Theorizing News Consumption

News Consumption in Belarus and Russia

In seeking to study the role of social rationality in news consumption in autocracies, Guriev and Treisman’s (Reference Guriev and Treisman2022) informational autocracy model again provides a useful entry point. The model distinguishes between two ideal types of autocracies: “fear dictatorships,” which rule through outright coercion and censorship, and “spin dictatorships,” which rule by maintaining an illusion of media choice and avoid overt repression where possible. At the time the field research in this study was conducted, in 2019–20, Belarus occupied a position between these ideal types. While retaining formidable coercive capacity and legal control over media, the regime also relied on a decades-long strategy of authoritarian co-optation of civil society, selectively tolerating apolitical collective action while suppressing overt opposition (Astapova et al. Reference Astapova, Navumau, Nizhnikau and Polishchuk2022). This configuration placed Belarus closer to the “fear” end of the spectrum, but with important caveats regarding the structure of social organization and the media environment. Russia at the time exemplified the “spin” variant of informational autocracy. Since the full-scale invasion of Ukraine in 2022, however, the regime has moved sharply toward fear-based rule, closing nearly all remaining genuinely independent outlets and criminalizing even minimal dissent.

Comparing these cases enables us to examine whether the social-rational logic of news consumption operates differently under these distinct styles of autocratic media control. By juxtaposing a “spin” and a “fear” dictatorship that are otherwise broadly comparable, we can probe whether the role of social rationality in news consumption varies in line with the regime’s mode of media control. Although Russia is considerably wealthier than Belarus and the two countries differ markedly in economic structure and geography, they remain substantially similar on factors likely to shape media consumption, including culture, urbanization, demographic structure, levels of literacy and education, and political and media histories, as well as the language in which the research was conducted.

The Russian and Belarusian media systems are generally understood to be subject to draconian restrictions from the state—but in different ways, and with different outcomes. Freedom House, a widely cited comparative monitor of civil liberties, assigns both countries the lowest possible score on its media freedom indicator, reflecting pervasive legal, political, and economic constraints on independent journalism (Freedom House 2021a; 2021b). In Russia, these constraints are enforced through a broad and flexible legal architecture—including extremism legislation, “foreign agent” laws, and criminal penalties for “discrediting the army”—which gives authorities wide discretion to suppress unwanted speech; attacks, arrests, office raids, and threats against journalists are common (Freedom House 2021b). In Belarus, repressive control is exercised more directly through a highly restrictive media law, widespread criminal prosecution of journalists, and the systematic designation of independent outlets as “extremist organizations,” with journalists subject to fines, detention, and imprisonment for their work (Freedom House 2021a). While the technical legal instruments differ, they serve similar repressive ends. In both cases, the analysis here focuses on the overall level of constraint imposed on media rather than on the specific statutory mechanisms through which repression is implemented. Nonetheless, independent and opposition-minded media outlets and journalists persist in both countries.

After a period of increased pluralism in the 1990s, Vladimir Putin presided over the gradual reestablishment of the Russian state’s influence—and eventually more or less direct control—over the country’s media (Becker Reference Becker2004). Directly or through proxies, the Kremlin controls virtually all of the country’s terrestrial broadcast media, most of the print media, and an increasing share of online media, as well as the core of the advertising market. By the time of Russia’s initial invasion of Ukraine in 2014, Russia’s media system was adapted to an authoritarianism that operated primarily via tacit power, rather than through command and control (Gehlbach Reference Gehlbach2010).

There was a clear relationship between the relative dominance of state television in Russia and electoral outcomes, as both individuals and communities exposed to more state messaging (and less independent messaging) were more likely to vote for Putin and the ruling United Russia party (Enikolopov, Petrova, and Zhuravskaya Reference Enikolopov, Petrova and Zhuravskaya2011). But such state-controlled media were understood to extend the Kremlin’s power not through the monomaniacal exposition of the party line, but by a more careful engagement, designed at once to maintain audience attention, implant convenient interpretations of current events, and prevent the emergence of inconvenient interpretations (Lankina, Watanabe, and Netesova Reference Lankina, Watanabe, Netesova, Bunce, Koesel and Weiss2020; Oates Reference Oates2016; Schimpfossl and Yablokov Reference Schimpfossl and Yablokov2014). Indeed, the Russian case was the archetype on which the “spin dictatorship” model was developed. Since the invasion of Ukraine in February 2022, however, the Kremlin has shifted toward fear-based control—closing nearly all remaining independent outlets, criminalizing “discrediting the army”—even as state television retains elements of its prewar audience-engagement strategies.

Unlike in Russia, state control of the media in Belarus never truly subsided in the post-Soviet period, and the state under Alexander Lukashenko has consistently sought to maintain direct control not only of the messengers, but also of the messaging. With the advent of media that the regime could not directly control, however, the government has expanded its legal authority to hem in independent journalists as well (Herasimenka Reference Herasimenka2016). Yet this pattern of tight media control coexisted, until 2020, with a distinctive form of authoritarian co-optation of civil society. As Astapova and colleagues (Reference Astapova, Navumau, Nizhnikau and Polishchuk2022) show, the Lukashenko regime redirected social energy into formally apolitical associational life, from welfare provision to cultural preservation, thereby sustaining dense horizontal networks under conditions of political demobilization. Following the mass protests of 2020, Belarus further intensified its reliance on coercion, designating major independent outlets as “extremist organizations,” imprisoning or exiling most leading independent journalists, and prosecuting even minor acts of online dissent by ordinary citizens.

The result of this focus on control rather than engagement has been the disengagement of many Belarusians from official media and the emergence, even earlier than in Russia, of a highly polarized media landscape—with state and nonstate media representing not two parts of an interpretive spectrum, but operating in two substantively disparate informational ecosystems with distinct audience structures and mobilizing capacities (Manaev Reference Manaev2014). In both Russia and Belarus, substantial shares of citizens report mixed media diets that cut across state and nonstate outlets. The critical difference lies instead in asymmetric dominance. In Belarus, independent online media became the most trusted and mobilizing outlets by 2020, despite continued state control over broadcast media and print periodicals, while the two media camps covered broadly disparate informational agendas. In Russia, by contrast, state television retained overwhelming audience dominance even as a broader range of alternatives remained accessible, and many media gravitated toward a “gray zone” of interpretation.

Moscow and Minsk, then, pursued divergent approaches to managing the news media—both highly restrictive, but with different degrees of rigidity, and with subtly different outcomes. While opposition media in both countries have largely been limited to the internet, and thus marginalized at least from a technological perspective, independent media seem to have been more effectively marginalized in Russia, where restrictions are more flexible, and where there is somewhat more fluidity in messaging between online and offline media (Smyth and Oates Reference Smyth and Oates2015). Indeed, by the time Belarus erupted into protest in August 2020, online independent media were in some respects more influential than the state-controlled television, radio, and print media (Greene Reference Greene2022). The comparison of Belarus and Russia, then, provides an opportunity to test the importance of social-rational factors in shaping media consumption. While remarkably similar in terms of their political regime as well as their culture and language, Belarus and Russia nonetheless have differently structured authoritarian media systems and patterns of media consumption.

News-savvy readers will have noted that the field research for this study was conducted at moments of significant political mobilization—in August 2019 in Russia, and September 2020 in Belarus. Both countries were in the midst of large-scale political protest waves at the time the data were collected. The implications of this context will be explored in more detail later. For now, however, it is important to provide a brief review of the circumstances.

Events in Russia began to heat up in the summer of 2019, as preparations began for local and regional elections across the country. After election officials blocked opposition leader Alexei Navalny from running against Putin in the 2018 presidential election, the opposition turned its attention to contesting smaller-scale elections, including through a new “smart voting” system devised to provide support to a coalition of candidates in disparate races. The biggest prize, however, was Moscow city council, the legislature of the richest—and most liberal—constituents of the federation. Navalny and his allies thus focused their attention on selecting the most viable candidates and mobilizing nationwide support (Dollbaum Reference Dollbaum2020).

In July of 2019, however, local electoral officials blocked two of the most prominent opposition candidates—Dmitry Gudkov and Lyubov Sobol—from running, through tactics that incensed opposition supporters in the city and elsewhere (Mukhametshina Reference Mukhametshina2019). The opposition responded by organizing a series of protests, which were banned by the authorities but nonetheless garnered significant turnout and attention from independent and international media. These culminated on August 3, 2019, and the police met the protesters with a level of force not previously seen at Moscow protests, arresting more than a thousand and sending dozens to the hospital (Dozhd 2019; Meduza 2019; Roth Reference Roth2019).

The protests that erupted in Belarus in August of 2020 were, at least ostensibly, sparked by similar grounds to the Russian protests of 2019: an attempt by the state to skew the electoral field. Officials banned leading opposition candidates—including Sergei Tikhanovsky and Viktar Babaryka—from running to challenge Lukashenko, and went as far as to arrest Tikhanovsky, Babaryka, and Babaryka’s son, among others. This deepened a sense of public resentment, fueled in large measure by corruption, economic malaise, and the government’s mishandling of the COVID-19 pandemic (Krawatzek and Rohava Reference Krawatzek and Rohava2020). As the election approached, momentum built behind the candidacy of Svetlana Tikhanovskaya, Sergei Tikhanovsky’s spouse.

While the August 2020 election served as the immediate trigger for mass mobilization, the protest wave reflected a far longer accumulation of grievances and organizational capacity. Recent work emphasizes the gradual crystallization of a new civic and national identity, the expansion of horizontally networked civil society, and the erosion of regime legitimacy well before the election campaign began. Onuch and Sasse (Reference Onuch and Sasse2022) show that protest participation was closely associated with durable value orientations and geopolitical self-positioning rather than short-term economic or epidemiological shocks alone. The regime’s co-optation of civil society had also unintentionally generated an associational infrastructure that could be rapidly repurposed for political action once repression escalated. As Astapova and colleagues (Reference Astapova, Navumau, Nizhnikau and Polishchuk2022) argue, this “exit into civil society” during the 2010s created precisely the organizational capacities that enabled the scale and persistence of mobilization after August 2020.

As a result, when the votes were counted and Lukashenko claimed victory with an alleged 80% of the vote, opposition supporters rejected the result and pointed to independent monitoring reports that suggested Tikhanovskaya had in fact won in a landslide. Tens and then hundreds of thousands of Belarusians—representing a remarkably broad demographic and socioeconomic spectrum—took to the streets in Minsk and other cities across the country. Mass street mobilization peaked in August–September 2020 and declined sharply by late autumn as repression intensified. Smaller-scale resistance persisted into 2021 primarily in dispersed, clandestine, and digital forms, aided by a combination of online reporting and mobilization and offline networks of mutual assistance, despite a violently repressive response from the state (Buzgalin and Kolganov Reference Buzgalin and Kolganov2021; Herasimenka et al. Reference Herasimenka, Lokot, Onuch and Wijermars2020).

Analytical Strategy

To test the core proposition advanced in this study—namely, that media consumption is driven powerfully by social-rational considerations—I draw on two original surveys, conducted in Belarus and Russia. The Russian survey was conducted in August 2019, involving 1,206 respondents. The Belarusian survey was conducted in September 2020, involving 1,276 respondents. Both surveys were administered online, and samples were stratified to approximate national representativeness. While neither sample is fully representative in a strict probabilistic sense, the analysis does not rest on point estimates of population parameters. Instead, the focus is on the structure of relationships among key variables and their interactions, which are less sensitive to modest deviations from representativeness.

The surveys—which used largely identical instruments, save for some country-specific questions asked after the questions discussed here—included batteries designed to capture media consumption at a higher level of granularity than is typical in political surveys. Respondents selected the media outlets they use regularly from closed lists of prominent state-linked and independent outlets—15 in Russia and 11 in Belarus—and reported their frequency of use. These responses allow for two distinct dimensions of media consumption.

Media breadth is operationalized as the count of outlets that a respondent uses on a regular basis. Media intensity is operationalized as the proportion of those regularly used outlets that the respondent reports consuming on a daily basis. Together, these measures distinguish between the extensiveness of a person’s media diet and the depth of their engagement with that diet.

To operationalize decision-rational considerations, the surveys include a composite measure of trust in formal media. Respondents were asked to indicate, on a five-point scale, the degree to which they agreed that different classes of media—print, television, radio, and online—are reliable sources of information when making a range of decisions in personal, professional, material, and political domains. These items were compiled into an index of generalized media trust. Rather than asking directly about “trust,” which is subjective and prone to misinterpretation, this approach allows for a utility-related measure of respondents’ orientation toward the media. These answers were compiled into an index of “media trust.”

This medium-level, rather than outlet-specific, trust measure is used for two reasons. First, it aligns with the decision-rational literature, which conceptualizes trust in the media as a general informational resource that lowers the perceived costs and raises the expected benefits of information acquisition. Second, it ensures comparability across Russia and Belarus, where the structure and brand ecology of media outlets differ substantially. A strict “horse race” between outlet-specific trust and socially embedded co-use is therefore not feasible within a fully comparable design. This represents a limitation of the present study and an important avenue for future work.

Social-rational considerations are operationalized through two measures capturing the social embedding of media use. First, respondents were asked to reflect on the frequency with which they converse about the news with friends, family, and colleagues, across a range of topics, including work, economics and finance, politics, pop culture, sports, food, and so on. This produces a measure of interpersonal discussion, capturing the conversational context of media consumption.

Second, the surveys include a novel measure of “co-use,” defined as the proportion of a respondent’s regular media outlets that the respondent believes are also consumed by members of their social circle. Respondents separately assessed the media use of friends, family members, colleagues, neighbors and online contacts. Co-use is calculated as the following:

$$ C=\frac{\mid {M}_R\cap {M}_S\mid }{\mid {M}_R\mid } $$

where C denotes “co-use,” M_R is the set of outlets the respondent reports using, and M_S is the set of outlets the respondent attributes to their social circle.Footnote ¹

Importantly, this measure is based on perceived rather than objectively verified patterns of others’ media use. This is not a measurement flaw for present purposes. The theoretical claim concerns what respondents believe about their social environment, since it is perceived shared consumption that structures conversational incentives and reinforces social alignment. Misperceptions about others’ media diets, to the extent that they exist, are therefore part of the substantive mechanism under study.

The survey also included standard questions on age, gender, levels of education, income, and geographical location.

In addition to these core measures, the surveys measured the Big Five personality traits. These traits have been shown to shape a wide range of social preferences, orientations, behaviors, and political outcomes (John and Srivastava Reference John, Srivastava, Pervin and John1999; Mondak Reference Mondak2010).

From the standpoint of social rationality, personality traits provide a way to assess whether socially embedded media behaviors reflect stable psychological predispositions rather than purely situational adaptation to the surrounding political environment. Prior work shows that openness and neuroticism correlate with media breadth (Gerber et al. Reference Gerber, Huber, Doherty and Dowling2011), while extraversion, neuroticism, openness, and agreeableness condition individuals’ propensity to discuss news, depending on the level of conflict associated with the subject matter (Gerber et al. Reference Gerber, Huber, Doherty and Dowling2012). Agreeableness has also been shown to shape attention to state media and evaluations of regime leadership and policies in Russia (Greene and Robertson Reference Greene and Robertson2017).

Relationships between the core independent variables (media trust, discussion, and co-use), the dependent variables (media breadth and media intensity), covariates (age, gender, education, and the Big Five traits) are evaluated through ordinary least squares regressions. Alternative analyses are presented in online appendix B.

In addition to direct effects, the theory also implies a sequential mechanism linking co-use, interpersonal discussion, and media intensity. Specifically, higher co-use is expected to increase opportunities and incentives for interpersonal discussion, which in turn should intensify daily engagement with existing media outlets. This implies a mediation structure: co-use → interpersonal discussion → media intensity. In addition, the theory anticipates that this indirect pathway may be conditioned by personality—particularly agreeableness, which is associated with conflict avoidance and sensitivity to social harmony. In Russia, where media environments are more interpretively fluid but still socially polarized, agreeable individuals may translate shared media environments into discussion more readily than less agreeable individuals. This yields a moderated mediation expectation.

To evaluate these indirect pathways, the analysis employs Hayes’s PROCESS framework with nonparametric bootstrap standard errors. While the data are cross-sectional, this approach is appropriate for assessing whether observed associations are consistent with theoretically specified indirect mechanisms, without making strong causal claims about temporal sequencing (full mediation analyses are presented in online appendix C).

Belarus and Russia are selected as critical cases of high-control autocracies with dense media ecosystems. Both regimes impose extensive formal and informal constraints on journalism and public communication, while nonetheless sustaining complex, multiplatform media environments. This makes them a hard test of social rationality: if socially embedded incentives for media consumption operate even where institutional control is extensive and political choice is constrained, they should operate at least as strongly in more pluralistic or hybrid settings.

At the same time, these cases define the scope conditions of the analysis. The data capture Russia prior to the full-scale invasion of Ukraine and the subsequent shift to overtly fear-based media control, and Belarus during the peak of postelection mobilization and repression. Implications for wartime Russia, China, and backsliding democracies are therefore addressed in the conclusion rather than generalized directly from the findings.

Discussion and Conclusions

At the outset, this article advanced a theory of social rationality in media consumption, in which individuals are seen to choose media that facilitate social connection, rather than in service of material utility. The aim was not to displace “decision-rational” explanations of media consumption, but to widen the aperture to make room for a more encompassing set of media-related choices. Moreover, it was proposed that the need for a wider aperture was particularly acute in the study of autocracy, where the assumptions of purposiveness and individuality that underpin classical models of media consumption are at their most tenuous.

The data from Belarus and Russia strongly support this theoretical claim, albeit with an important distinction between the breadth and intensity of media use. Breadth is linked to both decision-rational and social-rational factors, with trust emerging as a robust predictor in both countries. Intensity has no significant relationship to trust. Instead, the strongest predictors are measures of social rationality: co-use in both countries and, indirectly, interpersonal discussion. These results suggest that while decision-rational factors help to expand the range of outlets people include in their media diets, social-rational factors drive the depth of engagement with those outlets.

If media consumption is measured as the number of media outlets an individual turns to for news with at least some regularity, decision-rational logic remains salient, but it coexists with the demands of socialization. The belief that formal media—of whatever technology or political stripe—are reliable sources of information is a significant predictor of the number of outlets in respondents’ media diets, but it does not overshadow a person’s propensity to discuss the news with social interlocutors. If, however, we switch the focus from breadth to depth—measuring media consumption as the intensiveness with which the news is consumed—measures of social rationality come to the fore, while the role of trust recedes. The strongest predictor of daily engagement is the respondent’s tendency to consume the same news outlets as her social circle. In other words, the biggest news junkies in the surveys were those whose news consumption was most closely linked to their social lives.

Whether one prefers breadth or depth as the relevant indicator, the implication is clear: socialization features powerfully in the constellation of factors shaping media choice. This is not the same as saying that socialization causes media choice; the relationship is likely reciprocal, ongoing, and mutually reinforcing. Media and socialization choices are not a one-shot game. While one’s social circle may be more stable than one’s media diet, neither is fixed. If an important part of my social life revolves around talking about the news, I am likely to spend more time consuming news that supports enjoyable conversation with my usual interlocutors than news that puts me at odds with them (unless I am particularly disagreeable). Likewise, I may gravitate toward social ties who share my media habits to avoid disrupting existing conversational routines. Over time, such reciprocal causation can lock in both social networks and media repertoires, with decision-rational utility playing little role in sustaining high-intensity use. What matters, instead, is sustaining the flow of mutually reinforcing conversation and consumption.

The mediation models sharpen this picture. In both Belarus and Russia, co-use significantly predicts interpersonal discussion, which in turn predicts higher media intensity. The direct effect of co-use remains significant after accounting for discussion, indicating partial mediation. This suggests that co-use influences intensity not only by making content more socially salient, but also by stimulating the interactions in which that relevance is generated.

The moderated mediation results reveal sharply different dynamics in the two authoritarian contexts. In Belarus, the mediated effect of co-use on intensity through discussion is positive and significant across all levels of agreeableness, suggesting that the incentive to form and maintain communities of shared media consumption overrides intrinsic differences in prosocial orientation (cf. Greene Reference Greene2022). In Russia, however, agreeableness significantly moderates this pathway. Among highly agreeable citizens—those most oriented toward social harmony—intensive media engagement seems to be already sustained by their embeddedness in regime-aligned social networks (Greene and Robertson Reference Greene and Robertson2017), leaving co-use to have little additional effect. By contrast, for less agreeable Russians, co-use provides an extra impulse, integrating them into relevant conversations and driving more intensive media engagement.

These personality effects suggest that dispositions toward sociability, harmony, and sensitivity to uncertainty—shaped by factors deeper than, and thus causally prior to, political orientations—nonetheless play out in context-specific ways. In Russia, openness and neuroticism are both associated with broader media diets, while agreeableness predicts narrower repertoires; in Belarus, none of these traits significantly predict breadth, and agreeableness instead shows a weak positive association with intensity. The moderated mediation results further imply that in Russia, agreeableness dampens the role of co-use in stimulating news discussion, while in Belarus this damping effect disappears under conditions of mass mobilization. In both countries, however, the link between social interaction and intensive news use remains robust, underscoring that media engagement serves a social purpose even when the psychological routes to that engagement differ.

These findings do not overturn the logic of informational autocracy as developed by Guriev and Treisman (Reference Guriev and Treisman2019; Reference Guriev and Treisman2022). Instead, they suggest that informational autocracy operates through a social-rational microfoundation that remains undertheorized in existing formulations. Decision-rational accounts correctly emphasize how biased media can shape beliefs about regime competence and performance. The evidence presented here refines this account by showing how citizens enter, remain within, and intensify their engagement with the information stream in the first place. Social rationality helps to explain why citizens continue to consume biased or distrusted media even when its instrumental informational value is limited: doing so sustains shared conversational worlds, preserves social belonging, and maintains routine participation in collective meaning making. In this sense, social rationality specifies the mechanism through which informational autocracy gains traction and explains why regime propaganda often produces ambivalent, inconsistent, or “minimal” attitudinal effects rather than straightforward persuasion.

The findings reported here are consistent with a social-rational interpretation of media consumption, but the research design does not permit strong causal claims. Four major causal threats warrant explicit acknowledgment. First, reverse causation may run from high media use to higher perceived co-use: individuals who consume a great deal of news may infer that others do so as well. Second, intensive media use may itself generate more frequent interpersonal discussion rather than the reverse. Third, unobserved political interest may confound all three processes—media use, co-use, and discussion—simultaneously. Fourth, ideology and homophily may jointly structure media choice and social networks, producing the appearance of social coordination where ideological sorting is the deeper driver. The cross-sectional nature of the data precludes full adjudication among these possibilities. Accordingly, the results are best interpreted as documenting stable and theoretically coherent associations that are consistent with social-rational mechanisms, rather than as definitive causal proof of those mechanisms.

Future research can move beyond these limitations in several ways. First, survey experiments could exogenously manipulate information about friends’ or peers’ media choices and observe whether such cues alter respondents’ own consumption and political interest. Second, panel designs tracking the same individuals over time could directly test reciprocal causation between co-use, discussion, and media intensity. Third, field experiments embedded in digital platforms or messaging networks could introduce experimentally controlled variation in socially visible news exposure, allowing researchers to measure how changes in the social informational environment propagate through discussion and engagement. Together, these designs would provide the leverage needed to more cleanly identify the causal pathways implied by social-rational theory.

Several scope conditions bear on the interpretation of these findings. First, the data were collected during periods of heightened political mobilization in both countries. Under such conditions, socially embedded mechanisms of media use are likely to be especially visible. In calmer periods, the social payoffs to news consumption may be less intense, though there is little reason to expect them to disappear altogether. Second, while Belarus and Russia represent tightly controlled media environments, the logic of social rationality is unlikely to be confined to such contexts. In China, in newer autocracies, and in backsliding democracies where formal media pluralism coexists with intense social pressure, socially embedded media use may be at least as important in structuring engagement. Third, in wartime Russia, the regime’s shift toward overt fear-based control likely intensifies outright censorship while also pushing some citizens toward underground, diaspora, or encrypted networks of socially coordinated information exchange. These dynamics highlight the durability, rather than the fragility, of social-rational mechanisms under conditions of extreme coercion.

The findings carry clear implications for policy and practice. If media consumption is socially embedded, then interventions that focus narrowly on breaking censorship, expanding outlet access, or correcting factual misinformation will often fall short. What matters is not only what information is available, but also whether it can be integrated safely and meaningfully into existing social relationships. For independent journalists, international broadcasters, and digital platforms, this implies that content strategies should prioritize shareability within trusted social networks rather than simply maximizing reach. Information that can be circulated without imposing high social or political risk—and supports everyday conversational use—may travel farther than content that demands overt political signaling. Similarly, reaching new audiences is likely to be enabled more by building social embeddedness within untapped communities than by projecting greater factual accuracy.

More broadly, external efforts to promote pluralism in authoritarian settings may be more effective if they help to foster alternative communicative communities, rather than only alternative media outlets. Nongovernmental organizations, diasporic media producers, and digital intermediaries can play a role in cultivating socially safe spaces for shared interpretation, humor, cultural content, and soft political discussion that gradually expand the scope of permissible collective meaning making. In this sense, the most effective challenges to authoritarian media control may come not from frontal attacks on censorship alone, but from the slow construction of parallel social infrastructures of information sharing.

For scholars of Russia and Belarus, these findings reinforce the “coconstruction” model of authoritarian power (Greene and Robertson Reference Greene and Robertson2019) and broader accounts of the social foundations of post-Soviet authoritarianism (Forrat Reference Forrat2018; Morris Reference Morris2016; Sharafutdinova Reference Sharafutdinova2020). The fact that similar behavioral patterns yield divergent political outcomes—including different structures of media consumption and different scales of opposition—suggests that these outcomes are not products of contrasting political cultures. They also imply that political change will not come from simply breaking censorship: if citizens choose media for the socialization it enables, shifts in informational content must also sustain, rather than disrupt, those social relationships.

For authoritarianism studies more broadly, the results suggest that citizens’ support for regimes in “informational autocracies” may hinge less on the informational content of the media than on the tenor and structure of interpersonal conversations. Because such conversations are harder to monitor or censor than formal media channels, they represent a potential point of vulnerability for regimes—as Belarus’s 2020–21 protest cycle demonstrated. In Russia, however, the moderated mediation findings indicate that among highly agreeable citizens this conversational pathway is muted, potentially buffering the regime against socially driven dissent.

Finally, while the analysis here is confined to two authoritarian contexts, the argument carries implications for comparative media research. Models of political behavior in democracies have increasingly incorporated social influence, contagion, and pressure (Bond et al. Reference Bond, Fariss, Jones, Adam, Marlow, Settle and Fowler2012; Gerber, Green, and Larimer Reference Gerber, Green and Larimer2008; Nickerson Reference Nickerson2008). Media consumption models should follow suit—treating social rationality not as a distortion of decision-making rationality, but as an independent causal force. Authoritarian contexts offer a particularly demanding stress test for such models: if social rationality organizes media use even where informational returns are thin, models of media and politics that omit social incentives risk systematic misattribution by treating community maintenance as noise rather than as the signal.