability as an asset (Hockerts 2015). In terms of the media setting, we apply a comparative approach, conducting qualitative and quantitative analyses of national and regional newspapers in nine European countries. We recognize that social media such as Facebook and Twitter are becoming more and more important. Nonetheless, serious newspaper content (online as well as on paper) still plays an important, and arguably a more important, role for authoritative decision-making (Helmig et al. 2012, p. 368). The press set the agenda for public debate by highlighting different topics and indicating different degrees of priority and urgency (Andrews and Caren 2010). Even more important, the legacy press also wittingly or unwittingly communicate lack of political, economic or social importance by creating ''spirals of silence'' (Noelle-Neumann 1984) and ''spirals of cynicism'' (Cappella and Jamieson 1997). This has potential effects on resource availability and action capacity as outlined above. The level of agency and strategic action by journalists, newspapers and publishing houses as actors of high influence, power, and societal esteem, is not matched by communication on social media-or only if performed by communication by a limited number of celebrities, industry or political leaders. Thus while, social media would be an interesting and relevant angle, which we relate to in our prompts for future research, we chose to relate to institutionalized media.
The article proceeds as follows. We work out three more specific but interrelated research questions, and through agenda-setting theory, we explore them by framing analyses and quantitative content analyses, before we discuss implications of our results.

Research Questions
When presenting third sector organizations and their relation to innovation, journalists are not mere observers (see Bornstein's ''advocacy'' work, 2007). Within government and politics, the dominant players are cast in opinionated voices just as politicians seem to pursue ''mediatized politics'' based on their belief in the power of the media (Cohen et al. 2008). Other beats, e.g. in business, tend to be more neutral. Based on previous research (Thompson 2000;Taylor and Napoli 2009) we should expect newspaper coverage of third sector activities to be more like the latter than the former. ''[A]lthough non-profit coverage is generally quite favourable, it is often episodic and not thematically framed'' (Hale 2007, p. 465; see also Carroll and Hackett 2006). We relate to and extend Hale's work in relation to nine major American newspapers, by generating a European longitudinal sample of media content (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013). We analyse the tendency (positive/neutral/ negative) of the framing in the most recent year (2013 with N = 7842 news items). We further extend previous research by adding mediated social innovativeness for its timeliness and significance laid out above. This is summarized in our first research question.
RQ1 Which framing do European newspapers apply to third sector activities and how is the news framing of third sector activity related to social innovation?
In order to probe the issue of innovation, we further look at potential differences across seven primary fields of nonprofit activity (Salamon and Anheier 1992) that represent core thematic areas of the large-scale research project this study have been part of: arts and culture, social services, health, consumer protection, environmental sustainability, community development, and work integration. We also take into account that social innovation is often seen as a collaborative effort (Nicholls and Murdock 2012). This requires third sector organizations to co-produce, that is collaborate with government and business, and go beyond their typical standalone advocacy and service provision functions (Krlev et al. 2018a). 1 We would expect government and business to receive more attention than nonprofits in general. Backed by the supposition that collaborative cross-sector constellations might produce more effective solutions and thereby push the newsworthiness of stories, we would expect co-production to lever the third sector's news prominence. Therefore, we compare the trans-sectorial framing of co-production to the stand-alone roles of advocacy and service provision. We do this in order to analyse variations in mediated innovativeness and to answer our second research question below. Our analysis is again in relation to 2013 and based on a targeted subsample (N = 4187) of news items relating to the seven fields of activity listed above.
RQ2 How do European newspapers frame trans-sector co-production compared to stand-alone roles of nonprofit advocacy and service production?
In addition to the specific framing our interest is in how external events may make certain stories more ''newsworthy'' (Kepplinger and Ehmig 2006). One such pivotal event within the context of social innovation, representing a significant push in terms of social needs, is the financial crisis. After the financial crisis of the late 2000 s, the EU Commission and other European decision-makers have voiced high hope in relation to nongovernment and nonmarket actors as co-producers of social innovation (European Commission 2013; Krlev et al. forthcoming). We would expect the third sector to generate more attention, when markets and government fail, and decision-makers are under pressure to find alternative solutions to social problems, which are presented as results of the crisis by mass media (Trappel 2015). If this is true, we would expect media coverage to increase during times of crisis with catch words such as ''social innovation'' becoming mass mediated to a larger extent than during the everyday run of the mill (Scholte 2013). Within this context, we would also expect the role of the third sector, alone or within collaborative settings between sectors, to be highlighted. This leads us to our third research question which we explore in relation to the overall sample covering 2003-2013 and containing N = 108,535 items across all countries.
RQ3 How did European press coverage on third sector activities in general, and social innovation in particular, change over time?
Framing, Perceptions, and Action Capacity Third sector relevant media research is growing, but still in short supply (Helmig et al. 2012). Besides and as will be seen below few empirical analyses have been addressing nonconflict aspects of civic engagement (Malinick et al. 2013) or third sector activity that might be regarded as social innovation (Laurent-Ledru et al. 2011 on promoting vaccinations). This is despite the significant effects such reporting (or the lack of it) might have, not only in changing broader societal perceptions, but also in inhibiting the sector's action capacity. Media scholars have demonstrated that the diffusion of ideas and practices depends heavily on the media framing of news and views selected on the basis of professional criteria defined by journalists and editors (Reese 2007). Media professionals consider framing to be the essence of the craft of publishing (Rich 2007), realizing that news production is not merely a matter of reporting facts, but also a process of inter-views constructing meaning: Any analytic approach to journalism -indeed to the production of any mass-mediated content -must ask: What is the frame here? Why this frame and not another? What patterns are shared by the frames clamped over this event and the frame in different media in different places and different moments? […] And then: What difference do the frames make for the larger world? (Gitlin 1980:7) Since the 1980s, the framing paradigm has become the core of a more general agenda-setting theory (McCombs et al. 1997) and it still is, although calls for more nuanced versions of framing have become louder (Cacciatore et al. 2015). The framing paradigm demonstrates how professional perception of what is relevant news and what is not, defines not only the content of the mass media, but also frequently becomes the premise for political debate (Norris 2000). Besides, knowledge of framing processes is not only essential in order to evaluate the priority of certain types of messages, but more importantly to understand how pack journalism shapes public opinion (Mazzoleni and Schulz 1999;McCombs 2014). At the macro-level, media output may effectively be regarded as a proxy for public opinion. This is in spite of the fact that on the micro-level, most decision-makers and ordinary citizens know very well that media perceptions only prime, frame and spin selected parts of social reality (Lund 2006). On a day to day basis, however, the complexity of modern society legitimize the media to routinely translate multiplicity, e.g. third sector contributions to social innovation, into simple stories reflecting news criteria of relevance, identification, conflict, and sensation (Schudson 2003).
In most European countries, newspaper journalists working in print as well as online reporting, play a defining role within an ecology of professional media framing. In order to understand the dynamics of social domains, e.g. third sector activities, it is important to study how different fields of interest are communicated, perceived as more or less ''newsworthy'' (Galtung and Ruge 2016;Kepplinger and Ehmig 2006;Ö stgaard 1965). At first sight media coverage of third sector could be expected to be at par with the market and government. However, media research shows that while esteemed newspapers play a translating role in business and politics, this ''taken-for-grantedness'' (Marberg et al. 2016, p. 3) is not a given for nonprofit organizations. It is apparently easier to mediate public affairs related to commercial players and events taking place in institutions catering for the ballot box than covering activities of nonprofits (Curran and Seaton 2010;Andrews and Caren 2010, p. 841). We would expect if this holds more generally, it would in particular be the case with regard to the sector's involvement in social innovation.
The consequences may be significant: Agenda-setting proposes that mass media may not only determine what people think, but also what people think about (Dearing and Rogers 1996;Lund 1999) or how people think. This happens not only by way of more or less opinionated and biased gatekeeping, but also by priming (Kennedy 2008) and framing (Reese 2007) social reality. Of course the number of groups seeking news coverage continues to grow while the space and resources of newspapers do not. But or even therefore, journalists and their sources continue to be gatekeepers editing and spinning reality. Regarded from an agenda-setting perspective, the power of framing is not limited to the perceptual realm but can help understand how journalism influences what decision-makers act upon. Thereby, newspapers not only influence public opinion and decision-making by diffusing news and views, but also by ignoring claims and events not fitting into predefined frames. Significant developments in civil society may be victims of media neglect theorized as ''spirals of silence'' (Noelle-Neumann 1984). In agenda setting of this negative kind, activities may be excluded from public discourse. This may happen not because they are irrelevant to people's lives, but because they do not fit the current framework of professional journalism.
What comes in addition to this is that in routine coverage of associational life news media tend to fall back on simplified stereotypes. Most of these ''David vs. Goliath''stories (Lund 2013) are generated by press releases and other public relations initiatives by civil society organizations themselves. What is more, when voluntarism and active citizenship are dealt with more independently by critical journalists, it is usually done within a simplified conflict frame. Examples range from NGOs pressuring bad government (Cayli 2013;Taylor and Napoli 2009) or fighting big business (Scholte 2013), or spectacular coverups of scandals, where nonprofits and charities have erred and misbehaved (Allern and Pollack 2012; Thompson 2000). Although these accounts are generally legitimate as based on the ''evidence'' provided, they bear the risk of framing discourses towards a predetermined state affecting the very legitimacy of the social innovation imperative and the third sector's role in it. On the societal level this might overrule the positive framing driven by actors within the field (see Nicholls 2010). None of these antagonistic media frames invite educational stories relating to constructive aspects of how third sector organizations might contribute to societal change and renewal (Carroll and Hackett 2006). They also lead to a relative neglect of players that provide equally valuable and innovative work, which does, however, not stand out in the same way. Thus, communicating social innovation from a third sector point of view can neither be regarded as a simple top-down distribution of news nor as an open discourse with equal access.
There is previous research, which suggests ''under the radar'' activities, in particular of small and informally organized groups have significant potential and effectively contribute to social problem solving (Fechter and Schwittay 2019;McCabe and Phillimore 2017;McCabe et al. 2010). In view of their activities, more attention and scrutiny might even be harmful as it can expose seemingly uncoordinated and unprofessional activities without generating adequate benefit, which could stem from more public awareness. In a sense, such organizations seem successful, precisely because they are unnoticed. However, in view of major societal challenges and calls for collaborative action (Weber and Khademian 2008), it seems problematic when nonprofit partners receive relatively low attention. This connects to long-standing arguments that nonprofits need capacity to realize their innovative function (Alexander 1998) and for proving efficient operations and maintaining legitimacy (Prewitt et al. 2004). In the literature on cross-sector collaborations, both are said to be even more challenging, when nonprofits pair up with business (Bryson et al. 2015) as much as when they work together with or influence government (Lawrence et al. 2002). Financial resources and self-enforcing governance mechanisms are relevant antecedents for ''joint value creation'' for nonprofit organizations, but not in the same way for their partners from other sectors (Weber et al. 2017). This points at a generally higher level of efforts nonprofits need to invest to benefit from partnerships, which would likely be deteriorated by missing media attention and a lack of stimulus for public attention.
Crises in turn, we would expect, raise the profile of nonprofits. This connects to theories explaining nonprofit activity as reactions to state (Kendall and Brandsen 2009) and market failures (Rothschild 2009), both of which are likely to be exposed in and aggravated by crises, highlighting nonprofit models as an alternative. What is more, voluntary organizing and civic engagement are particularly prevalent and prominent under context conditions of crises. First, because of nonprofit initiatives' higher degrees of informality and flexibility, and second because of their prosocial vocation. Recent studies have, for example, looked at how civic initiatives in Western countries responded fast and effectively to the refugee crisis (Kornberger et al. 2018;Meyer and Simsa 2018). There is also rather longstanding documentation of the specific roles nonprofits take on in disaster relief (Simo and Bies 2007). Some even see an inherently higher potential for innovation in crisis resolution, if this is citizen-driven instead of government or business-led (Steiner et al. 2013). There is some evidence that specific nonprofits, in particular of a social movement type, are able to actively increase public awareness by strategically using social media (Fuchs 2014). Yet, we know little about whether this reflects clearly at the level of mainstream media news reporting, and whether it holds for the larger population of nonprofit organizations.

Methods
The review of agenda-setting theory in relation to nonprofit activity shows why it is important to understand mediated images of the sector in relation to social innovation. To do this we built a database of newspaper reporting from 2003 to 2013, in nine different European media systems (Hallin and Mancini 2004): the Czech Republic, Denmark, France, German, Italy, Spain, Sweden, the Netherlands, and UK. The country selection was based on the location of partners within a consortium of a large-scale collaborative research project. The consortium was composed to represent European diversity in terms of geographic regions, but also as regards variations in welfare state conceptions (Esping-Anderson 1990). Based on the database we performed the following analyses: First, guided by a joint codebook we performed a framing analysis of how the news items from 2013 within the database portrayed the third sector (negative, positive, neutral), with a particular focus on the news' link to social innovation (RQ1). We did this on a general level and in relation to seven sub-fields of activity. Second, we investigated the latter fields further in reference to whether third sector co-production with other actors made a difference in how positive and how innovative the third sector was framed (RQ2). Third, we mapped the occurrence of third sector news over the whole period 2003-2013 in a content analysis to address the prominence of reporting on the third sector and social innovation as well as shifts that have occurred over time (RQ3). Selective qualitative analyses of the news stories enabled us to understand whether the shifts we saw were related to the financial crisis.

Building the Database
A team in each of the partnering countries drew a comparative data sample from media monitors, containing third sector related content from two leading national and two leading regional newspapers, 2 judged by the public image of the outlets but also by their number of readers. The specific choices were guided by comprehensive data availability but have also been made to maximize editorial plurality of the sampled media. For example, researchers identified a spread of newspapers in terms of national geography as well as of political spectrum. The newspapers per country are summarized in Table 1. In addition to naming the newspapers and data sources, the table contains information on their ownership. The diversity of ownership displays that the newspapers we selected overall can be expected not to be systematically biased into one and the same direction in the views they feature. It also shows that despite some newspapers being owned by large media corporations cross-nationally, a significant share belongs to a publishing company, whose primary business is the newspaper itself.
Keywords sampled (in the national language) include ''Civil Society''/''Third Sector'', ''Voluntarism'', ''Social Innovation'', and ''Civic Engagement''. We cannot be absolutely sure that ''Civil Society'' in English, for example, carries the very same connotations as ''Zivilgesellschaft'' in German, ''občanská společnost'' in Czech, and ''sociedad civil'' in Spanish. In translating keywords some partners therefore conducted semi-structured interviews with stakeholders, identifying the most inclusive terms for the automated search, e.g. concepts such as ''nonprofits'' and ''associational life''. Accordingly, not all partners did proceed within the very same sequence of keywords, but comprehensive efforts were made to secure trans-national comparability.
The occurrence of keyword in selected newspapers over the years 2003-2013 was N = 108,535. In some cases two or more keywords may occur in the same article, which may distort the ''total'' sum. This possible discrepancy has been mitigated for the 2013 data, which has been used for the deeper framing analysis (7842 items for RQ1 and 4187 for RQ2).
The absolute numbers reported must be evaluated with caution, because we only capture a selective sample of all third sector activities. A catch-all sampling would have to include far more keywords, e.g. all major third sector organizations and informal actors. This, however, is a formidable task, way beyond the confines of the resources available. Items are almost equally distributed between national and regional media with some variations between countries. More than 75% of the Czech and Dutch items originate from regional media. Contrarily, 85% of the Italian items and 70% of the British items are generated from national media. These sample differences should be kept in mind when evaluating individual country contributions.

Framing analysis (addressing RQ1 &2)
Media framing, defined as ''selective views on issues (…) that construct reality in a certain way, leading to different evaluations and recommendations'' (Matthes 2012, p. 249), direct attention to or from particular aspects of a socially constructed reality. When successful, framing may impact not only normative valuations, but also cognitive scripts validating social and political facts.
Following the instructions described in a joint codebook, cross-country framing analyses were carried out on the data collected in the year 2013. Coders were instructed to exclude from the sub-sample redundant and inordinate items, e.g. a front-page note referring to an article inside the paper. After this vetting, the database for the framing analysis consists of 8463 breakdowns of stories. The first part of the analysis with regard to the framing of third sector activities in relation to social innovation was based on this whole set (RQ1). The examination of the effects of the reporting on ''co-production'' instead referred to a subsample of 4187 items (RQ2) relating explicitly to one of the mentioned seven fields of activity, deemed as central to the investigation of nonprofit activity and social innovation in the overall project this work stems from.
Partners were asked to randomly select 10% of all articles to be coded a second time by members of the team to grant for inter-coder reliability. All partners reported inter-coder inconsistency measured by means of Krippendorff's a, ranging from 0.94 in terms of mediated actors to 0.73 on organizational focus of discourse. Coders were asked to keep individual logbooks where all doubts and ambiguities would be addressed, especially ones that may give rise to biases or misinterpretations. Based on this feedback the 0.73-item (organizational focus of discourse) was excluded from the analyses presented here, i.e. all reported data are situated well above the 0.75 recommended threshold of reliability (Krippendorff 2013, p. 138).

Content Analysis (Addressing RQ3)
Editorial material (news and views) on third sector activities has been generated by quantitative searches in order to conduct a descriptive content analysis (Neuendorf 2002) on news coverage ranging from 2003 to 2013. Most of the news items focus on the micro-level in local communities, while public debates expressing more or less informed views on these issues usually take place at the macro-level of national policy making. Across these levels in order to find how third sector presence has increased or decreased in the media, and how it was influenced by the financial crisis, we tracked the occurrence of relevant news items over time. For both, the framing and content analysis, coders were additionally encouraged to take note of general trends, examples of particular interest, national peculiarities and the like to make out where the influence on news reporting came from.

Mediated Attention to Social Innovation in the Third Sector
In order to address RQ1, we combined two framing scales (positive/neutral/negative and innovative/noninnovative; see Table 2). Innovation was detected when the exact term was used, but also when an activity was described as neutrally as ''novel'' or ''unprecedented'', but also normatively ''better'', ''more effective'', etc. At first sight it might sound counter-intuitive that innovation can be framed negatively. However, in the social innovation context there are first, scholars that argue for a valueneutral concept where innovation refers mainly to new and not necessarily to ''better'' solutions (Howaldt and Schwarz 2010), and second, failed implementation of an innovative idea or counter-productive side effects may lead to a negative framing. Bike sharing offers, some civic, some commercial, might serve as a current example. While they would normally be seen as an innovative means to push alternative forms of mobility in cities, recently they mainly received news coverage in Germany when Obike, a Singapore-based company, went bankrupt and left the public with the open question of what to do with the bikes (Heuzeroth 2018). We not merely registered the specific use of the exact terms contained in the table's dimensions, but had coders evaluate the items at hand.
The table first shows that noninnovation (6476 items) is far more prevalent in media reporting than innovation (1366 items, or 17% of the overall reporting). When it comes to reporting about innovation, however, the overwhelming amount of media coverage in major European newspapers was framed positively (88%), while a more neutral stance was prevalent towards noninnovative items (52% neutral). The UK stood out in being generally less positive than the rest of the countries.
In addition to news reporting without a particular filter, we took a closer look at how innovation might be related to a number of fields within which third sector innovation may (or may not) take place. Table 3 lists these fields in the order of decreasing innovative framing. An important finding here is that innovativeness is more prevalent, when thematic frames are employed, compared to episodic framing of routine coverage where no particular field has been heralded by the newspapers (only 8% framed as innovative in the latter case coded innovative). The table also shows that the fields of community development, work integration, and healthcare form a group in which third sector activity is framed as more innovative than in other fields. The same applies to reporting where more than one field was mentioned in the news stories.

Co-Production, Advocacy, and Service Provision
Third sector activities can effectively supplement governments and market-players, especially by defining social needs and help to create trust and reciprocity, which leads to the generation of a sense of community. Our results in relation to the social innovation framing of third sector activity suggest that co-production might enhance it. We test this supposition in relation to RQ2 by generating a subsample of media content (N = 4187) consisting of thematically framed newspaper coverage of third sector actors cast in the co-production role compared to standalone roles of service provision and advocacy. Please note that more than one role may be framed in a single newspaper item, resulting in a total of 7125 observations of this kind, presented in Table 4.
It turns out that co-production is in fact mediated as more innovative (44%) than the stand-alone frames related to single third sector activities. Advocacy (34%) and service provision (33%), in particular, are mediated more noninnovative in spite of the fact that the former is supposed to be change-oriented regarded from most initiators point of view. This discrepancy of advocacy mediation was particularly prevalent in Spain and the UK. In terms of coproduction, we also find some differences between the countries coded: More than 50% of the French, German, and British items include this frame, while less than 20% of the Spanish and Italian items do so. It needs to be remarked that there was a considerable residual category (41% innovative framing) that circled around issues of volunteering, civic engagement and mobilization.
Surprisingly, mediated innovativeness, at least in the seven fields, did not necessarily translate normatively into more favourable media coverage-and vice versa. Table 5 applies the positive/neutral/negative tendency of Table 2 to third sector roles. It shows that co-production ranks four points lower than service provision, and pretty much at the same level as advocacy and other roles in terms of positive press coverage (all are located around 75%), while negative framing of the roles-and in the seven fields the data refer to here-is very rare (at 3-4%). This suggests journalists and editors may value other third sector contributions just as positively as innovativeness, e.g. voluntarism, self-actualization, and civic engagement.
Our additional qualitative analysis showed that most of the relatively few news stories explicitly addressing social innovation can be traced to press releases, websites, flyers and other materials provided by civil society organizations themselves. Especially in regional newspapers where critical gatekeeping is limited, such items do find their way to publication. Overall, however, this is not enough to link nonprofits to larger narratives and thematic frames of social innovation.

The Level of Press Coverage and the Influence of External Events
Finally and serving both as a background in terms of assessing the general prominence of press reporting and a test of the influence of external events, specifically the financial crisis, we performed a quantitative content analyses. Figure 1 shows a 29% increase in newspaper references to third sector activities from an average 2003-2013.
The timeline, however, is not strictly linear. Decreases can be found in the years 2007, 2008, 2010, and 2013,   while major leaps in this kind of media content take place in 2009 and 2011. Taken at face value this could be interpreted as supportive for a confirmation of RQ3. Statistically, however, the reported trends do neither confirm nor reject causal relationships between third sector media coverage and developments related to the economic crisis, culminating in 2008-9. Furthermore, we find marked differences in content development 2003-2013 between the nine European countries under study. A first group showed (increasing and decreasing) trends over time in response to changing government agendas but not the financial crisis. A second group showed an increase that could be tied to the financial crisis, but where we observed a difference between ''push'' and ''pull'' factors across countries. And a third group showed no trend, or did not enable a proper analysis: One set of changes over time can be connected to a shift in policy regimes or government agendas. In the Czech Republic the steady decline of third sector coverage is due to a policy shift from nonpartisan and value-based politics connected to ''civil society'' during the velvet revolution to a neo-liberal agenda after the year 2000. The steady increase in Denmark in contrast can be explained by rising levels of government and municipal concern about growing costs of welfare services, which made politicians turn more strongly to third sector actors for assistance. The positive tendency in the Netherlands stems from government action genuinely supportive of the third sector with key incidents around 2004, such as the Social Support Act and structural measures to support volunteering.
A second source of change seems in fact the financial crisis. The increase in press coverage in Italy was constituted mainly by needs driven reporting about ''volunteering'' and ''civil society'' in social services (push factor). In Germany instead, we saw more of an opportunity-driven development of reporting about the role of civil society in governing social progress (pull factor). An even stronger increase of this type could be seen in the UK. It reflects a growing political interest in third sector activities over time, sometimes called ''Big Society''-especially after the financial crisis of 2008-which, however, decreased again from 2011 onwards. Surprisingly, although Spain has been suffering widespread consequences of the financial crisis, and effects of this development were reflected in the newspaper content, e.g. calling for collective action and social innovation, there was even a quantitative decrease in third sector related items. In Sweden media do write about civil society organizations, but rarely conceptualize them as a sector as a whole. No quantitative impact on third sector newspaper content correlated with the financial crisis. In France around the middle of the period, the newspapers changed editorial policies by segmenting local coverage geographically, which resulted in a massive increase in the number of articles containing the keywords sampled. This, unfortunately, makes it impossible to isolate a potential impact of financial crisis awareness in relation to third sector activities.
On average, only 274 items were registered per newspaper per year for the period 2003-2013 on the third sector and/or social innovation. This means this kind of reporting cannot be regarded as an established field of reporting with day-to-day attention. Thereby, the specific references to ''Social Innovation'', quadrupled from 2003. This massive increase could, of course, be taken as proof of crisis impact, but the sample is very small (N = 1057), and the increase continues at the same growth rate from 2009 to 2013. A more likely explanation is the relative popularity of a rather novel term in journalism, not a one-way influence from the financial crisis on European newspaper preferences.

Discussion and Conclusions
Our main findings are as follows: When third sector engagement in social innovation is covered, it is covered within positive frames. Innovative framing is particularly prevalent within fields that might exhibit a high degree of co-production (community development and work integration as well as activities affecting multiple fields, and slightly less so health), that is settings in which no single actor or sector can manage the challenges alone. So innovation and co-production, often coinciding, may make journalists deviate from the ideals of nonengaged reporting they typically apply to the third sector. In deviation from previous research (Hale 2007), we might need to pay more attention to context and occasion specific framing practices, instead of a general attitude of journalists towards sectors. However, while the third sector's role of co-production is framed as more innovative than standalone service provision and advocacy, it is not assessed more positively. This suggests that media value other third sector roles equally highly, or that the ''positive glow'' of such reporting sticks with government or the market more than it does with the third sector. To unpack these presumptions by media studies, this suggests that in addition to studies which look at sectors in isolation (Taylor and Napoli 2009;Helmig et al. 2012), we need more research that is explicitly positioned at the intersection of actors from the different sectors.
By focusing strictly on singular events framed episodically as ''innovative'', journalists may block bridges to larger thematic frames, signalling to media users and decision-makers that the third sector in this respect is scant and of limited general importance. That a lack of communication can be harmful and directed communication beneficial cannot only be derived from framing theory. Recent research in organization studies has explored the link between communication practices and responsible action. Scholars find that organizations' communication has direct and formative effects on how organizations act (Schoeneborn et al. 2020). While Schoeneborn et al.'s view is from within organizations, it stands to reason that the way organizations are framed by the media also alters how they behave. This supposition should be investigated in more detail, and different lenses as to the behaviour being fostered could be applied. The analytic lenses we used here are clearly nonexhaustive. We focussed on service provision, advocacy, and innovation. But future studies could extend this to explore media coverage of, for example, selforganization, in situations of crisis ) but also beyond (Jang et al. 2013), or hybridized organizational practices at the crossroads of civility, commerce and policy (Lee et al. 2018). As these will be likely smaller in number than the frames we applied, locally focussed in-depth investigations are likely more powerful than quantitative analyses.
On the macro-level, we found that attention to the third sector in connection with social innovation is increasing across Europe but not in uniform ways. For example, it was surprising to see that the reporting on the third sector in the UK, where the discourse on social enterprise and social innovation is generally considered strong (Teasdale 2012), was generally less positive than in the rest of the countries. Although the assumption is typically that ''media messages are critical inputs to what becomes public discourse'' (Boykoff 2011, p. 3), this finding marks a disconnect between media reporting, public opinion, and also policy. The links are thus not always as uniform and directed as expected and the ''translation'' of discourse from one area to the other deserves more attention. Of course, this does not call into question the very foundation of framing theory (Dearing and Rogers 1996), but it suggests framing mechanisms might differ in relation to topic areas, organizational populations, or geographic contexts, considering these could bring more nuance into discourse analyses at the intersection of media, public awareness, and policy.
To our surprise, the financial crisis did not serve as a tipping force highlighting market and state failures, or promoting nonprofit models as alternative ways of organizing, or as superior problem solvers (Balassiano and Chandler 2010;Valentinov 2012). This is despite increasing evidence that crises require collaborative approaches (for example, Goldstein 2012), and the recognition of the power of civic organizing in such contexts (Kornberger et al. 2018;Meyer and Simsa 2018;Simo and Bies 2007). At the context level we might hypothesize that more concrete and immediate challenges, such as emergencies, or crises that polarize the public and media, such as the refugee crisis, have bigger effects on the attention that is given to nonprofits than crises with latent and long-term effects. At the actor level, this suggests there is a marked difference between media attention to social movementtype nonprofit action (Fuchs 2014) versus individualized action on the ground-even though local reporting should be sensitive to this. Comparative, multi-level research is needed to better understand who and what context is better able to mobilize media interest in nonprofit organizing.
On average, the third sector continues to play a niche role in news reporting. And in general, there is little controversy in news reporting about the third sector. Building on and extending our above considerations, we continue by discussing and problematizing in more detail what a lack of controversy might mean for the third sector and how that could affect its position in society.

Controversy and Relevance
The main tendency of European mediation of third sector activities is neutral to positive. On the rare occasion of derogative reporting in terms of micro-framing, the journalists focus on spectacular disclosure of fraud and hypocrisy. Such cases comprise, e.g. mission-driven volunteers not adhering to their said principles, specific cases of extraordinarily high transactions or unforeseen side effects related to charities and philanthropy. This episodic framing of spectacular scandals may momentarily turn press coverage more negative, but critical voices are rare. Based on arguments that perceptions diffused by media are crucial for third sector ability to mobilize volunteers (Helmig et al. 2012), public policy, and public support, this friendly superficiality may be a mixed blessing. Independent media not only provide regular information on voluntary and nonprofit activities. It may also contribute to improve quality of activities by naming and shaming players misusing public trust. This agenda-setting function is not fulfilled by the media in view of the third sector to date, whereas we have evidence that business misconduct is receiving more scrutiny with positive effects on actual behaviour (Jia et al. 2016).
On first glance and in the short run, positive media attention may be preferable to negative media framing (Kensicki 2004). In the longer run, to a large extent, lack of conflict-frames tends to make professional journalism ignore and neglect such practices. In this way, mass media indicate to politicians, philanthropists, and other decisionmakers that they do not need to do anything about these challenges (Koopmans 2004). This may lead to less attention from commercial and political decision-makers regarding the organizations involved as a marginal category of stakeholder with limited license to operate. And there is more to this: Evidence from the US context suggests that size and resourcefulness matter in how likely nonprofits are to get media attention, the simple formula being the bigger, the more attention (Jacobs and Glass 2002;Nah 2009). In combination with a relative lack of attention for the sector overall, this means that smaller nonprofits are likely to get no attention at all. This is juxtaposed to the perception in wider debates and publishing about what individual social entrepreneurs or social startups are capable of doing (Bornstein 2007).
Arguably, the disconnect between mediated attention and effective levels of innovation and impact achieved by the actors has contributed to the relatively slow evolution of awareness and misconception in the general public of nonprofits' role in effective social problem solving (Carson 2016). This is mirrored in academic research where ''innovation'' is still clearly an exception as an outcome in focus when studying nonprofit activities (Shier and Handy 2015). It is in rather stark contrast to how other new and business affiliated trends, such as impact investing, have been gaining traction in the last years (Barman 2015;Rangan et al. 2012). In the latter space even some individual impact investors ''[s]uddenly [found] themselves in the middle of a global media storm, lionized as saviors and vilified as greedy capitalists'' (Bugg-Levine and Emerson 2011, p. 9). And this is despite the fact that impact investing does not consist in and of itself but is meant to fuel the capacity of social problem solvers, among which we find many nonprofits or hybrid organizations (Mair and Hehenberger 2014). Thus, what is controversial and placed in some commercial setting is likely to get more attention, than what is being done on the ground to tackle salient issues. Policy makers and third sector executives would need to be aware of this potential attention trap.
Recent research suggests that social media might be an alternative way for nonprofits to gain attention (Guo and Saxton 2017), and by way of that, attract more resources. The idea behind it is that active use of social media represents a lower threshold than ''making the news''. Also, social media can be used in a more targeted fashion to reach out to beneficiaries, donors or other supporters. Social media clearly is a strategic tool for nonprofits, which they intend to use more intensely in the future (Young 2016b). A key strength nonprofits have on social media, in comparison with business or government, is their natural Voluntas (2020) 31:949-965 959 ''brand warmth'', which is more effective in gaining endorsement online than demonstrating other competencies (Bernritter et al. 2016). However, nonprofit but also other decision-makers need to be aware that ''[a]lthough the attentional dynamics of traditional and social media are correlated, evidence suggests that the rhythms of attention in each respond to a significant degree to different drummers'' (Russell Neuman et al. 2014). This means that neither is a substitute for the other. Nor is there a clear causal link of influence between mass media reporting and debates on social media. In the worst of cases, social media use by nonprofits would equal ''preaching to the converted'', while stakeholders equally important, but harder to reach directly, will remain ignorant of their contributions.
Does the Third Sector Remain a Mere Helping Hand?
Despite social innovation, third sector activities are not yet taken for granted as a professional news beat by journalists, but primarily framed episodically. When related to particular fields of activities, third sector actions are reported as somewhat more innovative, but in general, third sector activities are not reported as particularly innovation-driven.
The general trend at the macro-level is thus in line with equivalent content analyses on media data conducted on American data (Marberg et al. 2016). There could be substantial reasons for this specific framing. One reason could be that social innovation activities are marked by a high degree of local embeddedness (Bacon et al. 2008;Evers and Brandsen 2016). It could also have to do with the observation that nonprofits are critical enablers of social innovation, especially in the early stages, but often need strong partners to move the social innovation forward (Krlev et al. 2018b). Nonprofits' local restrictions, early presence and dependence on others, might make them easy to overlook by journalists. However, the media's task and purpose it to gather stories and provide insights across contexts. In particular when nonprofits are identified as innovation drivers where other actors might be unwilling or unable to engage in value-driven action (see Leca et al. 2018 on the role of nonprofits in cross-sector partnerships for work integration). As of now we lack systematic evidence to say whether journalists are missing part of the story, or whether third sector organizations are simply unable to make their innovative actions newsworthy and media relevant. Infrequent flares of media interest have been recorded, especially when politically elected officials turn to the third sector as a ''helping hand'' in the delivery of social services, healthcare, etc. This framing still dominates the media as newsworthy, although the increasing complexity and informality of government-nonprofit relationships have been at the centre of academic debate for a long time (Brinkerhoff and Brinkerhoff 2002;Gazley 2008;Gazley and Brudney 2016;Young 2016a). So, media reporting seems to lag behind in capturing this complexity. By ignoring the lion's share of voluntary and nonprofit activities, these esteemed gatekeepers may not only silence voices from the third sector, but also muffle news and views on social innovation writ large. As mentioned earlier, the role of co-production may be regarded as a mediating bridge. An increasing number of empirical findings underscore what used to be mainly a supposition, namely that social innovation necessitates action on multiple levels and multi-stakeholder engagement (Krlev et al. 2018b;Ometto et al. 2018;Phillips et al. 2017). Besides, the ''co''-prefix has different connotations than the more frequently used ''non-'' (-profit, -government, and -innovative). The practical and constructive implication from the research is that third sector executives might need to give co-production more attention, both for substantive reasons and for shaping their image. It might help them progress from being seen as a mere helping hand.

Limitations and Future Research
Despite the combination of qualitative and quantitative analysis, and the comparative nature of the research and its magnitude, we have only been able to scratch the surface of actual news stories. This did not enable us to track particular thematic areas or get a deep understanding of the trajectories and dynamics of individual innovation stories. Neither did it allow us to dive more profoundly into the causal mechanisms of changes after the financial crisis, or say with certainty that it was a relevant causal factor at all.
In an attempt to integrate our above considerations, we can derive a number of research questions for future agenda-setting research on the third sector: First, we know now that third sector coverage is of low intensity (when it comes to volume, controversy, and links to innovation alike), but the reasons still remain somewhat obscure. So, why exactly do news media neglect third sector stories when they seem to move more into the centre of societal problem solving? Has there been a shift over time, and if so has this shift been strategic or demanded by the circumstances, for instance the perceived severity of the state the world is in today? Also, do new media outlets that focus on long-term, deeper stories rather than fast daily news, such as ''The Correspondent'', 3 serve as game changers in this regard? Are these kinds of stories a better fit for capturing the complexities involved in nonprofits' stakeholder interaction and the generation of social innovation? Getting the voices of publishing houses and their executives would promise important insights. Unfortunately, we did not manage to tap those within our project. Future research should also investigate the links between policy discourse, for instance on social innovation, but also social entrepreneurship, impact investing, and SDG-related activities, and media reporting. Is the connection reciprocal or unidirectional? What are the mechanisms by which news and views are translated from one arena into the other? Do the media follow with a time lag rather than shape policy? And if so, why does the pattern deviate from how we see the media shaping policy discourse during election time, for instance (McCombs 2014)?
When it comes to nonprofit management and social innovation we should ask: What does the lack of media coverage of third sector activities more generally, and their involvement in social innovation more specifically, mean for the perception of their capacity by policy makers, business executives, and the general public? Combining media analysis and citizen survey data could be a potential strategy here, although in the latter we see a lack of dedicated investigation into social innovation too (Krlev et al. 2014). We would also suggest questions at the intersection of news media and social media, for instance: How are the audiences that the third sector can reach by either different? And which implications does that have for nonprofits' actual ability to raise funds, mobilize human resources, or gain and maintain legitimacy as social problem solvers? Can use and promotion of the one hedge deficiencies in the other, or does presence in both need to be fostered to gain exposure? A follow-up to that would be: In which cases is exposure more detrimental than beneficial, for individual organizations but also for the sector more widely? There likely is a tension between the individual level and the field level resulting from the ''selection of the best'' and a calling out of misconduct through increasing transparency and public attention. While as we argued earlier, this should be in the interest of the sector, and society, it will be to the disadvantage of some organizations. This circumstance probably results in mixed motives among organizations in striving for attention. Provided that nonprofits want to increase their news coverage, we still lack insights into what effective strategies might look like, in particular when it comes to settings of co-production. The only key pillar at present is that who does not shout out, does not get covered (Lund 2013). On such a background, it is unsurprising that business or government actors, with higher communication capacity will be better able to receive attention. If nonprofits want to get to par, it seems they would need to professionalize their communications. But how can they do so, given a persistent trend to reduce overhead costs (Berlin et al. 2017;Mitchell 2015;Parsons et al. 2017)? Our findings might provide the case for seizing to see communication as a mere add-on to effective nonprofit operations, and seizing it as strategy in its own right, a view that is more pronounced in responsible business action (Morsing and Schultz 2006). In particular, in collaborative settings which direction this should take is unclear. Should nonprofits strive for maximum distinction, should they pair up more actively with their partners in communicating joint efforts? Adding communication and media perception to the equation is likely going to further complicate the coordination of collaboration across sectors, which is already a challenging endeavour (Weber et al. 2017).
We hope our work will open up such future lines of inquiry and can serve as a reference point in terms of thematic focus, methodology, and analytical angles for investigating the framing of nonprofit activity, social innovation and collaboration across sectors.
Funding This study has received funding from the European Commission within the ITSSOIN project (613177).

Compliance with Ethical Standards
Conflict of interest The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/. Voluntas (2020) 31:949-965 961