More Bang for the Buck: Media Freedom and Organizational Strategies in the Agenda-Setting of Human Rights Groups

Abstract

Studies investigating the agenda-setting of human rights groups disagree on both their prospects of garnering political attention, and the factors that help them in that quest. This study makes the argument that we need to take account of both macro-institutional opportunity structures and actor-level strategies in order to gain a more complete understanding of the group-media interface. Specifically, it posits that the urgency of social problems only drives media attention toward groups if a country’s media landscape is sufficiently free, and that within these institutional constraints, groups themselves can enhance their media access by providing newsmakers with information subsidies. These claims are substantiated by way of a novel cross-nationally comparative data set of more than 1,000 domestic election monitoring and advocacy organizations. Findings show that media attention is structurally limited by the degree to which the news media serve as an open arena, and that even in countries with a free press, few groups achieve media access. At the same time, the most successful groups are not necessarily the most resourceful ones. Rather, strategic choices to invest in media effort, narrow policy engagement, and professionalization substitute for scarce resources, thereby giving groups “more bang for their buck.” The results clarify the causal mechanisms behind the dominance of resource-rich groups on the media agenda and reinforce calls for more globally comparative research into media agenda-setting.

Keywords:

• agenda-setting
• interest groups
• NGOs
• press freedom
• electoral integrity
• human rights

Introduction

“You can survive without direct access to legislators or the authorities. But you cannot survive without the press.”¹ This quote from the head of a Central American electoral monitoring and advocacy organization succinctly summarizes the rationale of outside lobbying, a strategy utilized by many human rights advocacy groups around the world.^2,3 Aiming to influence political outcomes, human rights activists tend to “go public,” frequently privileging media-centric strategies over direct access, and in so doing balancing their twofold goal of political influence on the one hand (e.g., expanding the conflict and convincing the public and policymakers that “their” issue is worth addressing) and organizational maintenance on the other (e.g., attracting members and funding).⁴ This attention-seeking imperative dictates that “if you are not covered by the media, you do not exist.”⁵ Yet, while media attention is vitally important to human rights advocates, it is a scarce resource (Gamson & Wolfsfeld, 1993; Jones & Baumgartner, 2005; Koopmans, 2004).

This study seeks to determine the factors that drive media attention to human rights groups. In a departure from established scholarship on the impact of human rights norms (e.g., Hafner-Burton, 2013; Keck & Sikkink, 1998; Ron, Golden, Crow, & Pandya, 2017; Simmons, 2009), I treat these actors as a sub-category of interest groups.⁶ Of course, human rights groups are unique in that they advocate for public (rather than private) interests, habitually come into direct conflict with governments, and are often connected to transnational advocacy networks and donors headquartered in the Global North. Yet there are important similarities between them and interest groups at large. They seek political influence using a repertoire of strategies, and they depend on the crucial resource of political attention. In focusing on attention-getting, this article thereby brings together disconnected scholarships on human rights, interest groups, and agenda-setting.

A number of studies investigate the drivers of media attention to groups (e.g., Andrews & Caren, 2010; Binderkrantz, 2012; Binderkrantz, Bonafont, & Halpin, 2017; Binderkrantz, Christiansen, & Pedersen, 2015; Danielian & Page, 1994; De Bruycker & Beyers, 2015; Elliott, Amenta, & Caren, 2016; Rohlinger, Kail, Taylor, & Conn, 2012; Seguin, 2016; Thrall, 2006; Thrall, Stecula, & Sweet, 2014). Yet scholars disagree on both the prospect of groups garnering attention, and the factors that help them in that quest. Some point to the remarkable agenda-building successes of public interest groups in a range of issue spaces (Baumgartner, De Boef, & Boydstun, 2008; Berry, 1999). Others, however, regard the prospect of outside voices gaining media prominence, let alone political influence, as exceedingly low (Danielian & Page, 1994; Schlozman, Verba, & Brady, 2012; Thrall, 2006). In such studies, it is the differing level of organizational resources that is seen as a key reason for skewed media access (Andrews & Caren, 2010; Binderkrantz et al., 2015; Thrall, 2006). Resourceful groups amass attention, while cash-strapped ones are left in the dust.

Despite the development of compelling arguments, a problem with both views in the existing literature is their empirical reliance on single-country case studies in established postindustrial democracies.⁷ This effectively limits the scope of the inquiry to group-level explananda, while ignoring important variation in the functioning of the public communication systems in which groups operate. What is more, the causal mechanisms behind the impact of resources on groups’ attention-getting ability—corroborated by many studies—are unsatisfactorily explained.

In this article, I posit a twofold argument: First, that macro-institutional opportunities outside groups’ control crucially structure their ability to garner media attention. Specifically, the urgency of social problems in the real world drives journalists’ attention toward human rights groups in that issue space only if a country’s news media landscape is sufficiently free. Second, within the confines of these institutional constraints, groups themselves can and do compensate for scarce resources by substituting strategies such as media involvement and narrow policy engagement, thereby maximizing the “bang for their buck.”

I empirically substantiate these claims through a case study of domestic election monitoring and advocacy organizations⁸—a subclass of human rights groups who focus on the public interest issue of electoral integrity.⁹ Electoral integrity relates explicitly to a large number of human rights treaties and standards (Davis-Roberts & Carroll, 2010), placing election monitors firmly within the study population of interest. Election watchdogs may present a “most likely” case to find effects of media freedom on attention-getting, as “their” issue is pivotal in distributing access to power and political offices. As such, outside lobbying by election monitors is likely more sensitive to variation in media freedom than that by other human rights groups.

I draw on a novel cross-national data set of 1,176 organizations across 118 countries, measuring attention to each group via news content analysis, and monitoring organizational strategies and basic characteristics through a bespoke organizational survey. Macro-institutional constraints are tracked by data on freedom of the press, while the “objective” urgency of the social problem these groups advocate on (electoral malpractice) is monitored via an expert survey (Norris, Wynter, & Grömping, 2017). For illustrations and additional evidence, I draw on interviews and background talks with election watch activists from 15 countries.

The results not only support the argument in this article, but they contribute more generally to studies of interest groups by highlighting the need for a joint appreciation of media system-level factors and group-level strategies. This calls for a much broader comparative approach to media agenda-setting. In addition, the results also shed light on the policy processes in non-democracies, suggesting that complex ways of interest intermediation may exist, even under authoritarianism. Finally, the study has ramifications for research into human rights impacts, by specifying one vector of norm-building, namely the domestic pressure group-media nexus.

The article proceeds with a section discussing different theories of groups’ media attention-getting. I develop the theoretical model of institutional constraints and informational subsidies. I then present data and methods, and discuss the empirical results. The article concludes by drawing out wider implications of the findings.

Theories of Attention-Getting

Attention is a necessary (if not sufficient) condition for achieving political influence (Jones & Baumgartner, 2005). It consists of resources (time and otherwise) actors dedicate toward discussing or thinking about an issue, and is produced and expended in different arenas (i.e., the news media, the legislature, or the executive branch; Hilgartner & Bosk, 1988). Its quantitative aspect (e.g., the number or length of news stories) makes attention a useful yardstick by which to measure agenda-building success (Green-Pedersen & Walgrave, 2014, p. 10).

But as we know, the carrying capacity of individuals and institutions is limited, making attention a scarce resource—a bottleneck through which would-be agenda-builders need to pass. For human rights groups, as for others engaged in outside lobbying, the first hurdle to pass is to gain attention in the news media (Gamson & Wolfsfeld, 1993; Koopmans, 2004). The pertinent question is therefore: Why do some groups achieve media access while others remain obscure? There are two broad approaches scholars have used to explain attention to groups.

Issue Characteristics

On the one hand, scholars contend that issue-specific factors provide the backdrop of attention dynamics in a way in which groups have little say. This might range from the competition of issues in a congested agenda space (Boydstun, 2013), or a group’s position vis-à-vis a specific issue (for or against?; De Bruycker & Beyers, 2015), to the political leanings of news outlets regarding an issue (Rohlinger et al., 2012).

Seminal work on the “issue-attention cycle” supposes that objective conditions drive the early stages of public interest in an issue (Downs, 1972). Although such attention cycles can later take on a life of their own, real-world problems and events have been recognized as driving forces of news coverage (Lawrence, 2000). Scandals, disasters, and other “focusing events” (Birkland, 1997) afford different actors opportunities to interpret and frame what occurred. When such disruptions happen, independent voices and activists may have greater chances of attracting attention because they will be sought-after sources (Bennett, Lawrence, & Livingston, 2006). For example, studies have found that the news media dedicate more attention to economic issues if there are negative economic developments (i.e., when the issue is more urgent; Behr & Iyengar, 1985). The same is true for a range of issue spaces from violent conflict (Miller & Albert, 2015), or unemployment (Soroka, 2012), to immigration (Vliegenthart & Boomgaarden, 2007). Contradictory results (Kleinnijenhuis & Rietberg, 1995) may point to the persisting challenge of measuring the “objective” urgency of a social problem. Still, if such a measurement could be devised, a relationship between issue urgency and issue attention in the media is plausible. And since issue attention increases the news value of actors that are “thematically relevant” (Wolfsfeld & Sheafer, 2006), issue urgency should also funnel attention to them.

This approach expects attention to groups to be higher when their cause resonates well with audiences, which is more likely where “their” social problem or issue is more urgent.

Resources

On the other hand, a number of studies link media access to variation in organizational characteristics, primarily groups’ resources. Many such studies would agree that attention is dramatically skewed, with only few organized interests receiving the bulk of attention while the majority remain unnoticed (Binderkrantz, 2012; Binderkrantz et al., 2017; Danielian & Page, 1994; Thrall, 2006; Thrall et al., 2014). Chief among concerns is the view that there are structural biases in news coverage that are simply too difficult to overcome for non-elite actors. “Official” voices are routinely privileged, as newsmakers tend to “index” coverage to the views expressed in mainstream government discourse (Bennett, 1990). Only groups that are well-endowed with financial and human resources manage to cut through this skew in coverage. Typically, such groups are business lobbyists or trade unions, but not human rights groups.

A number of studies have found organizational resources to be a strong predictor of media attention (Andrews & Caren, 2010; Binderkrantz et al., 2015; Thrall, 2006). As Fenton notes, “[w]ealthy organizations can inundate the media[…] while the attempts of resource poor organizations quickly become marginalized” (Fenton, 2010, p. 159). So much so, that some have called the “outside initiative” or outside lobbying a “myth” (Thrall, 2006).

This argument therefore holds that attention will flow mainly toward resource-rich groups rather than resource-poor ones, and that all else being equal, only the best-financed human rights groups will achieve media access.

Institutional Constraints and Informational Subsidies as Drivers of Attention

Being cognizant of the important contributions just outlined, I argue that limitations in the empirical and theoretical scope of these studies have led scholars to disregard relevant variation in the drivers of attention at both the contextual level and the group level.

This twofold argument builds on related accounts of the sociology of news production (Reese, 2001) and exchange perspectives of group influence (Berkhout, 2013). Reese’s (2001) “hierarchy of influences” approach stipulates that there are different layers influencing media content, encompassing (a) social systems and ideology, (b) social institutions, (c) media organizations and routines, and (d) media workers’ socialization and attitudes. The layered nature of these influences on media content is underappreciated in studies of group attention-getting. I argue that two layers in particular need to be taken into account: social institutions and media routines.

The Functioning of the Media as an Arena

Just like other political actors, human rights groups rely on the news media as an arena (van Aelst & Walgrave, 2016). However, news coverage is structured strongly by social institutions exogenous to the media and pressure groups themselves. In one of the few comparative studies of attention to groups, Binderkrantz and colleagues (2017) find variation in attention across different European media systems. Yet, their media system typology based on Hallin and Mancini (2004) might be too narrow for a truly comparative look (Stier, 2015). For the functioning of the media as an arena, de jure and de facto restrictions on press freedom are arguably more important than political parallelism and media marketization, factors emphasized by Hallin and Mancini. The exclusive focus on Organization for Economic Cooperation and Development (OECD) democracies in studies of groups and media, however, disregards relevant variation in this regard because it assumes a permeable public communication process. Thus, a broader empirical scope is needed.

State control of the media—through regulation, financing, or coercion—presents an overarching extra-media influence that impacts on the very spectrum of topics that newsmakers are able to cover. In authoritarian and hybrid polities journalists may not report critically—or not at all—on occurrences of fraud, corruption, or human rights abuses. Often, oppressive libel laws stifle critical stories (Stanig, 2015). Or worse, violence directed against journalists muzzles political reporting (Hughes et al., 2017). In the specific issue industry of electoral integrity, for instance, “endemic political pressures[…] are strongly and directly detrimental to a fair coverage of elections” (Nai, 2017, p. 192).

This logic holds that, all things being equal, the “real-world” severity of social problems is likely related to higher media attention (see earlier discussion), so that where the urgency of human rights issues is high, media attention to human rights groups will be high (H1a). At the same time, where newsmakers face restrictions in their work, they will be less willing or able to report critically, meaning that only where freedom of the press is respected, media attention to human rights groups will be high (H1b). Connecting these arguments suggests that the linear relationship between issue urgency and attention will be broken where the news media face harsh restrictions. Put differently, I expect an interactive effect of issue urgency and media freedom (i.e., where freedom of the press is respected, media attention to human rights groups will increase with issue urgency, but where there is little press freedom, this will not be the case) (H1c).¹⁰

Group Strategies as Information Subsidies

Faced with institutional constraints, can groups actively influence their media access? I contend that agency does indeed matter. Regardless of the aforementioned institutional constraints, strategic choices afford advocates with opportunities to maximize attention. Specifically, I expect that attention is related to a group’s ability to produce “information subsidies”—“[e]fforts to reduce the prices faced by others for certain information, in order to increase its consumption” (Gandy, 1982, p. 8). Certain group decisions regarding media strategy, policy engagement, and professionalization in turn structure their ability to provide information subsidies that match well with journalistic news routines.¹¹

While Reese’s layer of “routines” is incorporated in some of the literature (e.g., Tresch & Fischer, 2015), it remains underappreciated and disconnected from the institutional layer. Routines are the unwritten rules guiding newsmakers’ decisions to turn a select few of the manifold occurrences happening in the world into news stories. They are shaped by norms regarding the journalistic profession, the media’s role in politics, and the imperatives of for-profit newsmaking (Bennett, 1996). Three important routines derived from these norms are “defensive” routines such as fact-checking and objectivity (Schudson, 1978), the adherence to news values such as conflict, timeliness, or negativity (Galtung & Ruge, 1965), and economic considerations such as deadlines and resource efficiency (Tuchman, 1978). Rather than challenging such routines, human rights groups commonly accept them as given and try to emulate them as closely as possible (Powers, 2018, p. 131). As sources, they engage in a transaction with newsmakers, in which they provide information subsidies that mimic news routines in exchange for latent public opinion support, and favorable issue expansion or containment (Berkhout, 2013, p. 240). In other words, resources are only a crude proxy for the many ways in which groups can enhance their exchange relationship with newsmakers. Rather than just counting money and staff we should take a closer look at what these resources are used for (i.e., actor-specific strategies and capacities; Halpin, 2014, p. 180).

Information subsidies are effective if and when they “clone” news routines (Fenton, 2010)—that is, if they match well with the economic concerns, news values, and objectivity norms of newsmakers mentioned above. This in turn is plausibly related to groups’ decisions about the strategies and scope of their advocacy.

Groups may involve themselves in different strategies of influence, from the mobilization of members, or the direct contacting of legislators or bureaucrats, to media-centric strategies (Binderkrantz, 2008; Kollman, 1998). Insights from journalism studies suggest that groups who closely match the news routine of economic constraints (efficient news production with ever-smaller budgets, shrinking newsrooms, and accelerated news cycles) make themselves indispensable (Fenton, 2010; Powers, 2018; Waisbord, 2011). This requires frequent and proactive interactions with journalists on the part of the group, as this alleviates journalists’ time and budget constraints. The story comes to them instead of them having to chase it. Frequent contact can be fabricated through pseudo-events such as press conferences, writing op-eds, or through directly contacting newsmakers. This type of involvement—termed here “media effort”—has indeed been found to relate positively to media attention to groups (Binderkrantz & Pedersen, 2017). I therefore hypothesize that all else being equal, human rights groups that put a great deal of effort into media contacts will attract a lot of media attention (H2a).

Furthermore, I contend that newsmakers will be more amenable to information subsidies received from groups with a narrow policy engagement, rather than from policy generalists. Specialist groups develop exclusive expertise on a select few core issues, while generalists work on many different issues (Halpin & Binderkrantz, 2011). Narrow policy focus signals in-depth expertise and commitment to an issue, making it a valuable exchange good in the group-journalist relationship. Importantly, narrow policy engagement matches well with the news routine of objectivity and fact-checking, a core characteristic of which is reliance on “expert” sources (Albæk, 2011). Specialization signals expertise, and newsmakers will seek out groups whom they expect to be the most knowledgeable source on the given subject. This spares the journalist the arduous translational work of interpreting and backgrounding arcane policy issues.¹² This suggests that all else being equal, human rights groups with a narrow policy engagement will attract a lot of media attention (H2b).

In addition, well-conceived information subsidies recognize the importance of news values and package the group’s information accordingly. This is easier for more professionalized groups, as professionalization includes the creation of dedicated public relations positions (spokesperson, social media officer, graphic designer). Oftentimes, former journalists are hired, as they know precisely how to satisfy news values (Fenton, 2010; Powers, 2018). Professionalization also entails the introduction of hierarchical structures, increased division of labor, formalization of processes, a shift of professional self-identities from activists to employees, and increased demands placed on formal education credentials (Klüver & Saurugger, 2013; Maier, Meyer, & Steinbereithner, 2016). To newsmakers, this signals a “costly effort” that only a group “truly committed to social change would be willing to pay” (Gourevitch & Lake, 2012, p. 15). This may increase the credibility of that group as a source. On the other hand, the opposite may also be true. Professionalization may decrease groups’ credibility with audiences, because they are seen as inauthentic checkbook manifestations (Maloney, 2015). Still, professionalization has been linked to expanded media access of environmental groups (Andrews & Caren, 2010), as well as to heightened impact of European Union (EU) lobby groups (Klüver, 2012), international election monitors (Hyde, 2012), or Latin American social change NGOs (Waisbord, 2011). On balance, therefore, this suggests that all else being equal, human rights groups that are highly professionalized will attract a lot of media attention (H2c).

Finally, are information subsidies equally potent under different media regimes? On the one hand, news routines may be contingent on certain conditions relating back to social institutions. In particular they may be less manifest under “dictatorial conditions where[…] journalistic autonomy is not respected” (Høyer, 2005, p. 16). In this view, the limited range of “allowable” topics and the lack of economic competition among news outlets reduces the salience of news routines and thereby the power of information subsidies. On the other hand, repressive regimes—notoriously ill-informed about public sentiment or optimal policy solutions—may encourage a modicum of “consultation” or “deliberation” via the group-media interface (He & Warren, 2011; Teets, 2014), implying that news routines such as objectivity and news values still hold salience to some degree. And as media commercialization introduces economic pressures even under authoritarianism (Stockmann, 2012), efficient news production is still a prime concern. Information subsidies are per se unrelated to the content of that information, be it benign or regime threatening, which is why I expect that all else being equal, the effects of media effort (H3a), narrow policy engagement (H3b), and professionalization (H3c) hold equally in free and unfree press environments.

Evidence and Methods

I investigate the merits of these theoretical expectations by taking a globally comparative look at groups in a specific human rights issue industry—electoral integrity. Previous research has recognized election monitoring groups as important actors in the governance network “producing” electoral integrity—next to electoral authorities, political parties, the judiciary, and of course the voters themselves.¹³ Empirically, I draw on a novel global data set of domestic election monitoring initiatives (DEMIs).¹⁴ These data are based on a comprehensive mapping of citizen-based electoral observation and reform groups in all countries around the world. The DEMI data set covers 1,176 groups in 118 countries, and is used to track media attention and organizational characteristics of groups. I also develop illustrative evidence from interviews with election watch activists from 15 countries.¹⁵

Dependent Variable

I measure attention to groups as the count of news items in which a group’s name appears. To increase robustness, I use two alternative measures, one for domestic and one for international news attention, both indexed to the period two months before until two months after the last national election in the country in which each group is based.¹⁶ The domestic news attention indicator draws on daily newspapers in each country that receive everyday all-page archiving in the Factiva news archive service.¹⁷ In the 118 observed countries, a total of 449 newspapers adhere to this criterion.¹⁸ The number of sources per country range from zero to 15, with a mean of 4.5 and a median of three sources. Due to a lack of Factiva coverage, 20 countries and 226 corresponding groups are dropped from the data set.¹⁹ The measure for international news attention is based on the same time frames and search terms, but is sourced from international news wires such as Associated Press, Agence France-Presse (AFP), Xinhua, or Deutsche Presse-Agentur (dpa).²⁰ These services structure “Western” newspapers’ reporting, and may therefore tap into different aspects of attention. Only four countries need to be excluded from the data set for this variable. Due to the different measurement criteria, while the DEMI data set encompasses 1,176 groups in 118 countries, domestic news attention is only measured for 950 groups in 98 countries, and international attention is tracked for 1,109 groups in 114 countries.

Independent Variables

The role of institutional constraints—in particular the functioning of the media as an arena—is monitored via the Reporters sans frontières (RSF) Press Freedom Index, indexed to the election year.²¹ For the analysis here it is reversed and normalized to a scale of 0 to 1. For the countries under observation, media reporting is most restricted in Yemen (press freedom score of 0.00) and most free in Sweden (0.92). Tracking issue urgency requires a real-world indicator, “a variable that measures more or less objectively the degree of severity or risk of a social problem” (Dearing & Rogers, 1996, p. 28). This is achieved through the inverted 10-point electoral integrity rating (1 = no electoral malpractice; 10 = very high malpractice/failed election) of the Perceptions of Electoral Integrity (PEI) expert survey (Norris et al., 2017). This measure is indexed to the same election as the attention measure. In the countries under observation, problems of electoral malpractice were measured as least urgent in Denmark (malpractice score: 1.2) and most urgent in Burundi (9).

I derive measures for organizational strategies from a bespoke survey administered to groups in the DEMI data set (response rate 41%). All indicators are additive indices standardized to a range from 0 to 1, with higher values denoting higher media involvement, narrower policy engagement, and higher professionalization. First, the media effort index is based on how often groups publish reports and analyses, how frequently they contact journalists, and how often they write opinion pieces for the press. Second, specialization (narrowness of policy engagement) assesses how many policy areas other than electoral integrity the group is involved in, and the percentage of their resources specifically committed to work on elections. Finally, the professionalization index covers whether the group has established a number of specialized staff positions, issues a membership card, hires leadership positions on the open job market, and has fixed meeting schedules for the board.²²

Controls

To control for the possibility that organizational strategies are themselves limited by the resources a group possesses, an ordinal measure of the group’s annual budget in an election year, the number of paid full-time staff employed, and the number of volunteers working for the group are combined to a resources index ranging from 0 (minimum) to 1 (maximum). In addition, attention might result simply from having been around for a while, and the group having established itself in the policy community. I therefore control for organizational longevity, operationalized as the age of the group, combined with an ordinal measure of how many elections the initiative has observed thus far. Furthermore, groups that are internationally recognized may have better standing and therefore enjoy higher media access. This possibility is accounted for via a dummy coding membership in the Global Network of Domestic Election Monitors (GNDEM) and a number of regional networks. Group strategies might also be endogenous to attention-getting itself, because media coverage may lead to a group attracting more funds and training from donors, which then allows more expenditure on these areas. To control for a spurious relationship between groups’ capabilities and media attention, the modeling therefore accounts for lagged attention during the previous election cycle.

Analysis and Results

The analysis proceeds in three steps: First, I describe the distribution of the dependent variables at group and country level. Second, I look at the association between institutional constraints and attention, and third at the relationship between group strategies and attention, given the institutional constraints. To these ends, I use hierarchical logistic and negative binomial regression models to identify the factors explaining the crossing of the threshold between none vs. some attention and the absolute amount of attention beyond that threshold. In addition to the statistical modeling, I provide some illustrations from the interviews.

The Power Law of Media Attention

I find an exponential power-law distribution of attention, which concurs with other studies on groups’ media attention (Binderkrantz et al., 2017; Thrall, 2006). Figure 1 graphs the cumulative distribution of attention across all groups in all countries. For cross-national comparability, both dependent variables are standardized by the total number of news articles/wires covering the election in question.²³ This is necessary given that election coverage varies drastically, from only 39 news articles in Factiva covering the 2011 Liberian election to 197,556 articles on the Indian election of 2014. Controlling for this variation, the x-axis of Figure 1 therefore represents the share of attention a group receives on the media agenda, as a proportion of total election coverage. The y-axis depicts the percentage of groups remaining below a given proportion of attention.

Figure 1. Cumulative distribution of media attention at group level. Note: Proportion of attention = articles mentioning a group as percentage of all articles within ± 2 months of election, N = 950 (domestic), N = 1,109 (international).

There are very similar dynamics in both attention arenas in that 71% of groups receive not a single mention in the domestic press, and 72% remain unnoticed in international news. At the same time, there is a small percentage of extremely successful groups capturing large shares of attention. For instance, 14 organizations (1.5% of all groups) manage to surpass 5% of domestic news attention, meaning that they are named in every twentieth newspaper story about the election in their respective country. Six groups (0.5%) achieve the same in the international attention arena. The middle section of the graph suggests that the power law of attention is more pronounced in the international arena, and less pronounced in the domestic press.

There is also evidence for a similar power law of attention within countries, as Figure 2 shows. The graph plots two common indicators of divergence from normality for the distribution of media attention to groups in each of the countries under investigation. The upper pane shows the skew of media attention by country. Only about 5% of countries display a “fairly symmetrical” distribution of attention (skewness < 0.5) whereas the distribution is “highly skewed” (>2) in at least 40% to 50% of countries.²⁴ Second, the lower pane plots the cumulative distribution of kurtosis, which measures the combined sizes of the two tails of an empirical distribution. The distribution of attention indicates “peakedness” or infrequent extreme deviations (kurtosis above 1) in all but 10% of countries.

Figure 2. Cumulative distribution of skew and kurtosis of media attention at country level. Note: N = 98 (domestic), N = 114 (international).

In combination, Figures 1 and 2 demonstrate that media attention remains elusive for most organizations across the international ecology of groups as a whole, and within almost all national group ecologies.

Outside Lobbying in Free and Unfree Media Systems

Given the striking gap between successful and unsuccessful attention-getters, what factors explain a group’s ability to pass the threshold between none and some attention? And what drives the absolute amount of attention beyond that threshold? I use logistic regression models to tackle the first question (dichotomizing the dependent variables) and negative binomial count models to approach the second one. Standardized coefficients of four models are reported in Figure 3. The first two explain the crossing of the attention threshold in domestic (M1) and international (M2) news, and the other two explain the absolute amount of attention in the domestic (M3) and international (M4) news arena, respectively.²⁵

Figure 3. Explaining media attention to human rights groups. Note: M1: Dependent variable (DV) = dichotomized domestic attention (1 = yes, 0 = no); M2: DV = dichotomized international attention; M3: DV = count of domestic attention; M4 = count of international attention; M1 and M2 are generalized linear mixed-effects models (GLMM), logit link function, random effect for country; M3 and M4 are negative binomial count models, random effect for country, offset for total election-related attention. N = 188 (M1), 273 (M2), 190 (M3), 265 (M4). Variables centered and z-transformed. Depicted with 95% confidence intervals.

Turning first to institutional constraints, the results show that, other things being equal, neither issue urgency nor press freedom by themselves predict media attention. However, press freedom facilitates media access for human rights groups, if and when “their” issue is an urgent problem in the real world, as the positive and significant coefficients for the interaction term suggest. The substantive size of the relationship is modest, but consistent across all four models.²⁶ Figure 4 further explores this relationship, based on the results of models M1 and M2. Each pane shows on the y-axis the predicted probability of receiving attention from 0% to 100%. The x-axis varies issue urgency from 4, where electoral malpractice is not a major problem (e.g., Morocco) to 9, where malpractice is endemic (e.g., Burundi). The indicator for a free environment for journalists is set at two levels, low = 0.4 (ca. the level of Sri Lanka) and high = 0.9 (level of Austria).²⁷

Figure 4. The effect of institutional constraints and issue urgency on the predicted probability of achieving media attention. Note: Based on M1 and M2 (Figure 3), at median random effect, other variables held constant at mean, 90% confidence interval. N = 188 (domestic), 265 (international).

The theory of event-driven news would expect that where election fraud, gerrymandering, or voter intimidation are common, election watchdogs should be sought-after sources. And indeed, this is the case in countries where the media approach the ideal of an open arena (depicted by triangles in Figure 4). Here, an increase in issue urgency beyond a certain threshold of about seven virtually ensures attention in both attention arenas. However, the situation is dramatically different in countries where reporters face repression (depicted by circles). Here, the likelihood of coverage is about equal to countries with a free press, when issue urgency is low. But when wrongdoings around elections rise, this is actually associated with a lower likelihood of attention, decreasing to a point of being negligible in the worst cases. Rather than exposing problems, and devoting more coverage to monitors as key sources, the press either self-censors or bows down to repression. Outside lobbying becomes exceedingly difficult under such conditions.

Election advocates interviewed for this study repeatedly flagged censorship, intimidation of journalists, or government ownership of news outlets as major obstacles. An interviewee from an African country with a “difficult situation” for journalists according to RSF said for example that “it is rare to be covered unless you are a government entity[…] media pluralism is on top of our wish list for reform.”²⁸

Maximizing Attention

All three indicators of information subsidies are predictors of attention, but to varying degrees. Media effort is associated with a higher likelihood of attention and a higher absolute amount of attention in both news arenas (M1, M2, M3, and M4), whereas specialization predicts the threshold and the count of domestic attention (M1 and M3), but only the count of international attention (M4). Professionalization finally is associated only with domestic attention (M1 and M3).

Media effort is indeed the strongest of all model predictors. The interviews corroborate that the vast majority of groups recognize media effort as a strategy with high payoffs. Interviewees talked of the “one-minute pitch” and said they felt “like an advertisement agency trying to get content to be viral.”²⁹ There was an understanding that the more frequent and institutionalized contacts with newsmakers become, the better.³⁰ Some organizations sign a memorandum of information-sharing with news outfits every election. This mutually beneficial exchange relationship “works both ways,” as news organizations “don’t have employees throughout the country. So, if we tell them [about malpractices], they can either send a reporter, or our own volunteers can even serve as their citizen reporters on elections.[…] In return, we get our voter education program in for free.”³¹ The groups with the strongest involvement in media effort convene “situation rooms” on election day, where journalists can observe the NGOs work in action, are provided with regular press updates/conferences, and are quite literally subsidized with an infrastructure of printers or Internet access.³²

Specialization, niche expertise, and superior data were also noted as crucial assets by interviewees. Some were highly confident of the newsworthiness of such narrow policy focus, saying things such as “if you read anything about how much a campaign costs in Guatemala, you know it through our data. And this is reflected in the media.”³³ Most groups strive to showcase their specialization whenever possible, presenting data at press conferences that no one else could produce (e.g., about malfunctioning voting machines, which would “definitely” make them “front-page material”).³⁴ The advantage in expertise coming with specialization is sometimes so large that groups would convene trainings for journalists whenever new electoral issues arise, “so that they know what we’re talking about: the pros and cons. This is in order to skill them.”³⁵

As for professionalization, numerous interviewees were of the view that an organizational division of labor and staff dedicated exclusively to specific tasks aids attention-getting. One group that was particularly lauded at international conferences for their professionalism has “a centralized communication strategy.[…] If regional staff, or office staff are asked to give an interview, they contact management[…] or topic expert, and we agree on the messages.”³⁶ Another interviewee pointed out that the NGO has “a stand-by graphics designer, a trained social media officer, and also a freelance journalist.[…] They help us make our information visually appealing, to condense it, make it easy to understand.”³⁷

Importantly, the statistical models also show that these three group strategies relate to attention in free and in unfree media environments alike. Interaction terms of press freedom with all three organizational indicators are not significant.³⁸ At the same time, the findings are robust to some alternative explanations such as organizational longevity and membership in international networks, which are not associated with more attention. And crucially, a group’s resources are unrelated to news coverage, once lagged media attention is factored into the equation.³⁹ This finding concurs with De Bruycker and Beyers (2015), who also reject the hypothesis that resources facilitate media access, but stands in contrast to some influential studies on groups’ attention-getting, which typically find that resources facilitate access (Binderkrantz et al., 2017; Thrall, 2006).

In summary, human rights groups’ uphill battle for scarce attention is exacerbated by the constraining effect of institutions where they cannot rely on a fair media arena. In free and unfree media landscapes alike, groups’ involvement in media effort relates to higher media access in all arenas. The likelihood and the absolute amount of domestic attention is also associated with group specialization and professionalization, while in the international arena, professionalization does not matter, and specialization only relates to the absolute amount of attention. The results are summarized in Table 1. Taken together, the statistical models as well as the qualitative interviews suggest that information subsidies—facilitated by organizational strategies—give human rights groups “more bang for their buck.” They substitute for resources and provide an effective way to maximize the chances of news coverage, even under conditions of media repression. This gives strong evidence supporting hypotheses H1c and H2a, and partially supporting H2b and H2c. There is also evidence upholding H3a,b,c, because media freedom does not affect the efficaciousness of information subsidies.

Table 1 Summary of the drivers of media attention to human rights groups

Hypothesis	Mechanism	Indicator	Domestic News	Int’l News
H1a	Issue urgency	Electoral malpractice
H1b	Institutional constraints	Press freedom
H1c	Issue urgency under institutional constraints	Press freedom X Electoral malpractice	++	+
H2a	Information subsidies: Involvement	Media effort	++	++
H2b	Information subsidies: Narrow policy engagement	Specialization	++	+
H2c	Information subsidies: Professionalization	Formalization	++
H3a/b/c	Information subsidies under institutional constraints	Press freedom X Media effort/Specialization/Formalization

Note. “+” = variable is significant predictor for either threshold or count; “++” = predictor for threshold and count.

Conclusion and Discussion

Grabbing media attention is the sine qua non of human rights pressure group influence around the world. Activists interviewed for this study made this abundantly clear, saying that “the only way that [our] message gets to the stakeholders is through the media,”⁴⁰ “interaction with the media is really key,”⁴¹ “without presence in the media we would not have achieved what we have achieved,”⁴² and “the only way for us to pressure the government is to make something public.”⁴³

In this article, I presented the twofold argument that in order to develop a more complete understanding of the interest group-media interface beyond the OECD world and beyond counting money and staff, we need to take account of macro-institutional opportunity structures and actor-level strategies. The empirical evidence drawn from a globally comparative study of human rights groups—in particular election monitoring and advocacy NGOs—confirmed this argument. First, the real-world urgency of social problems is only associated with increased attention to human rights groups if a country’s news media landscape is sufficiently free. A certain permeability of the public communication system is thus a necessary condition for agenda-building. Second, groups who are good at producing information subsidies are better at garnering attention. Strategies focusing on media effort, narrow policy engagement, and professionalization—rather than purely resources—allow groups to maximize their exchange relationships with journalists, thereby getting more “bang for their buck.” Third, these information subsidies are effective in free and in unfree media landscapes alike.

Some important limitations of the research should be acknowledged. As a study of attention in two offline arenas (newspapers and wires), one particular group type (citizen groups) in one particular issue industry (human rights), there are certainly limits to the generalizability of results. For instance, where print media are constrained, public attention may merely be displaced to online social media, rather than completely muted. Furthermore, results may not travel to other types of organized interests (e.g., business groups or professional associations). As human rights advocacy routinely attributes blame to the government, and may even threaten the regime’s legitimacy in more authoritarian contexts, caution is also advised in applying the results to other public interest issue areas. Electoral integrity in particular is the ultimate “redistributive” issue (Lowi, 1964), in that it affects the distribution of access to political office and power. The stakes are much higher than in other issue areas, as one interviewed activist noted: “Whenever we publish anything, there is a reaction. I stayed in jail as a result of this reaction.[…] There are 3,000 NGOs in Azerbaijan.[…] Only I and three colleagues—also human rights defenders involved with the election process—were arrested.”⁴⁴ The media, in turn, may shun such groups due to the high stakes involved. Finally, some findings—for instance, the one on narrow policy engagement—may not hold if one considers the whole spectrum of issues rather than only one. Generalist groups may well receive more overall attention than specialists on the media agenda as a whole, simply because they have the potential to be thematically relevant to more than one issue.

Keeping in mind these limitations, the study contributes to several existing research programs, drawing strength from its globally comparative design. First, for studies of interest groups, the lesson that social-institutional variation matters, is one of significance. Outside lobbying is much harder under hybrid and autocratic regimes with limited media freedoms. This calls for an expansion in the empirical scope of interest group and agenda-setting studies. It also raises interesting questions about how groups may circumvent institutional constraints. In this regard, I was able to show that strategies matter, perhaps more than resources once one controls for previous attention success. This adds nuance to studies showing a positive relationship between resources and attention (Binderkrantz et al., 2017; Thrall, 2006), and clarifies the causal mechanisms underlying that finding. Indeed, it is good news for public interest advocates, because—somewhat contrary to previous research (Tresch & Fischer, 2015)—I show that typical outside lobbying strategies facilitate media access. Clever public relations work, niche expertise, and shrewd allocation of specialized staff can lift them out of obscurity, even in repressed media environments. Having said that, the findings do not per se challenge the well-established understanding that interest systems on average tend to be biased in favor of rich groups. They merely establish—in principle—the substitutability of resources among an overall resource-poor sample of interest groups.⁴⁵ The idea of “more bang for the buck” is perhaps most applicable to the underdogs among interest groups.

Second, the findings shed light on the policy process in non-democracies. Even where the press is not free, some citizen groups apparently gain access to the media agenda. Attention may drop off exponentially and may be highly skewed toward the best information subsidizers, but this is not so different in democracies. This suggests that complex ways of interest intermediation exist even under authoritarianism. Autonomous citizen groups may indeed expand and contribute to policy agendas while remaining within the boundaries of subtler, yet tightening state control (Teets, 2014). The extent and limits of this interface certainly calls for further investigation.

Third, there are ramifications for research into the diffusion of human rights norms. The findings add nuance to the “boomerang” model of norm diffusion, in which domestic groups circumvent repression by appealing to international allies and transnational advocates (Keck & Sikkink, 1998). The domestic vector of pressure “from below” in the boomerang model is perhaps less systematically studied. As it turns out, media-savvy human rights activists can influence the media’s agenda, thereby expanding human rights language and values. This creates avenues for mobilization around these issues and for the embedding of human rights norms.

In conclusion, the present research has shown that there is merit in expanding the empirical scope of political agenda-setting research beyond the OECD world, and that there is still much to be discovered about bias in interest representation toward resource-rich groups. It has necessarily left many questions unanswered, providing fertile ground for future research. For instance, do the findings bear out for other issue industries, different group types, or in different media? Under which conditions does media attention precipitate preferred framing and/or political attention? And how do technological affordances mediate attention-getting? All this makes for an exciting comparative research agenda on human rights groups and the media.

Disclosure statement

No potential conflict of interest was reported by the author.

Supplemental material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/10584609.2018.1551256.

Additional information

Funding

This work was supported by a PhD scholarship provided under the Australian Research Council Kathleen Fitzpatrick Australian Laureate Fellowship Award [FL110100093], and by The University of Sydney’s Faculty of Arts and Social Sciences’ Doctoral Research Travel Grant; University of Sydney [FASS Doctoral Research Travel Grant].

Notes on contributors

Max Grömping

Max Grömping is a lecturer at the Institute for Political Science at Heidelberg University, Germany.

Notes

1. Interview with Guatemalan electoral reform advocate. See online Supplemental Material, interview GTM-01-UMM.

2. “Outside” lobbying consists of the mobilization of public opinion via news media and public actions, and is often juxtaposed with “inside” strategies, such as direct interactions with policymakers (Beyers, 2004; Kollman, 1998).

3. Human rights groups are here defined as locally headquartered nonprofit collective actors other than political parties or bureaucracies seeking the achievement of human rights-related legislation and/or practices that, if materialized, would benefit all of society, not only the group’s members or constituency. They use a human rights frame in their mission and draw on international norms, but also domestic laws (see also Ron et al., 2017, p. 3). The definition used here pertains to local actors and does not encompass international nongovernmental organizations (INGOs). However, the category may contain actors studied elsewhere under the rubrics nongovernmental organizations (NGOs), social movement organizations (SMOs), or civil society organizations (CSOs).

4. Outside lobbying by (international) human rights groups is often discussed under the heading of “information politics” (Keck & Sikkink, 1998; Ramos, Ron, & Thoms, 2007; Ron, Ramos, & Rodgers, 2005; Thrall et al., 2014). Interest group studies also note outside lobbying as the strategy of choice of citizen groups (Dellmuth & Tallberg, 2017; Dür & Mateo, 2013; Hanegraaff, Beyers, & De Bruycker, 2016; Junk, 2016; Tresch & Fischer, 2015).

5. Interview GTM-01-UMM.

6. Human rights groups are most akin to the category of “citizen groups” in interest group studies (Berry, 1999). See also Beyers, Eising, and Maloney (2008) for a discussion of the overlap of different conceptual frameworks in this space.

7. See Binderkrantz and colleagues (2017), Dür and Mateo (2013), or Tresch and Fischer (2015), for comparative perspectives, which are nevertheless situated in OECD democracies.

8. These are nonstate, nonprofit, nonpartisan, and nonmedia collective actors that witness and document electoral malpractice in their own country, and/or advocate for legislative or procedural changes in the way elections are conducted (Grömping, 2017). Political party observers, international election monitors, news media, and for-profit electoral assistance providers are excluded.

9. Defined here as electoral conduct adhering to international norms applying universally to all countries and throughout the whole electoral cycle (Norris, 2013, p. 564).

10. Most likely, institutional constraints will affect different issues in differential ways. Issues that readily allow or even imply blame attribution to the government (such as human rights) will likely be curtailed first, while the news gates may remain open longer for more “benign” issues. However, this differential effect is not tested in this study.

11. It should be noted that the concept “information subsidies” is unrelated to the types of information (e.g., expert knowledge or knowledge about encompassing interests) that groups provide (Bouwen, 2004). Indeed, information subsidies reduce the costs journalists face to discover either type of information.

12. This argument has most traction when focusing on only one clearly delineated issue industry (human rights in this case). The opposite argument has been made in broader studies of interest groups’ lobbying success across many issue areas, where generalists may be at an advantage (Binderkrantz & Pedersen, 2017).

13. See Kelley (2012) on international and Grömping (2017) on domestic election monitoring.

14. See Footnote 8.

15. The online Supplemental Material details the sampling procedure, measurement of variables, and the interviews.

16. All elections under observation took place between 2011 and 2016.

17. https://global.factiva.com.

18. The availability of news sources in Factiva may have led to the nonrandom exclusion of local newspapers, or those in vernacular languages. What is presented here is thus a conservative measure of attention. If it registers attention to a group, there are likely many more articles about the group in the real world. Still, the possibility of false negatives cannot be ruled out.

19. Conducting the content analysis around a high-salience political issue (election integrity that simultaneously receives a spike in attention (an election)— two factors predicting a convergence of the agendas of different news outlets into one national news agenda (Atkinson, Lovett, & Baumgartner, 2014, p. 374)—justifies using only a few newspaper sources per country (up to a minimum of one). However, as a robustness check, all analyses were duplicated with a minimum requirement of two sources per country. The results are substantively the same.

20. Operationalized through the Factiva autocodes “Reuters news wires” (trtw) and “World wide news wires” (twww).

21. https://rsf.org/en/ranking.

22. The three indicators are not colinear with each other (Pearson’s r between –.14 and .21) or with the other group-level controls (Person’s r between .12 and .35). They thus likely tap into different group characteristics. Missing data in the survey responses are treated with multiple imputation using chained equations (Rubin, 1987), allowing observation of the independent variables for 298 groups. Details are found in the online Supplemental Material.

23. That is, articles containing the terms “election*” OR “elector*” and their equivalents in other languages.

24. See Hair, Black, Babin, and Anderson (2009) for a discussion of these rule-of-thumb thresholds.

25. All models are fully reported in the online Supplemental Material. Figure 3 shows the best-fitting ones as judged by log likelihood, Akaike information criterion (AIC), and the root of the mean squared error (RMSE).

26. In M4, the interaction is slightly less significant than in other models at p < .1, but just barely outside the .05 threshold.

27. Some countries in the data have even harsher conditions for journalists (e.g., Yemen or Sudan with scores of below 0.2). However, there are only few survey responses and furthermore insufficient variation of electoral malpractice in these countries. Therefore, a cutoff of 0.4 is chosen for the depiction in Figure 4.

28. Interview ZWE-01-VCR.

29. Interview JOR-01-FAD.

30. Interview LBN-01-LKJ.

31. Interview PHL-01-QMD.

32. Interviews JOR-01-FAD, KEN-01-OMM, KHM-01-EPK.

33. Interview GTM-01-UMM.

34. Interview PHL-01-QMD.

35. Interview ZWE-01-VCR.

36. Interview GEO-01-WLN.

37. Interview JOR-01-FAD.

38. Except for M4, where press freedom interacts significantly with media effort, but not the other two strategies.

39. See the full models in Tables A.8 through A.11 in the online Supplemental Material. Resources are significant when entered on their own, but drop out once lagged attention is controlled for (domestic news), or when the group strategies are introduced to the model (international news), except in M4. Furthermore, Table A.12 in the online Supplemental Material reports all four full models excluding lagged attention. Resources are a significant predictor in all of these models when lagged attention is ignored, whereas the effect of all group strategies remains broadly the same. This suggests that resources are endogenous to attention-getting, but how these resources are spent is not, meaning that strategies can substitute for resources.

40. Interview ZWE-01-VCR.

41. Interview LBN-01-LKJ

42. Interview PHL-01-QMD.

43. Interview GEO-01-WLN.

44. Interview AZE-01-YMA.

45. The average election-year budget of the studied groups is only USD $50,000 to $100,000 and the median staff category is 5–9.