Introduction

Timely access to high-quality information is considered a key prerequisite for making good decisions, and any organization—private or public—hinges on effective information sharing to pursue organizational and policy goals (Levinthal & March, 1993; Simon, 1997; Spreitzer, 1995). In today’s information age, social media and other forms of electronic communication substantially increase the amount of information that is available on a daily basis and is used to inform decision-making processes. In the context of public policymaking, the increased availability of informational resources is mostly considered in terms of enhancing a government’s analytical capacity (Howlett, 2015), that is, its instrumental rationality as pursued through evidence-based approaches (see also discussions on algorithmic depoliticization, e.g., Rouvroy & Berns, 2013; Weiskopf & Hansen, 2023). For example, emphasis is placed on the various opportunities of policy analytics (De Marchi et al., 2016) such as geospatial prediction to help allocate resources or agent-based computational modelling to simulate complex social dynamics.

Yet, in order to design feasible and effective policy programs, policymakers not only require analytical expertise on the policy substance at hand. They must also possess political judgment on the needs and interests of societal groups and stakeholders and effectively interact with their political–administrative environment to articulate, negotiate, and assert a given policy (Wu et al., 2018). Such political proficiencies form an indispensable pillar of policy capacities, since even the most technically sophisticated policy alternative is no use if it fails against the unanticipated influence or resistance of policy stakeholders (Wu et al., 2018). Here, too, policymakers can benefit from data and information about the political opinions, attitudes, perceptions, and behaviors that are carried via social networking services (SNS) and are arguably conducive to political capacity, defined as “the ability to effectively understand stakeholders and others at work, and to use such knowledge to mobilize others towards achieving a desired policy outcome” (Saguin et al., 2018, based on Ferris et al., 2005).

Although the concept of policy capacity is “among the most fundamental” (Wu et al., 2015, p. 165) components of public policy, its political dimension only plays a marginal role in the public administration’s digitalization debate that first and foremost looks at the use of ICT for analytical and managerial tasks. Accordingly, research on public sector digitalization mainly concentrates on target-specific digitalization projects that transform analog into digital services or change given organizational processes—neglecting that within the broader cultural context of digital transformation dynamics, the work of policy officials is affected by a trend of acceleration and hyperconnectedness, where the boundaries between private and professional communication and information retrieval are blurred. SNS have become an indispensable part of our daily lives—especially in terms of political news coverage—and these changes in the trends of seeking and encountering information affect the work of policy officials drafting laws and regulations. Yet, they have found their way into the working world rather subtly; perhaps for this reason, they have been overlooked by public administration and policy scholars.

This paper aims to contribute to a more comprehensive and holistic understanding of policy formulation realities in the digital age, which must include the implications of SNS for government employees’ individual-level information use and political capacities. In fact, the lack of systematic empirical evidence on this issue neglects the political sensitivity that policy officials require. In the policymaking process of Western democracies, policy programs are prepared and shaped by ministerial bureaucracies in which the political leadership relies on the bureaucratic-professional expertise of civil servants (Aberbach et al., 1981; Page & Jenkins, 2005; Page, 2012). These policy officials anticipate and integrate politically relevant aspects into their day-to-day functions (e.g., Hustedt & Salomonsen, 2014, p. 750), often pre-negotiating and coordinating policy matters. In doing so, they also filter and reduce the amount of information because information processing capacities decline at higher hierarchical levels. As Wu et al. (2017, p. 8) point out, “contrary to what is commonly believed, political capacity is not only essential for senior policy-makers and officials, but also for policy analysts and experts.”

This paper investigates the effects of SNS for policy officials’ information use in the daily routines of policy work and their implications for individual political capacities. Based on survey data collected between June and October 2020 in the central ministerial bureaucracies of Germany, Italy, and Norway, a moderated mediation analysis was conducted to test the effects on the amount of information utilized in in legislative drafting. The paper also considers the variability of information sources as well as a concentration measure (i.e., the Herfindahl–Hirschman Index) to check whether SNS lead to a more concentrated use of specific information types compared to others. Furthermore, it scrutinizes the presence and effects of social media stress on information use and political capacities in order to address the increasingly discussed tension between SNS as an empowering tool and the mental exhaustion that can result from overexposure to SNS (Bright et al., 2015). On the one hand, SNS reduce the costs of communication, documentation, and information gathering, allowing for the communication and exchange of more information in the same amount of time. They thus facilitate access to the information needed to weigh the consequences of policy choices and also to appraise social actors’ preferences, missions, and visions, which are essential components of organizational empowerment structures (Spreitzer, 1995). On the other hand, the development toward faster yet shorter communication with greater actor involvement may tempt people to produce more abridged input, potentially exceeding policy officials’ processing capacities and hampering their understanding of political priorities and preferences. This paper examines such individual-level consequences of SNS in central government administrations and adopts a comparative perspective to allow for conclusions across the Roman-Scandinavian, Roman-Germanic, and Roman-French traditions. The opportunities and pitfalls of ICT for government-to-government, government-to-society, and government-to-citizen collaboration have mostly been considered from an organizational-level perspective, using qualitative and small-N studies to examine the impact of purpose-built ICT tools on analytical know-how or collaboration dynamics. SNS in particular are studied with regard to how governments can mine social media sentiment to gauge public opinion or forecast social unrest (e.g., Driss et al., 2019). In contrast, this paper addresses the lack of systematic empirical evidence on the impact of SNS on government employees’ individual-level information use and political capacities.

The rest of this paper proceeds as follows. The subsequent section introduces the theoretical framework and hypotheses, drawing on communication and behavioral research. The paper introduces the methodological procedure that was applied to test the hypotheses, and then it presents the results. The last section discusses the empirical findings as well as the implications of SNS use for the political capacities of policy bureaucrats.

Theoretical framework

Social media and information use

As the “memory of the state” (Bull, 2008, p. 215), ministerial bureaucracies steadily accumulate policy field-specific expertise, including legal, technical, and financial competencies but also the institutional knowledge on administrative procedures resulting from their experience in service. According to Leibholz (1967, p. 318), “Without the ministerial bureaucracy, neither the government nor legislation can fulfil their functions.” After all, political executives’ capacities are limited, and they usually rely on the strongly specialized knowledge of ministerial directorates—or, rather, the ministerial units they are composed of—to initiate and draft laws and regulations. Much of these units’ knowledge is generated through regular exchange with external stakeholders themselves, such as interest groups; they rely on stable policy networks to accomplish their tasks while also considering the political signals of the ministry’s top echelon. Adequately performing these knowledge-providing functions requires fast access to information, a task strongly facilitated by digitalization dynamics.

Private messenger services allow for rapid exchanges with colleagues, and many ministries even provide their own in-house messenger services to enhance the sharing of government documents, legal advice, and opinions. Together with social media such as Facebook, Twitter, and LinkedIn, these tools are usually summarized under the term “social networking services” (SNS), which can play an important role for policy officials to disseminate and obtain different types of information. For example, the content communicated via social media is not only comprised of journalistic work, sensational information, and political and partisan ideas and opinions but also of information provided by interest groups, survey and statistical data, expert opinions, and much more. These resources are shared in real time, processed in an interactive and collaborative manner, and often include voices that are not typically covered by conventional or dominant media; they thus provide a rich array of information. The reasons for engaging in social media are manifold, ranging from active information seeking or expressing individual opinions to passing the time and entertainment functions. Accordingly, people may encounter certain information not only from purposeful action but also rather incidentally—for example, when the newsfeed offers cross-cutting content or when followed social media accounts share various types of informational resources.

Such passive exposure to information has long been discussed by communication researchers who, for instance, distinguish between routine (passive) and non-routine (active) information seeking (Wilson, 1997) or between information search and information receptivity. The latter describes “receptive encounters with topic-related cues during routine scanning of the message environment” (Atkin, 1972, p. 191, as cited in Metag, 2020). Bates (2002) classifies information seeking in terms of “activeness” and “directness,” that is, between incidental information encounters where the recipient had no prior information needs (passive and undirected), encountering information that is interesting and useful to the recipient (passive and directed), unfocused information seeking (active and undirected), and active, purposeful information seeking (active and directed). Compared to traditional forms of information seeking and news consumption, SNS offer more incidental encounters; they especially facilitate passive forms of informational behavior. Recently, communication researchers have pointed to a phenomenon called “News-Finds-Me” perception (e.g., Gil de Zúñiga et al., 2017), which describes the increasingly widespread belief that one can stay informed without actively seeking out news—a perception that is closely tied to time spent online, especially on social media. Social media also provides greater opportunities to interact with a given piece of content, so that even passive news consumption evoked by algorithmic suggestions or content shared by friends and colleagues may facilitate further information seeking and engagement with the matter at hand (cf. Geise et al., 2021). Overall, social media offers high choice in information resources; it is independent of time and location, and it can enhance policy officials’ capacity to find work-relevant information, including through passive consumption. Faster to be shared across messenger service.

H1

The higher policy officials’ SNS use, the higher their information use to accomplish their policy tasks.

Information use and individual political capacities

Insofar as social media and messenger services provide policy officials with additional platforms to find, exchange, and encounter various types of information, they also enable a better “read” of policy actors’ positions and interests. As policy capacity research emphasizes, being able to assess the political feasibility of a given policy and understand stakeholders’ needs and logics of action is a key prerequisite for effective policymaking (Mukherjee & Bali, 2019). Such political acumen not only involves knowledge about the policy process itself—e.g., the way certain actors interact with each other or when certain information is needed to effectively engage in coordination and consensus building (Wu et al., 2015). Policy bureaucrats also require knowledge about the political leadership’s mission and goals, as well as the ability to articulate their department’s medium- and long-term priorities (Riddell, 1998). Indeed, political proficiencies involve some kind of “tacit knowledge” (cf. Polanyi, 1958) about how to pursue particular goals in particular situations, both in terms of intra-organizational dynamics—e.g., briefing policy managers and decision makers such as cabinet ministers or senior managers—and external actors involved in policymaking—e.g., negotiating with stakeholders or consulting the public on policy matters (Ramesh et al., 2016).

As such, the political component of administrative policy work entails considerable leeway in drafting laws and regulations, and research on ministerial bureaucracies has repeatedly confirmed their central role in the policy formulation process. Although the room for maneuver depends on the legislative matter and is smaller for politically salient, attention-grabbing policies, most regulatory initiatives offer a lot of space for ministerial policy officials to exert influence. In particular, politicians crucially depend on bureaucratic expertise if the issues are more technical in nature. Of course, ministerial officials can use their information advantage for tactical purposes as well, for example, by selecting, framing, deferring, or even withholding certain information in an attempt to influence policy choices (Korte, 2019). Apart from normative debates on the politics–administration dichotomy, empirical studies remind us of the importance of ministerial “backstage policy advocacy, where policy bureaucrats in a “(silent) dialogue” (Wegrich & Bach, 2020) with the upper political echelon anticipate potential support and resistance for policy alternatives and include party and power-political considerations in inter-ministerial coordination processes. This study does not aim to contribute to the functional politicization literature but instead investigates political capacities in an administration context, which requires some remarks on why such capacities are an important prerequisite in the first place. It is within this ministerial bureaucracy context that an information advantage and policy bureaucrats’ expertise are expected to shape, and thus precede their political capacities instead of a reversed causal pathway. The survey and experimental studies of Blom‐Hansen et al. (2020) demonstrated the importance of policy information for political capacities and how they can shape political decisions, showing that the majority of bureaucrats agreed to anticipate political decisions that affect the policy information they emphasize. The authors also found evidence that bureaucrats prepare information in a way that increases the chances of politicians choosing the bureaucrats’ preferences—the professionally superior options, in the bureaucrats’ views—so that political decisions were susceptible to policy proposals being framed in a positive, neutral, or negative way. Politicians also spent significantly more time on information provided by bureaucrats than by other policy-relevant actors (think tanks, researchers, etc.); it was, in fact, the first source of information to be assessed before consulting other sources, as well as the information they relied on the most.

In order to accomplish their policy tasks, ministerial officials thus require quick access to informational resources, such as policy-relevant data, proposals, protocols, and information and opinions provided by interest groups, think tanks, and consultants. SNS constitute a further platform for officials to find and exchange information or even identify key policy actors. On social media, stakeholders affected by or involved in policy processes communicate their needs and views, often providing signals about resources and strategies. In the diplomatic context, Ashbrook and Zalba (2021) demonstrated the importance of “anticipatory” social media use and how failure to effectively engage in digital listening may have highly disruptive effects on negotiations at the table, negotiations away from the table, and in the negotiated policy outcome. Having access to a variety of information sources is crucial to improve one’s analytical power but also to respond to policy opponents in a timely manner, convincing others of a policy’s value and thus enforcing policy ideas.

Furthermore, informational signals from within the ministry are crucial for employee empowerment. Such resources facilitate individual sensemaking and reduce ambiguity, which is particularly important in the context of complex, dynamic, and uncertain policy issues. Having access to information helps policy bureaucrats create meaning and purpose in their tasks—understanding how the political components of their work can contribute to achieving ministerial objectives and ensuring that their decisions and actions are properly aligned with those of the political leadership. Simon (1997, p. 177) stressed the importance of organizations reducing uncertainty and complexity by providing their members both with factual and value premises, where value premises describe information on the desirable outcomes, objectives, and decision criteria as specified by the organization. Additionally, social cognition theory proposes that the provision of informational resources enhances bureaucrats’ self-efficacy, a concept used to describe “people’s judgements of their capabilities to organize and execute courses of action required to attain designated types of performances” (Bandura, 1986, p. 391). Applied to the context of policy bureaucrats’ political capacities, those scoring high in their information use are likely to feel better equipped for their everyday political tasks.

Access to information may also constitute a competitive advantage, providing employees with the knowledge necessary to engage in boundary spinning, coordination, and conflict management and even allowing them to operate as a means of social power in that well-informed persons are perceived as centrally connected and more influential. In this sense, informational empowerment allows policy bureaucrats to develop expertise and coordinate individual action (cf. Mintrom, 2015).

H2

The higher policy officials’ information use, the higher their political capacities.

Expecting that SNS increase the amount of information, which, in turn, increases officials’ political capacities, one can also assume an indirect, positive effect of SNS use on political capacities via increasing the amount of policy-relevant information. It is also conceivable that SNS directly enhance individual political capacities not only by relying on the actual information retrieved from social media but also by anticipating public opinion. Within the mediatization debate, for example, media awareness and media pressure were found to increase politicalpolitical logics of action within ministries (Salomonsen et al., 2016), arguably because mediatization dynamics raise both the inclination and ability of civil servants to perform their increasingly politicalized functions. The growing relevance of social media may shift dynamics in the organizational environment of ministries, thus affecting behaviors in policy formulation (see also the social information processing perspective). However, such a perspective eventually boils down to prior experiences with social media debates (i.e., information from and about policy-relevant actors). This study therefore focuses on the indirect effect and expects that.

H3

Policy officials’ information use mediates the positive relationship between their SNS use and their political capacities.

The moderating role of social media stress

While SNS create fewer obstacles to accessing information and are likely to enhance political capacities, too much information can also have overwhelming effects, causing cognitive burden, uncertainty, and diminished judgmental capability. In particular, the concept of information overload points to a state in which the sheer volume of information affects peoples’ ability to make (good) decisions or stay informed, given the limited capacity of the human brain and the limited time available to process it (Zhang et al., 2016). Similarly, research on communication overload has shown that SNS can cause stress by distractions, with an excess of messages or information, as well as by compromising message quality and generating pressure for fast decisions and responses (Stephens et al., 2017). An oversharing of pertinent, redundant, or duplicated information can be especially stressful when consumed passively, while the active use of SNS to search for and exchange information is less likely to have adverse effects (cf. Kross et al., 2021). With accelerating SNS dynamics, information and communication overload can occur more easily, impeding policy officials in their ability to (effectively) integrate that “extra” information into their policy tasks. Social media stress may also impair officials in judging the importance of certain information or prioritizing them based on the organizations’ values and objectives. In other words, social media stress may alleviate both stages of the indirect effect between SNS use and individual political capacities—i.e., attenuating the positive effect of SNS on information use as well as the positive effect of information use on political capacities (see Fig. 1).

Fig. 1
figure 1

Conceptual model

Full size image

Although the harmful effects of social media stress in the context of ministerial policy formulation have hardly been explored, the phenomenon is well known in the behavioral sciences. A 2013 survey among United Kingdom SNS users showed that two thirds of Twitter users reported seeing too many posts, and more than half of the respondents wished for a tool to filter out irrelevant content (Bontcheva et al., 2013). Those using Twitter in a professional capacity were more likely to feel overloaded with posts but were also less likely to un-follow accounts they regard as uninteresting. With SNS constantly offering new content and updating users’ newsfeeds every few minutes, policy bureaucrats may feel obliged to keep up with the provided information or fear that they might miss the latest news—i.e., sensing a fear of missing out (“FOMO”): the concern of missing a rewarding experience in one’s absence (Przybylski et al., 2013), a feeling of anxiety that one is “left behind” (Salem, 2015), which evokes the desire to stay constantly informed and connected with others. Colleagues’ use of SNS may especially trigger pressure to do so, leading to an upwards spiral of who knows first; it is a fear of missing the latest information to one’s own disadvantage in policy coordination processes. Moreover, people must read a considerable number of posts to encounter interesting content, yet they do not want to sacrifice the total amount of interesting tweets by reading less (Bontcheva et al., 2013). Aggravating this situation, information retrieved from social media usually underlies the recency effect, according to which people attribute great weight to the most recent information. In the fast-moving political environment, shared content can cause continuous interruptions; however, due to recency of exposure and FOMO, each new message or piece of information is attached with great importance, which may ultimately pose a threat to policy officials’ productivity. An abundance of information can even have “paralyzing” effects, inducing people to take a back seat, which may also apply to ministerial coordination processes of policy formulation.

Although SNS are also considered as resources or coping tools to buffer stress, such a perspective is based on an understanding of SNS as social capital—i.e., a way for establishing and maintaining social contacts or seeking distraction from negative emotions. Within the professional capacity of ministerial bureaucracies, such mechanisms are arguably of less importance. Indeed, the invasion of colleagues into the private sphere outside working hours demonstrably constitutes a techno-stressor with negative effects on various performance outcomes. However, it is interesting to note that those who experience high levels of stress are not necessarily those who use SNS more often: rather, information overload is reported by less frequent users, those often, seldom, or never engaging in SNS activities (Bontcheva et al., 2013).

Lastly, information provided via SNS can contain misinformation or fake news. While the condensed or concise feature of content on SNS may facilitate the comprehension of most current news, it may not similarly enhance an in-depth understanding of policy problems, impact chains, and their interdependencies. Extensive information, instead of providing reliable guidance, may thus hamper policy officials in understanding and mobilizing others in policy formulation. In other words, social media stress may act as an attenuating moderator that hampers the beneficial effects on bureaucrats’ information use and political capacities. Thus, the last hypothesis assumes the following:

H4a

The effect of policy officials’ SNS use on their information use is moderated by social media stress, such that the effect is lower under high levels of social media stress.

H4b

The effect of policy officials’ information use on their political capacities is moderated by social media stress, such that the effect is lower under high levels of social media stress.

H4c

The indirect effect of SNS use on policy officials’ political capacities via their information use depends on the degree of social media stress, such that the indirect effect is stronger when social media stress is low rather than high.

Methodology

Sample

The data used to test the hypotheses were collected as part of an online survey about digitalization dynamics, conducted across all the central ministerial bureaucracies of Germany, Italy, and Norway. These three Western European countries are bureaucratic meritocracies that belong to the Roman legal tradition of civil law, where codified statutes predominate. However, they represent different subtypes of the Roman tradition and thus distinct structures in terms of the ministries’ position in the political process (i.e., bureaucratic influence) and the civil service systems (Schnapp, 2004, p. 47). As part of the Roman-Scandinavian tradition, Norway is a consensus- and citizen-oriented society (Painters & Peter, 2010, p. 20), which values quality and the rule of law over individualism and efficiency. Accordingly, reform elements of digitalization, transparency, and e-participation are attached with greater importance compared to other European countries or formerly neglected NPM efforts (Greve, 2020). Norway has an open civil service system with a high permeability between the public and private sectors, and there are many opportunities for the senior executive service to shape policy formulation through strategic interaction, though parliamentary constellations may narrow ministries’ bureaucratic agenda-setting potential. In contrast, the administrative culture of the Roman-Germanic tradition is characterized by a closed civil service system with rigid career tracks and a high status of civil servants who are united by a cross-sectoral “class consciousness” and who enjoy few restrictions on political activities (e.g., party membership, public political statements). Along with Italy, Germany is one of the most legalistic states in Europe, and its national bureaucracy is dominated by administrative lawyers. Moreover, Germany is less citizen-oriented but more state-oriented, with a high acceptance of governmental regulation authority. Like in Germany, reform efforts in Italy (the Roman-French tradition) are usually embedded into the juridical paradigm, that is, rather hampered by the juridical way of thinking. However, the Italian bureaucracy does not rest on a strong state tradition; Italians have a low acceptance of central state interference, and politics and administration are strictly divided (Kickert, 2005).

The central government officials of the three countries were identified by drawing on the ministries’ websites, organizational charts, and social networks. The sample encompasses all civil service positions below the upper political echelon (e.g., state secretaries), as these officials are involved in drafting laws and regulations—they oversee and coordinate policy work or deal with the substance of policies. In total, 3,701 officeholders were invited via an e-mail that provided them with a link to the survey; 382 respondents answered all the survey items required for this study. Mirroring the actual government workforces, the participants in the final sample are 53.56 years old on average, with 61% being male and 39% being female. Approximately 20% lead a directorate or sub-directorate (i.e., higher hierarchical levels), whereas 80% percent are the heads of the smallest ministerial unit or their direct subordinates. Moreover, the respondents are roughly representative across the policy fields of interior public order and safety (22%), economic affairs (25%), welfare (24%), foreign affairs (14%), and environment and natural resources (15%).

Measures

Political policy capacities were measured with three items (see Table 1) partially adapted from Saguin et al. (2018). Respondents were asked about their confidence in (1) briefing politicians on different policy alternatives, (2) explaining their unit’s medium- and long-term priorities to policy-relevant stakeholders (e.g., parliament or opposition parties), and 3) communicating policy-related content to relevant stakeholders (e.g., interest groups, NGOs). In order to adapt these items to the ministerial bureaucracy context and provide a reliable measure for government officials’ policy work, the statements were introduced with “For my work in policy and program development, I feel confident that I can” and thus built according to the commonly employed structure of self-efficacy scales. Such claims of self-belief are concerned with respondents’ perceived capability and are phrased in terms of “can do” rather than “will do” (Bandura, 2006, p. 308), capturing policy officials’ sense of empowerment as arguably strengthened through informational resources. To allow for the possibility that not all tasks covered by the political capacity items are actually performed by each respondent, the introductory note stated: “In case you do not perform these tasks, we kindly ask you to indicate how confident you would feel to perform them and to additionally tick the box ‘Not part of my tasks’. For the same reason, three of the four questions answered were considered adequate to calculate the mean of policy officials’ political capacity scores. Answers were given on a 5-point response format (1 = strongly agree; 5 = strongly disagree). Cronbach’s alpha for this scale was 0.78, with an average variance extracted (AVE) of 0.56, indicating good convergent validity.

Table 1 Means, standard deviations and factor loadings of the study variables
Full size table

SNS use was measured by asking respondents how often they use the following services for their work: (1) social media (Facebook, LinkedIn, Twitter, etc.), (2) internal messenger services (provided by the organization), and (3) private messenger services (WhatsApp, Threema, Signal, etc.). Again, the analysis took into account the possibility that not all of these three tools are used by each respondent. The mean was constructed with at least two available items, especially since organization-specific messenger services may not be provided by each ministry. Each tool’s frequency of use was indicated on a 5-point scale: 1 = never, 2 = seldom (a few times per month), 3 = sometimes (a few times per week), 4 = often (once a day), 5 = very often (several times a day).

Information use was assessed by asking respondents how often they rely on the following information sources to accomplish their tasks: (1) interest groups (industry, non-governmental organizations, etc.), (2) expert opinion (think tanks, consultants, etc.), (3) survey and statistical data (data provided by the World Bank, OECD, etc.), (4) traditional media (print, radio, TV, etc.), (5) government documents (policy/cabinet proposals, white papers, diplomatic cables, protocols, etc.), and (6) laws, regulations, and legal opinions. The answers were given on a 5-point scale ranging from 1 = never to 5 = very often (several times a day) and transformed into three distinct scores of information use.

First, the mean of all eight items was computed to indicate the average amount of information used in the policy work of ministerial bureaucracies. Because such an indicator does not factor in the variability of information sources, that is, the absolute number of information type categories that enters the work of policy officials, a second score neglects the frequency of each information use and only reports the count of categories as ranging from zero to six. A third indicator considers both variability and balance, thus accounting for the (un)even distribution of information types in the frequency of their utilization (Schaeffer, 2016): the Herfindahl–Hirschman Index (HHI) is a concentration measure that uses categorical shares to calculate the relative frequency—in this case, among each of the eight information types. It was computed for each policy official by summing the squares of the utilization share of each information type. The HHI is widely known as a measure of market concentration, but it is also suitable for describing the diversification of information sources because, as a fractionalization index, it reflects how homogeneous or diverse the landscape of information use is in consequence of SNS applications.

Lastly, social media stress was measured with five items, four of which were adapted from the communication overload scale of Cho et al. (2011)—see Table 1. The items were introduced with “When using messengers (WhatsApp, Threema, Signal, etc.) and/or social networks (Facebook, LinkedIn, Twitter, etc.) for work” and included statements such as “I receive too many messages from my colleagues.” A fifth item asked about respondents’ level of agreement with the statement that they often spend more time with these tools than they initially intended to; it was included to account for the implications of passive SNS use (e.g., online vigilance, availability stress, or even FOMO that can be triggered by the microtargeting algorithms and push notifications of social media and messenger services). Cronbach’s alpha for this five-item scale was 0.91, with the four items from Cho et al. (2011) scoring an alpha of 0.89, and an AVE of 0.66 indicates that 66% of the items’ variance is explained by the latent variable.

Control variables

Several control variables were included in the analysis—most importantly, policy bureaucrats’ position, which may affect both the information used and the confidence in briefing and negotiating policy matters with internal and external actors. Moreover, age and gender were included as covariates, as well as a dummy variable indicating whether respondents have a legal professional background, which is likely to affect the sources of information used to accomplish policy work. Furthermore, ICT use attitudes may shape overall digitalization and information and data use dynamics; they were measured with three items, such as “In general, I like to try out new ICT tools,” and achieved a Cronbach’s alpha of 0.86 and an AVE of 0.67. Because the survey was conducted during the acute COVID-19 crisis, the respondents’ primary working location (home office or office) was included in the analysis, thus accounting for the crisis-induced promotion of telework practices which may alter perceptions of political-administrative-processes. Relatedly, policy capacity research emphasizes the importance of operational autonomy for a targeted use of information, especially in a political context; therefore, three items by Spreitzer (1995) measuring job autonomy were factored in, e.g., asking respondents if they feel to “have significant autonomy in determining how I do my job”. Lastly, country dummies were included as covariates because the administrative tradition may shape political dynamics and information use as well as social media stress and ICT use attitudes. For example, more open civil service systems may increase the extent of prior ICT exposure.Footnote 1

Confirmatory factor analyses

Because the study variables comprise four latent constructs that may entail theoretical proximity (between social media stress and ICT use attitudes or between individual political capacities and job autonomy), discriminant validity was tested by applying confirmatory factor analysis (CFA) on these variables. The hypothesized four-factor solution provided a very good model fit and yielded significantly better fit indices than alternative three-factor models in which social media stress and ICT use attitudes loaded on one factor, political capacities and autonomy loaded on one factor, or social media stress and political capacities (which significantly correlate) made up one factor (see Table 2). Subsequently, measurement invariance was addressed by conducting a multi-group confirmatory factor analysis across Germany, Italy, and Norway. First, configural invariance was established, confirming structural equivalence across the countries. In a second step, metric invariance was tested to examine measurement equivalence, revealing that factor loadings—i.e., scale intervals—are equal across groups (see Table 2). Factor loadings did not fall below the often-recommended threshold of 0.5 for any of the variables in any of the countries. In fact, the vast majority of cases exceeded 0.7, which means that more than half of the variance is accounted for by the respective factor (Hair et al., 2010).

Table 2 Summary of the hypothesized and alternative measurement models
Full size table

Overall, Norwegian and German respondents reported equal levels of political capacity (M = 4.12 and M = 4.09, respectively), while Italian respondents were slightly less confident in performing political tasks (M = 3.79). However, looking at each task separately, the confidence levels of Norwegian officials vary more than among the other groups: they feel very confident in briefing politicians on different policy alternatives and—though more confident than Italian and German officials—less certain in communicating policy-related content to relevant stakeholders such as interest groups or NGOs. They feel even less confident than German and Italian respondents in explaining their unit’s medium- and long-term priorities to stakeholders such as the parliament or opposition parties (see Fig. 2). This comparatively high uncertainty in the communication with parliament can be attributed to the political constellation of the cabinet and the parliament (i.e., low reliability of parliamentary support due to minority governments and, thus, less room for effective influence in parliament), as well as the decentralized and citizen-oriented administration. In contrast, Germany’s administration is characterized by a strong acceptance of governmental regulation authority and less citizen orientation, which might explain why German officials feel more confident in briefing politicians than in interacting with other stakeholders.

Fig. 2
figure 2

Individual political capacities

Full size image

Italian self-assessments across different tasks do not differ as much and are lower overall, which may be attributed to the stricter separation of politics and administration, as well as to the formally strong and resourceful position of parliament that lowers ministerial agenda-setting potential (Schnapp, 2004, p. 294). Altogether, the findings of this descriptive analysis match the results of corporatism indices that characterize Scandinavian ministries by strong relations with interest groups. In German ministries, these relationships are known to be a bit weaker, but they are nonetheless very strong, while corporatist relations in Italian bureaucracies are less institutionalized, further constraining the exertion of influence in policy formulation.

Information provided by interest groups are used more by Norwegians than by Germans or Italians, while Italian respondents stand out with an extensive use of legal sources. The overall average uses government documents most often, followed by legal sources and traditional media content. With an HHI of 0.19, the respondents’ information use is moderately concentrated.

Results

The intercorrelations of the study variables are listed in Table 3 and show that the relationships between the main study variables point in the expected directions. Moreover, favorable attitudes towards ICT are positively correlated with social media use and negatively correlated with social media stress, whereas older policy officials tend to report more negative attitudes and higher stress levels. Older respondents are also less likely to work from home. Perhaps unsurprisingly, job autonomy is significantly correlated with political capacities, with older respondents reporting higher and women reporting lower job autonomy. Though autonomy is not correlated with the amount or concentration of information use, it reveals a tendency to allow for a more diverse information use. High intercorrelations further emerge from the country dummies, with Italian officials reporting better ICT use attitudes and higher social media use than the German reference group but also higher social media stress and lower political capacity. Norwegian respondents indicate a tendency of lower social media use and stress but also poorer ICT use attitudes. Altogether, each criterion variable is associated with at least one control variable, so all control variables are considered in the main moderated mediation analysis.

Table 3 Intercorrelations of the study variables
Full size table

The hypotheses were tested using the PROCESS macro in R (Hayes, 2022). The macro carries out conditional path analyses and allows for the implementation of non-parametric bootstrapping procedures, a resampling technique that does not assume normality and is recommended for testing the significance of indirect effects (MacKinnon et al., 2004). In a first step, a simple mediation model (PROCESS Model 4) was specified to test Hypotheses 1–3. As outlined in Table 4, policy officials’ SNS use positively relates to their information use (B = 0.29, p < 0.001), supporting Hypothesis 1. Hypothesis 2 predicted information use to positively affect individual political capacities, which was also supported by the data (B = 0.24, p < 0.001). The indirect effect of SNS on political capacities via officials’ information use (Hypothesis 3) was estimated using 95% bootstrap intervals with 10,000 bootstrap samples. An indirect effect can be confirmed when the bootstrap confidence intervals (CIs) do not contain zero (Hayes & Scharkow, 2013). The results provide evidence for a mediating effect (B = 0.07, 95% CI [0.4, 0.12]), though the effect size is small, and both direct and total effects were found to be insignificant (see Table 4).

Table 4 Results of the OLS regression analyses: Average information use
Full size table

In a second step, a moderated mediation model was constructed, using PROCESS model 58 to integrate social media stress into the indirect effects model. The first and second stage moderation hypotheses propose that social media stress moderates the relationship between SNS and information use (Hypothesis 4a), as well as the relationship between information use and political capacities (Hypothesis 4b). The results at the right-hand side of Table 4 indicate a significant negative interaction between SNS use and social media stress (B = − 0.11, p < 0.05), thus supporting Hypothesis 4a. Figure 3 illustrates the attenuating effect of social media stress by plotting the relationship between policy officials’ SNS and information use at high and low degrees of social media stress (i.e., one standard deviation above and below the mean). It can be seen that the positive effect of SNS on information use is greater for policy officials reporting lower social media stress—or, put differently, the benefits of SNS are mitigated by officials feeling strained about the increased demands to process the information communicated via SNS and to effectively integrate them into their work routines.

Fig. 3
figure 3

Moderating role of social media stress for average information use

Full size image

No moderating effect was found in the association between information use and political capacities, rejecting Hypothesis 4b. However, social media stress yielded a significant negative effect on policy officials’ political capacities (B = − 0.13, p < 0.05), suggesting that perceptions of information and communication overload can inhibit political capacities, independent of the frequency of SNS use. Lastly, the overall moderated mediation hypothesis was tested, investigating whether the indirect effect of SNS use on political capacities via policy officials’ information use depends on levels of social media stress (Hypothesis 4c). The bootstrap results as depicted in Table 4 indicate that the indirect effect is stronger when social media stress values are 1 SD below the mean than under the conditions of the mean and 1 SD above the mean.

Because the average amount of information does not consider the variability and balance of information sources, the moderated mediation analyses were repeated with two further information use measures: the absolute number of information type categories, ranging from zero to six (see Table 5), and the HHI to account for the (un)even distribution of information types in the frequency of their utilization (see Table 6). The results show that SNS use also facilitates the variability of information use (B = 0.19, p < 0.001) and leads to a less concentrated use of different information sources (B = − 0.16, p < 0.01), providing further support for Hypothesis 1. The variability indicator also yielded a significant regression coefficient on political capacities (B = 0.17, p < 0.001), whereas the HHI indicator was found to have no effect on political capacities. Furthermore, there was a significant moderating effect of social media stress on the relationship between SNS use and the variability of information sources such that the positive relationship is stronger under conditions of low social media stress (see Fig. 4). No further interaction emerged from the data in either the variability or the HHI analyses. Lastly, the bootstrap confidence intervals indicate a conditional indirect effect in the information source variability model (see Table 5) such that the indirect effect is significant at low and average levels of social media stress but insignificant at high levels of social media stress. However, the effect sizes of the (conditional) indirect effects are negligibly small.

Table 5 Results of the OLS regression analyses: Variability of information use (∑)
Full size table
Table 6 Results of the OLS regression analyses: Concentration of information use (HHI)
Full size table
Fig. 4
figure 4

Moderating role of social media stress for diversity of information use

Full size image

Because the effect of SNS on information use was higher at lower levels of social media stress, a curvilinear relationship between SNS and information use was considered, thus testing whether the positive effect of SNS would reach a point of saturation and then negatively affect information use. The polynomial regression was tested by entering a quadratic predictor of SNS use into the model. The model was indeed significant (F(3.71) = 4.32), p < 0.05), but there was no support for a non-linear relationship regarding the variability and HHI measures, and there was no evidence for a non-linear relationship between information use and political capacities. The results thus provide only weak support for an inverted U-shaped relationship.

A look at the controls revealed that favorable ICT use attitudes yielded a positive effect on political capacities, which may be related to the information-facilitating effect of SNS. Furthermore, lower-ranked respondents used a significantly lower average of information, but the effect shrunk or disappeared when measuring information in terms of variability and balance.Footnote 2 Similarly, job autonomy positively affected political capacities as well as the average use of information but neither variability nor balance. Working from home positively affected the variability and balance of information sources but not the average amount of social media used.

Lastly, Italian respondents reported significantly lower/less variable/more concentrated information use; similarly, Italians were lower in their self-assessment of political capacities than the German reference group, which resonates with the high level of political control in Italian ministries. This finding may hint at the important role of historical path dependencies that become particularly vivid in the case of Italy, which has repeatedly struggled to provide lower-level bureaucrats with discretion in formulating policy and to create a fiduciary relationship between the political leadership and top civil servants. The post-war period rested on a “bargain” between politicians and administrators in which civil servants gave up all their discretionary power for job security, and corruption scandals in the early 1990s gave rise to administrative reforms that sought to strengthen the role of upper echelon bureaucrats (Stolfi, 2011, p. 87). However, the reform simultaneously established the temporization of lower-level manager contracts, with their renewal being dependent on politically determined performance criteria and policy objectives. As a result, the reform has reasserted political control over routine decision making (Stolfi, 2011),Footnote 3 and the fact that ministerial officials are considered to have little influence on policymaking processes (Kuhlmann & Wollmann, 2013, p. 41) is mirrored in the political capacity assessments of this contribution.

Conclusion

This paper investigated the political capacities of government officials engaged in policy formulation and how these capacities are affected by accelerating information dynamics in the social media age. In the context of Norwegian, Italian, and German ministerial bureaucracies, the use of social media (Facebook, LinkedIn, Twitter, as well as private and organizational messenger services) positively related to policy officials’ information use, which, in turn, increased their self-reported political capacities. These effects apply to information use measured not only in terms of the average amount of information but also the variability and concentration of information sources, except for the concentration not yielding a significant effect on political capacities. Hence, the availability of many and more variable information contributes to political proficiencies within ministries, whereas a balanced use of information sources seems less decisive. Emphasizing the detrimental of social media stress, however, the positive relationships between SNS and both the amount and variability of information use were found to be diminished when levels of social media stress are high rather than low. Against expectations, social media stress did not moderate the association between information use and political capacities but directly impaired political capacities. These findings emphasize individual differences in stress perceptions and that there is no objectively determinable information level at which policy officials feel overloaded or stressed. The point where more information inhibits political capacities due to distractions and accelerated decision-making processes is highly individual and may also depend on preferences for type and manner of communication.

Amid a dynamic digital transformation debate that is usually linked to expectations of more rational and less bounded decisions, the findings of this study highlight the information-increasing effect of SNS and its potential for improving individual political capacities but simultaneously point to increased demands stemming from dynamics of acceleration and hyperconnectedness. The resulting social media stress can weaken the positive effect on information use and political proficiencies, perhaps even evoking satisficing rather than optimal decision-making behavior, which was hoped to be improved by ICT in the first place, and thus adding to the dynamic complexity of political–administrative coordination and negotiation. In a form of the “‘acceleration paradox,”paradox’ the value of increasingly available information for policy officials seems to be relative to the tools available to process them, but as Entschew (2021, p. 273) points out, “An alleged solution […] to implement better technologies to further optimize time […] would lead to an acceleration cycle resulting in a perceived lack of time.”. Acknowledging phenomena like social media stress in a debate that seems to be dominated by a “rational enablement” perspective opens the view to the race between the Hare and the Hedgehog, as well as to decision-making implications from a mental-emotional perspective that psychology has taken for years, which helps to understand why overcoming bounded rationality in policymaking remains a challenging endeavor despite constant technological innovation.. This applies even more since there are requirements for policy tasks that cannot be solved by more information; in this regard, the effects of ICT on individual political proficiencies below the state secretary would deserve more attention, whether from a policy capacity perspective based on the conceptual work of Wu et al. (2015) or from a functional politicization perspective.

Although this paper does not aim to contribute to the normative debate on the politics–administration dichotomy or to the functional politicization literature, its empirical findings may encourage a “digital mediatization” argument that discusses civil servants’ “social media awareness” or “social media pressure” (cf. Salomonsen et al., 2016) and guides empirical research on politicization dynamics. After all, the increase in demands for policy officials preparing laws and regulations may lead to higher levels of functional politicization in policy work. Equivalent to the enablement versus curtailment theses applied to street-level bureaucrats (Buffat, 2015), informatization and advanced analytical capacities may either weaken ministerial officials’ information advantage—and thus their ability to select, prioritize, and present information—or relieve them from the routine work of gathering information, providing them with more time and resources to consider political implications. From a public administration perspective, the increased use of information shared via social media may also be discussed from a legitimacy perspectives, if the voices of frequently posting actors or of those whose contributions are pushed by algorithms are more apparent than others.

This paper has focused on the effects of SNS in general, including organizational and private messenger services. Future research could take into account the differences between socializing platforms such as Facebook, micro-blogging media such as Twitter, and professional networking services such as LinkedIn (cf. Bontcheva et al., 2013). Twitter, for example, may have much more overload potential compared to LinkedIn, which is widely use to document and advance one’s network but is increasingly used to inform one’s work. Similarly, Facebook is primarily used for social purposes but perhaps with a more select circle of accounts to follow, and it is primarily used to retrieve (political) opinions, attitudes, and perceptions. In addition to distinguishing between different types of SNS that may result indifferent types of information being used,future studies may examine the role of social media for distinct types of political tasks in a more nuanced way and unravel the conditions and effects of SNS in terms of conscious and unconscious information processing in policy formulation, including how officials prioritize information or which information they regard as useful and interesting. This may interact with stress-induced reactions and behaviors as well, bearing important implications for public sector organizations to take preventive mental health measures.

With its focus on the individual level of political capacities in the social media age, this paper draws attention to the rarely discussed group of policy officials in central ministerial bureaucracies that belong to the executive branch of government but whose tasks, by nature, take place under the primacy of politics. The political component of their tasks will continue to play an important role, as the complex nature of social, economic, and environmental challenges continues to leave policymakers with poorly defined and ambiguous policy problems. Although information about policy alternatives and their consequences may become less incomplete, deciding between competing interests and value conflicts remains at the heart of policy formulation and becomes reflected in the preparatory work of policy officials. In this regard, this study has confirmed the country differences found in earlier studies, particularly those of Italian respondents reporting lower political capacities than the German and Norwegian reference groups (cf. Schnapp, 2004).

As a cross-sectional survey study, this paper cannot infer causal effects. Following research on ministerial bureaucracies that derives potential influence from bureaucrats’ information advantage, a reversed relationship between political capacities and information use is expected to be less likely, yet there might be some circular component in that policy officials with greater confidence in the political components of their tasks seek out and consider more or other information than their colleagues. The consideration of alternative routes could also be linked to the role of administrative traditions. For instance, not being involved in the political processes of policy formulation may cause Italian policy officials to satisfy their need for information via social media more than Norwegian officials who play quite an active political role and whose relationship to the political leadership is marked by high levels of mutual trust. Although SNS use is not necessarily associated with higher social media stress (see Table 3), in Italy the latter was actually higher while their information use was significantly lower. Hence, a greater need for retrieving information via social media combined with fewer opportunities to use that information and a lower efficacy in influencinginfluence policy formulation processes may induce feelings of social media stress. On a similar yet different stress dimension related to the mediatization debate, research could investigate if social media content in the small and homogenous nations of Scandinavia, which have few cross-cutting cleavages and high trust in the government, is comprised of comprise less conflicting or heated societal debates, thus resulting in lower social media stress for policy officials.

Lastly, this study is limited in its assertion on political capacities, which were measured via self-report. However, conceptualization via self-efficacy captures individual role perceptions, which is enlightening from a theoretical point of view. While it may be a valuable path for future studies to approach such research via experiments or panel data, central ministerial bureaucracies are difficult to access, especially when it comes to explaining the genesis and content of policies. Rooted in experimental design but avoiding such practical constraints, implicit mediation analysis which relies on scenarios designs may provide an answer for future research (Bullock & Green, 2021). This paper offers interesting initial insights into the work and information use of policy officials in the social media age, including a granular picture that considers not only the amount but also the variability and concentration of information sources.