1370943 research-article2025 CRXXXX10.1177/00936502251370943Communication ResearchHoffmann and Boulianne 11) Check for updates Original Research Article The Role of Political Interest in the Relationships Between Privacy Concerns, Social Network Size, and Political Expression on Facebook, Twitter, and Instagram 0 0 0 Christian Pieter Hoffmann1 Communication Research 1­–33 © The Author(s) 2025 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/00936502251370943 https://doi.org/10.1177/00936502251370943 journals.sagepub.com/home/crx Ssage e and Shelley Boulianne2 Abstract Studies on the predictors of using social media for political purposes reveal some unexpected complexities: users often disregard institutional privacy concerns to discuss politics online, and the size of social networks positively correlates with political expression on social media. Building on the privacy calculus theory, we explore how political interest interacts with privacy concerns and social network size when users decide to engage in political expression on social media. This study utilizes survey data from four countries (the US, UK, France, and Canada) collected in 2019 (n = 6,291), encompassing three social media platforms: Facebook, Instagram, and Twitter. We find that privacy concerns are negatively related to expression on social media. Larger social networks positively relate to political expression, especially on Twitter. Political interest plays an important moderating role: highly politically interested users discount privacy concerns and opt to post political content. These findings replicate across all three platforms. Keywords social media, political expression, political interest, privacy, social network 1 University of Leipzig, Leipzig, Germany Mount Royal University, Calgary, Canada 2 Corresponding Author: Christian Pieter Hoffmann, Institute of Communication and Media Studies, University of Leipzig, Nikolaistrasse 27-29, Leipzig 04109, Germany. Email: christian.hoffmann@uni-leipzig.de 2 Communication Research 00(0) Introduction In the digital age, the most common forms of online political behavior are expressive, as users engage in communicative actions such as posting or commenting on content (Lane et al., 2022; Ruess et al., 2023). Both vertical/institutional and horizontal/social privacy considerations influence users’ willingness to engage in online political expression (e.g., Dal, 2024; Quinn et al., 2019). Building on the privacy calculus theory (Dinev & Hart, 2006), this study examines two predictors previously shown to relate to political expression in unexpected ways: institutional privacy concerns (vertical dimension) and social network size (horizontal dimension). We highlight the role of a stable personal characteristic, political interest (Prior, 2010; Russo & Stattin, 2017), in moderating the relationships of institutional privacy concerns, network size, and political expression across three social media platforms in four countries. Political interest helps to resolve some apparent puzzles in the research on privacy and online political expression. While privacy research suggests that institutional privacy concerns (in relation to platforms) should inhibit online political expression (Ahmed & Lee, 2025; Dienlin et al., 2023), several studies have found a positive correlation (Best & Krueger, 2011; Hoffmann et al., 2015; Hoffmann & Lutz, 2023). Similarly, arguments derived from research on “context collapse” (D. Boyd, 2006; Marwick & Boyd, 2011) or the “spiral of silence” (Sheehan, 2015; Stoycheff, 2016; Zerback & Fawzi, 2017) indicate that large social networks would inhibit political expression (Schulz, 2025; Weeks et al., 2024). Yet, a meta-analysis found that self-censorship is more prevalent in political discussions about obtrusive issues with known others, rather than mere acquaintances or even strangers (Matthes et al., 2018). In fact, large networks—both on- and offline— positively relate to political engagement (Eveland & Hively, 2009; Gil de Zúñiga & Valenzuela, 2011; Weeks et al., 2024). To account for these counterintuitive and conceptually challenging findings, we focus on the role of political interest. Based on the privacy calculus theory (Dinev & Hart, 2006), we argue that highly politically interested users are more likely to value the benefits of political expression on social media and are willing to discard potential risks or disadvantages (i.e., privacy concerns). Examining the role of political interest through the lens of the privacy calculus is important because past research has shown that political interest is unevenly distributed throughout the population (Denny & Doyle, 2008; Van Deth, 1990), and that it tends to be a relatively rare but stable individual characteristic (Prior, 2010; Russo & Stattin, 2017). Our analysis is based on a large-scale cross-sectional international survey of Internet users in the US, UK, Canada, and France, examining three major social media platforms: Facebook, Twitter, and Instagram. We heed recent calls for more comparative research on privacy (Masur et al., 2025). Citizens use different platforms for political activities (Ruess et al., 2023), albeit most of the research on political expression on social media does not consider the platform (Lane et al., 2022). When it does, the focus has been on Facebook (Lane et al., 2022). We examine the platform-specific privacy and networking dynamics that shape political expression on Twitter, Instagram, Hoffmann and Boulianne 3 and Facebook. A platform-specific perspective is crucial, given the variation in privacy settings across these platforms and the differences in their network structures (Bossetta, 2018; Evans et al., 2017). Yet, our study highlights that a personal characteristic, such as political interest, can play a consistent role in privacy considerations across these contexts. We address the following research questions: RQ1: What is the role of political interest in moderating the relationships between privacy concerns, social network size, and political expression on Facebook, Twitter, and Instagram? RQ2: How do these relationships vary by platform? Literature Review Political Expression on Social Media and Political Interest In the context of social media, political expression has been defined as “behaviors that involve communication of one’s political views, beliefs, or identities to others” (Lane et al., 2022, p. 5). However, not all social media platforms may be equally conducive to political expression. Mitchelstein et al. (2021) find that Twitter is regarded as more of a “political” platform, Instagram as more of a “non-political” platform, and Facebook as being somewhere in between. The Digital News Report (Newman et al., 2024) shows that across 47 markets, 26% of users use Facebook for news, 14% use Instagram, and 11% use Twitter. Boulianne et al. (2024) find that across different Western markets, of those using each platform, a sizeable share post political content: 33% on Facebook, 35% on Instagram, and 43% on Twitter. In the United States, Instagram users tend to be more left-leaning, while Facebook is used across the political spectrum (Vogels et al., 2021). Before Twitter’s takeover by Elon Musk, left-leaning users were more likely to post political content on Twitter, but that may have changed under the platform’s new ownership (McClain et al., 2024). Privacy research has highlighted that obstacles to online political expression can be both institutional and social in nature (Quinn et al., 2019). Using social media for politics necessitates sharing data with platform providers (Hoffman et al., 1999; McKnight et al., 2002) and exposing oneself to potential institutional surveillance (Dencik & Cable, 2017). In addition, political expression on social media is inherently social. It constitutes social interactions as users are exposed to or even strive for the attention of others, attempting to generate interest, mobilize, and/or persuade (Lane et al., 2022). Individuals tend to mitigate social risks by engaging in online political conversations with like-minded others rather than seeking out cross-cutting discussions (Wojcieszak & Mutz, 2009). Yet, as discussed below, the roles of institutional privacy concerns and large social networks in potentially inhibiting political expression on social media are insufficiently understood. While some institutional and social influences may impede political expression on social media, some user characteristics positively predict it. Political interest is 4 Communication Research 00(0) a key predictor of online political expression (Bimber et al., 2015). It can be defined as the “degree to which politics arouses a citizen’s curiosity” (Van Deth, 1990, p. 278). Political interest is conceptually distinct from political behavior. Political interest develops during adolescence but then attains a trait-like stability over time (Prior, 2010; Russo & Stattin, 2017). Van Deth (1990) points out that interest also connotes advantage or profit. Political interest motivates political involvement, as politically interested individuals derive subjective or objective gratification from engaging in politics. Kim et al. (2021) find that individuals who frequently comment on politics on Facebook tend to be highly interested in politics and more politically polarized. Focusing on moralizing political talk, Grubbs et al. (2019) find that status-seeking personality traits bolster users’ propensity to engage in online political interactions, particularly with members of the political outgroup. Status-seeking, in turn, is correlated with political interest (Bor & Petersen, 2022). Similarly, Boulianne and KocMichalska (2022) highlight that extraverted individuals are more likely to engage in political discussion online, both generally and in cross-cutting discussions. Extraversion is also positively related to political interest (Denny & Doyle, 2008). Politically interested individuals, thus, are uniquely motivated to engage in political expression and seek out political exchanges with others online. In this study, we apply a privacy calculus lens to examine how political interest interacts with both institutional and social privacy considerations across platforms, and how it may contribute to resolving some conceptual puzzles when it comes to the role of privacy in political expression on social media. Privacy Concerns Privacy concerns are based on the assessment of the likelihood and extent of adverse consequences from information disclosure (Dinev & Hart, 2004; Malhotra et al., 2004). On social media, when users post content, they share personal data with platform providers and with other users (Dienlin & Trepte, 2015; Young & Quan-Haase, 2013). Using social media for politics—professing a political stance, supporting political causes, and reaching out to others to inform and mobilize—is associated with norms that encourage self-expression, resulting in both vertical and horizontal privacy risks. Vertical privacy risks refer to risks emanating from (commercial or government) institutions. For political expression, especially, the political system is a key contextual influence on vertical (or institutional) privacy concerns (Masur et al., 2025). Various studies examine how users engaging in political expression in authoritarian contexts struggle to safeguard their privacy—for example by restricting the visibility of their content or profiles (Lokot, 2020; Mak et al., 2024; Pearce et al., 2018). The present study focuses on privacy concerns in relation to platform providers within Western democratic contexts. These concerns are informed by public discourse, the salience of specific platforms, recent privacy breaches, or current policy discussions (Miltgen & Peyrat-Guillard, 2014; Morrow, 2022; Turow et al., 2018). Facebook, for example, has repeatedly been the subject of well-reported privacy breaches, such Hoffmann and Boulianne 5 as the Cambridge Analytica Scandal. Instagram was recently fined 405 million euros by the European Union for mishandling the personal data of minors (European Data Protection Board [EDPD], 2022). After Elon Musk’s acquisition of Twitter (now “X”), the US Federal Trade Commission launched an investigation into the platform’s privacy policies. In a recent survey of US citizens, Business Insider (2022) found that few people trust social media companies to protect their privacy and data, with a steady decline for all platforms since 2020. Facebook emerged as the least trusted platform (18%), followed by Twitter (23%), with only slightly more survey participants trusting Instagram (25%). Platforms differ in their privacy affordances or architectures, such as visibility, identifiability, searchability, and persistence (Kakavand, 2024). Bossetta (2018) notes that on Twitter, user posts tend to be visible and openly accessible by default. At the time of data collection, this was true, but recently, X has changed, requiring a sign-in to access posts. On Facebook, personal profiles are visible to friends or followers by default but can be made openly accessible. Profiles on Instagram and public profiles on Facebook are set to be open by default. All three platforms require some personal data for sign-up, but X and Instagram allow the maintenance of pseudonymous profiles (Boulianne et al., 2024). Facebook and Instagram offer the option to post ephemeral content as a “story.” Privacy concerns do not entirely preclude online interactions (Dienlin et al., 2023; Hoffman et al., 1999; McKnight et al., 2002). Instead, in a “privacy calculus” users weigh the privacy risks of an online interaction against its expected benefits (Dinev & Hart, 2006; Trepte et al., 2017). As a result, individuals apply a variety of tactics to limit online self-disclosure while still reaping the benefits of online interactions (Dinev & Hart, 2006). The privacy calculus posits a cost-benefit trade-off in users’ exercise of privacy practices (Dienlin & Metzger, 2016; Krasnova et al., 2012). A number of user characteristics can affect this calculus, such as users’ privacy literacy or values (cf. Kezer et al., 2022; Trepte et al., 2017). In general, however, heightened risk perceptions should be associated with privacy protective behavior, such as reduced self-disclosure (Dienlin & Metzger, 2016; Krasnova et al., 2012). Ahmed and Masood (2025), as well as Ahmed and Lee (2025), find that privacy concerns relate negatively to online political participation in both Western and non-Western countries. We therefore propose a first hypothesis: H1: Privacy concerns will be negatively correlated with political expression on Facebook, Twitter, and Instagram. However, some evidence points to a “privacy paradox” in the context of online politics. The “privacy paradox” has evolved to denote an intrapersonal discrepancy between stated privacy concerns and a lack of commensurate privacy protection behavior (Dienlin & Trepte, 2015; for criticism of the concept see Dienlin et al., 2023). In a study of German Internet users, Hoffmann et al. (2015) found that institutional privacy concerns exert a weak but positive effect on political content creation. Best and Krueger (2011) found that anxiety about monitoring decreased and feeling anger 6 Communication Research 00(0) increased political participation among US respondents. This pattern was replicated using single-item measures for contacting officials, signing petitions, donating, and persuading others; the latter is most relevant for our study about political expression. Blank (2013) finds that, among Internet users in the UK, the level of comfort with disclosing personal information online is not related to online political expression. As such, there may be cross-national differences in the role of privacy concerns on online political expression. None of these studies explored differences between platforms. Hoffmann and Lutz (2023) argue that the effect of institutional privacy concerns on online political engagement may differ by the effort required for online political action, as those especially high in political interest may be drawn to high-threshold forms of political behavior. Given previous findings on the characteristics of individuals engaging in online political expression, we propose that political interest may play a significant role in the relationship between individuals’ privacy concerns and their political expression on social media. Highly politically interested individuals, who also tend to be more partisan, extraverted, and polarized (Davis & Dunaway, 2016; Rogowski & Sutherland, 2016), may discount privacy risks when weighing them against the perceived benefits of engaging others in political talk (cf. Bimber et al., 2015). The perceived benefits of online political expression may be especially pronounced for those who are highly interested in politics. We thus expect the negative relationship between privacy concerns and political expression to grow weaker the higher a user’s level of political interest (contributory moderation; Holbert & Park, 2020). H2: Political interest will moderate (i.e., weaken) the negative relationship between privacy concerns and political expression on Facebook, Twitter, and Instagram. Social Networks Network size appears to play a central role in the social privacy considerations in political expression on social media platforms. Some users prefer to engage with small, politically like-minded audiences, while others reach out to members of the political outgroup and engage in wide-ranging, cross-cutting interactions (Boulianne & Koc-Michalska, 2022; Brady et al., 2021; Kim et al., 2021). On social networking sites (SNS), establishing ties is intertwined with horizontal privacy protection behavior, as tie formation tends to be the basis for information sharing (D. M. Boyd & Ellison, 2007). Young and Quan-Haase (2013) describe “friending practices” as a salient form of privacy protection among young Facebook users. Individuals with heightened privacy concerns may prefer to maintain a smaller network. Social media platforms offer different options for network maintenance and securing horizontal privacy. Facebook, Instagram, and Twitter allow users to maintain private profiles that are only visible to those whose connection requests are accepted (Bossetta, 2018; Boulianne et al., 2024). Some platforms, such as Facebook, allow different audiences for posts. Others, such as Twitter, tend to prompt more public forms of communication by default, with profiles and posts visible. Generally, Twitter can be considered to be composed of more weak ties among acquaintances and Hoffmann and Boulianne 7 strangers (Valenzuela et al., 2018). For Instagram, the size of one’s social network depends on one’s age group; this platform is especially popular among young adults (75% use), who tend to cultivate larger networks than other users (Boulianne & Hoffmann, 2022). Lee and Yuan (2020) compare the formation of ties and privacy concerns among Facebook and Instagram users, finding that higher privacy concerns are associated with a preference for small networks characterized by close ties. They argue that Facebook is perceived as more threatening to users’ privacy due to its large user base. Mitchelstein et al. (2021) distinguish political expression on Twitter, Facebook, WhatsApp, and Instagram. They highlight the differences in the scale of privacy and the degree to which content is public versus private (Mitchelstein et al., 2021). They use this framework to explain how the frequency of Facebook and Twitter usage predicts posting political opinions, but the frequency of Instagram and WhatsApp usage does not. Unfortunately, their study did not include platform-specific measures of posting political content. Research on network characteristics and political expression tends to focus on network composition and size (Zhang et al., 2024), which are often closely related. Larger social networks are commonly associated with politically more heterogeneous networks, as larger networks tend to include more weak ties (cf. Gil de Zúñiga & Valenzuela, 2011) with more diverse viewpoints (Eveland et al., 2013; Eveland & Hively, 2009; John & Dvir-Gvirsman, 2015). Settle and Carlson (2019) find that Internet users tend to shy away from political conversation topics depending on the composition of the group in which they would converse (cf., Weeks et al., 2024). Political unfriending is thus more common among those maintaining larger online networks (Skoric et al., 2018; Yang et al., 2017). Research applying the “spiral of silence” theory to online platforms finds that many avoid discussing contentious issues online for fear of social isolation (cf. Hoffmann & Lutz, 2017; Matthes et al., 2018; Sheehan, 2015; Stoycheff, 2016; Zerback & Fawzi, 2017). Similarly, the “context collapse” concept addresses the challenges arising for privacy management due to deviating social contexts (D. Boyd, 2006; Marwick & Boyd, 2011). Context collapse occurs when users maintain ties embedded in different social contexts based on a single social media profile (Marwick & Boyd, 2011; Vitak, 2012). Context collapse is more likely to occur with larger social networks, and it tends to inhibit online self-disclosure (Vitak, 2012). Across the political spectrum, individuals engage in political self-censorship when engaging with political outgroups (Schulz, 2025). However, Zhang et al. (2024) found that political network diversity, when controlling for size, was positively related to political expression. They argue that large networks tend to offer more opportunities for political conversations. Since larger networks also tend to be more politically heterogeneous, they offer more opportunities for cross-cutting political discussions (Barnidge et al., 2018). Matthes et al. (2018) conducted a meta-analysis of research on the spiral of silence, finding that self-censorship is more likely to occur in conversations with known others compared to those with strangers. A large social network can thus be conducive to political expression as 8 Communication Research 00(0) it contains more mere acquaintances or even strangers (Eveland et al., 2013), which may disinhibit political expression. Indeed, network size has been found to positively relate to political expression, both on- and offline (Eveland & Hively, 2009; Gil de Zúñiga & Valenzuela, 2011; Weeks et al., 2024). We propose the following hypothesis: H3: Network size will be positively correlated with political expression on Facebook, Twitter, and Instagram. As we have discussed above, individuals differ in whether they are attracted to or shy away from political expression (Bor & Petersen, 2022). Some may even enjoy engaging in political controversy; studies suggest that gains in ingroup prestige may motivate engaging with the political outgroup (Brady et al., 2021; Grubbs et al., 2019). Applying a privacy calculus lens, we propose that, again, political interest plays a key role in these relationships, as the politically interested may derive more enjoyment from engagement with larger, potentially more politically diverse audiences. Because political interest is related to partisanship, extraversion, and polarization (Davis & Dunaway, 2016; Rogowski & Sutherland, 2016), these social incentives are likely to be particularly salient among the politically interested. Hampton et al. (2017) find that those with strong issue attitudes are willing to engage in online political discussions even if they expect little agreement from their social network. Highly politically interested individuals may find political expression on social media more beneficial as their social network and, thereby, their audience becomes larger. We thus expect the positive relationship between network size and political expression to grow stronger the higher a user’s level of political interest (contributory moderation; Holbert & Park, 2020). We propose: H4: Political interest will moderate (i.e., strengthen) the positive relationship between network size and political expression on Facebook, Twitter, and Instagram. Figure 1 summarizes the hypotheses for this paper. To address RQ2, we will examine how these relationships vary between Facebook, Twitter, and Instagram. Data and Methods This paper utilizes survey data collected in four countries between September and November 2019 (n = 6,300). The sample is based on an online panel with quotas to ensure representation of the population in each country (sex, age, education). The quotas ensure a near-perfect match (97%–100% weighting efficiencies) between the sample characteristics and the population characteristics. As such, we did not weight the data. Kantar/Lightspeed administered the survey to their online panel: 1,700 people from the United States, 1,542 from the United Kingdom, 1,510 from France, and 1,539 from Canada. Before submitting the data to the researchers, Kantar conducted various quality checks. The quality checks were decided at the pretest stage; these Hoffmann and Boulianne 9 Political interest H2 (+) Privacy concerns H4 (+) ,. . r Hl (-) ,r Network size H3 (+) . r Political expression on Facebook, Twitter, and Instagram (RQ2) Figure 1. Research model. checks included removing respondents who completed the survey too quickly and respondents who provided nonsensical answers to open-ended questions. The survey received ethics approval before data collection (File No. 101662), in accordance with Canada’s Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (TCPS). The dependent variable—political expression on social media—was assessed based on the question “During the past 12 months, how often have you posted this political content on Facebook/Instagram/Twitter?” (never, rarely, time to time, often), posed for each platform used. The question was preceded with an explanation of “political content” as “current events in the world, news about elections, information about political figures, information about government performance, debates about public policy, and other political issues,” reflecting advice from Guess et al. (2019) about how to improve the accuracy of recall on political posting. Lane et al. (2022) discuss feature-focused measures of political expression. They report that 62% of measures of political expression focus on “posting,” 50% include “sharing/reposting,” 41% focus on “liking,” and 39% examine commenting. Political posting on social media platforms is rare (Table 1; Guess et al., 2019). Political interest was assessed based on responses to a single-item question (“How interested would you say you are in politics?” (Response scale 1–4), mirroring the question used in the World Values Survey. We measured institutional privacy concerns based on the level of concern related to the following topics (Response scale 1–5): (a) Social media companies insufficiently protecting personal data (information leakage); (b) Social media companies sharing personal data with government agencies; and (c) Social media companies selling personal data to third parties, such as political groups 10 Communication Research 00(0) Table 1. Descriptive Statistics. Variables Responses Average (SD) or % Females Education (high school or less; some colleague, Bachelor’s degree, more than Bachelor’s degree) Age in years Center/moderate/neither left nor right Left-wing Right-wing Political efficacy (α = .741) Privacy concerns (α = .905) Political interest Facebook network size: 0–30 (26%), 31–100 (25.4%), 101–200 (20.4%), 201–400 (15.2%) and more than 400 (9.7%) Facebook exposure to political information Facebook posting of political content Twitter network size: 0–15 (43.3%), 16–100 (25.4%), 101–200 (14.0%), 201–400 (8.6%), and More than 400 (8.7%) Twitter exposure to political information Twitter posting of political content Instagram network size: 0–15 (34.3%), 16–100 (25.9%), 101–200 (16.9%), 201–400 (8.6%), and more than 400 (11.5%). Instagram exposure to political information Instagram posting of political content 0.1 1–4 51.24% 1.94 (1.05) 18–100 0.1 0.1 0.1 1–4 1–5 1–4 1–5 48.45 (17.32) 51.32% 21.04% 27.64% 2.44 (0.70) 3.46 (1.12) 2.61 (0.97) 2.64 (1.36) 1–4 1–4 1–5 2.50 (1.09) 1.65 (0.98) 2.14 (1.30) 1–4 1–4 1–5 2.56 (1.12) 1.84 (1.07) 2.40 (1.36) 1–4 1–4 2.01 (1.03) 1.65 (1.00) (cf., Lutz & Ranzini, 2017). We combined these responses and then averaged the scores (Table 1). For network size, we collected participants’ responses to the question “How many friends do you have on Facebook?” (Lu et al., 2018). The response options for Facebook were 0 to 30, 31 to 100, 101 to 200, 201 to 400, and more than 400. For Instagram and Twitter, we asked about the number of followers they had on the platform. The response options were 0 to 15, 16 to 100, 101 to 200, 201 to 400, and more than 400 (see Table 1). The response categories for Facebook and Twitter were based on a pilot test conducted in 2017 in the US, UK, and France. During the pilot testing, we offered an open-ended response about network size. We devised categories based on natural breaks in the distribution but more importantly, the move to categories, instead of an open-ended response, reduced the burden of response and likely improved accuracy. To isolate the distinct roles of our key variables on political expression, we account for several other variables. Political efficacy was measured based on agreement with the following three statements: (a) People like me can influence government; (b) I Hoffmann and Boulianne 11 consider myself well-qualified to participate in politics; and (c) Working as a group, people can influence government (Response scale 1–4). These measures were chosen based on a systematic review of 193 studies on political efficacy (Boulianne et al., 2023). Political ideology was assessed using a self-placement question on a scale of 0 (left) to 10 (right; Jost et al., 2009). We coded responses 0 to 2 as “left-wing” and 8 to 10 as “right-wing.” Seeing political information on the platform was asked based on the question, “Please think about current events in the world, news about elections, information about political figures, information about government performance, debates about public policy, and other political issues. During the past 12 months, how often have you seen this type of content when you are using Facebook/Instagram/Twitter?” (never, rarely, time to time, often). Respondents reported more exposure to political information on Facebook and Twitter compared to Instagram (Table 1). We also asked if respondents identified as: females = 1, males = 0, non-binary = missing. Education was assessed based on a series of categories: high school or less, some college, bachelor’s degree, and more than a bachelor’s degree. Age was measured in years. To test our hypotheses and address our research questions, we first ran ordinary least squares regression models of political expression on each platform with pooled country data (Table 2). Second, we examine whether the relationships differ by country (Table 3), and finally, we examine interaction effects (Tables 4 and 5). Data and replication files are available here: https://osf.io/ksaud/. Results For all three platforms, we find that institutional privacy concerns are weakly but negatively related to political expression on social media (H1)—consistently across the three platforms. For Instagram, the higher the privacy concerns, the lower the frequency of posting political views on this platform (b = −0.066, p < .001). A similar pattern is observed for Facebook (b = −0.046, p < .001) and Twitter (b = −0.048, p = .014). Network size is positively related to political expression (H3). The coefficient for network size differs across platforms. Twitter is distinctive in that the size of one’s network is a larger factor (b = 0.239, p < .001) in predicting the likelihood of political expression than it is for Facebook (b = 0.095, p < .001) or Instagram (b = 0.077, p < .001). Figure 2 illustrates the distinctiveness of the Twitter network effect on posting on the platform. The confidence interval for Twitter network effects does not overlap with the confidence intervals for the effects of networks on posting for the other two platforms. As such, this figure illustrates the distinctiveness of the Twitter findings. Respondents based in France are more likely to engage in political expression on social media compared to those in the US. Canadian and UK users do not differ from US users in terms of their Facebook and Twitter posting habits. However, for Instagram, UK users are more likely to post than US users. Political interest positively correlates with posting on Facebook and Twitter but is not significant for Instagram. Political 12 Communication Research 00(0) Table 2. Ordinary Least Squares Regression Model of Political Expression on Each Platform. Variables Facebook b Twitter SE p-Value b 0.215 0.036 <.001 Francea United Kingdoma 0.059 0.036 .097 Canadaa 0.007 0.035 .835 Femalesa −0.152 0.026 <.001 Education −0.030 0.012 .016 Age −0.010 0.001 <.001 Left-winga −0.010 0.033 .760 Right-winga 0.288 0.030 <.001 Political efficacy 0.269 0.023 <.001 Exposure to political 0.270 0.013 <.001 information on the platform Political interest 0.130 0.017 <.001 Network size on the 0.095 0.011 <.001 platform Privacy concerns −0.046 0.013 <.001 R-squared = .347 n = 4,249 0.299 0.033 −0.029 −0.112 −0.018 −0.008 −0.081 0.347 0.323 0.216 Instagram SE p-Value b 0.053 0.049 0.049 0.037 0.017 0.001 0.047 0.044 0.033 0.018 <.001 .498 .553 .002 .298 <.001 .086 <.001 <.001 <.001 SE 0.306 0.046 0.164 0.046 0.025 0.044 −0.164 0.033 −0.021 0.015 −0.007 0.001 −0.166 0.042 0.289 0.039 0.235 0.029 0.440 0.018 p-Value <.001 <.001 .568 <.001 .162 <.001 <.001 <.001 <.001 <.001 0.101 0.025 <.001 0.239 0.015 <.001 0.040 0.022 .074 0.077 0.014 <.001 −0.048 0.020 .014 R-squared = .397 n = 2,229 −0.066 0.017 <.001 R-squared = .451 n = 2,256 a The reference groups for the analysis are US respondents for the series of country variables, males, and those in the center/moderate/neither left nor right in terms of political ideology. Table 3. Key Variables by Country.* Variables Facebook Twitter b SE p-Value b 0.172 0.124 −0.009 0.035 0.020 0.026 <.001 <.001 .742 0.104 0.076 −0.044 0.037 0.022 0.026 0.070 0.119 −0.073 0.150 0.059 −0.068 Instagram SE p-Value b SE p-Value 0.101 0.048 0.260 0.027 −0.001 0.038 .035 <.001 .982 0.082 0.124 −0.009 0.044 0.025 0.035 .064 <.001 .788 .004 .001 .099 0.131 0.050 0.197 0.029 −0.077 0.037 .010 <.001 .039 0.049 0.046 −0.135 0.044 0.028 0.034 .269 .102 <.001 0.033 0.023 0.025 .034 <.001 .003 0.030 0.057 0.239 0.040 −0.092 0.048 .595 <.001 .053 −0.017 0.071 −0.075 0.045 0.032 0.038 .714 .028 .051 0.034 0.021 0.024 <.001 .005 .005 0.127 0.049 0.230 0.031 −0.046 0.037 .009 <.001 .212 0.050 0.055 −0.055 0.044 0.028 0.033 .265 .049 .093 USA Political interest Network size on the platform Privacy concerns (α = .907) UK Political interest Network size on the platform Privacy concerns (α = .899) France Political interest Network size on the platform Privacy concerns (α = .892) Canada Political interest Network size on the platform Privacy concerns (α = .918) *Full models are available in Appendix A1 and A2. Hoffmann and Boulianne 13 Table 4. Moderated Effects.* Variables Facebook b SE Twitter p-Value b Instagram SE p-Value b SE p-Value Privacy concerns × Political interest 0.051 0.012 <.001 0.077 0.019 <.001 0.053 0.016 .001 R-squared = .350 R-squared = .402 R-squared = .453 b SE p-Value b SE p-Value b SE p-Value 0.089 0.010 <.001 0.098 0.015 <.001 0.077 0.013 <.001 Network size × Political interest R-squared = .360 R-squared = .409 R-squared = .460 *Full models are available in Appendix A3. Table 5. Key Variables by Level of Political Interest.* Variables Facebook b 1 = Not at all interested Network size on the 0.030 platform Privacy concerns −0.057 2 = Not very interested Network size on the 0.073 platform Privacy concerns −0.062 3 = Fairly interested Network size on the 0.070 platform Privacy concerns −0.069 4 = Very interested Network size on the 0.176 platform Privacy concerns 0.019 Twitter Instagram SE p-Value b SE p-Value 0.019 .113 0.085 0.040 .035 0.027 0.026 .316 0.019 .002 −0.082 0.043 .059 −0.067 0.027 .015 0.019 <.001 0.180 0.031 <.001 0.037 0.025 .140 0.022 .006 −0.135 0.038 <.001 −0.144 0.031 <.001 0.017 <.001 0.240 0.024 <.001 0.067 0.024 .005 0.022 .002 −0.075 0.032 .019 −0.084 0.031 .007 0.026 <.001 0.300 0.030 <.001 0.034 0.028 0.043 .512 .568 b SE p-Value 0.151 0.032 <.001 0.030 0.042 .478 *Full models are available in Appendix A4 and A5. efficacy and exposure to political information positively correlate with political expression on the different platforms. On all three platforms, male, younger, and rightwing users are more likely to engage in political expression. Education negatively relates to political expression on Facebook. Overall, the model fit is good but is best for Instagram (45.1% explained variance). We also examine whether the key variables have different roles in each country. The sample sizes for this platform-specific and country-specific analysis range from 427 (Twitter users in France) to 1,164 (Facebook users in the US). The full set of results is available in the Appendix, but Table 3 highlights the key variables. In each case, network size remains a statistically significant predictor of political expression on each platform (except for the UK and Instagram). In all countries, the role of 14 Communication Research 00(0) 0.50 0.40 0.30 0.20 0. 10 + 0.00 t + t + -0. IO + + Privacy TW Privacy JG -0.20 -0.30 -0.40 -0.50 Political Political Political interest FB interest TW interest JG Network Network Network size FB size TW size JG Pri vacy FB Figure 2. Marginal effects (unstandardized regression coefficients) with 95% confidence intervals. networks is strongest for Twitter-based political expression (see Figure 3). The role of privacy concerns, instead, differs by country and platform. Specifically, privacy concerns are significant and negative for UK-based Instagram users, but this pattern is not significant in other countries. Privacy concerns are also significantly and negatively related to the posting of political content on Twitter for UK-based respondents. Privacy concerns are negatively related to posting political content on Facebook in France and Canada. To assess the differential role of our three key variables, we create a series of interaction terms using the pooled sample, then add these one at a time to the model outlined in Table 2. First, we examine whether political interest and privacy concerns interact (H2) regarding the likelihood of posting political information on different platforms (Table 4). Across the three platforms, we find that this is indeed the case— the interaction term is statistically significant (H2). To clarify these findings, we split the sample into four groups based on political interest (Table 5). For each group, we examine the coefficients for privacy concerns related to posting on each of the three platforms. We find that for the first three groups (not at all interested, not very interested, and fairly interested), privacy concerns negatively relate to posting to the various platforms. Across all three platforms, for the “very interested” group, privacy concerns do not have a significant relationship to posting. Finally, we examine an interaction term for political interest and network size (H4). These results are presented in Table 4. The interaction terms are statistically significant across platforms. To illustrate the differential effect, we again split the sample into groups based on political interest (Table 5). The moderation effect is most apparent for Twitter, where the coefficient for network size becomes larger as political interest increases. In the case of Facebook, network size is not a significant predictor of Hoffmann and Boulianne 15 0.,0 0.40 0.30 0.20 + 0.10 0.00 -0.10 --0.20 -0.30 -0.40 -0.,0 Network Network size FB size FB Network size FB Network size FB Network size T\V Network size TW Network size T\V Network size T\V Network size JG Network size IG Network size JG Network size IG Figure 3. Marginal effects (unstandardized regression coefficients) of network effects by country with 95% confidence intervals. posting for individuals who are not at all politically interested. It is significant (at roughly the same effect size) for those who are “not very” and “fairly” interested, and has the strongest relationship for those who are “very interested.” In the case of Instagram, the relationship is not statistically significant for the “not at all” and “not very” interested, it is significant for the two more politically interested group, with the largest coefficient for the “very interested.” Thus, despite some variance across platforms, these findings align with H4. Discussion This study contributes to current research on the complex role of vertical (institutional) and horizontal (social) privacy considerations in political expression on social media platforms. It applies a comparative lens, examining the use of three major platforms across four Western countries, to highlight the consistent moderating role of political interest in the relationships between institutional privacy concerns, network size, and political expression. We demonstrate that taking political interest into account can contribute to resolving some apparently contradictory findings on the roles of privacy concerns and social networks in online political expression. Applying the “privacy calculus” (Dinev & Hart, 2006; Trepte et al., 2017), we argue that politically interested individuals obtain more gratification from political engagement on social media. Highly politically interested individuals, in particular, set aside their institutional privacy concerns and decide to post their views on different social media platforms. Political interest, thus, may partly explain why studies have come to inconsistent results when it comes to the role of privacy concerns and online political expression. We confirm findings that privacy concerns generally negatively relate to political expression (Ahmed & Lee, 2025; Ahmed & Masood, 2025). However, highly politically interested individuals, such as activists engaging in resource-intensive forms of 16 Communication Research 00(0) political expression, are likely to discard their privacy concerns (Hoffmann & Lutz, 2023). This pattern has also been observed in the context of political activism in authoritarian countries (Lokot, 2020; Mak et al., 2024). It speaks to a contingent rather than the expected contributory moderation (Holbert & Park, 2020), as the phenomenon seems to be limited to the most highly interested individuals. Concerning the size of one’s social network, our findings indicate that the highly politically interested may not be deterred by “context collapse,” as large social networks could afford greater benefits from political expression on social media. As networks expand, they tend to become more heterogeneous and composed of weak ties with strangers as well as known others (Eveland et al., 2013; Eveland & Hively, 2009). Large networks offer more opportunities to engage in lively political discussions (Zhang et al., 2024), which will be especially attractive to the highly politically interested. Research on the “spiral of silence” reveals that individuals tend to engage in more self-censorship when participating in political conversations with known others (Matthes et al., 2018). This research is enriched by our finding that, for the less politically interested, privacy concerns more strongly inhibit political expression on social media. Highlighting the moderating role of political interest in the relationships between institutional and social predictors and political expression on social media is important due to the nature of this personal characteristic. First, research has shown that political interest is unevenly distributed throughout the population. It is positively related to socioeconomic status, especially education, cognitive ability, or male gender (Denny & Doyle, 2008; Van Deth, 1990). Political interest, thus, contributes to digital inequalities in the context of political expression on social media. Second, political interest is a relatively stable individual characteristic, forming during adolescence, but changing little after (Prior, 2010; Russo & Stattin, 2017). Political interest, accordingly, is difficult to influence. Third, political interest correlates with characteristics that tend to shape the quality of online political discourse, such as extraversion, assertiveness, or status-seeking (Bor & Petersen, 2022; Denny & Doyle, 2008). Regarding our second research question on platform differences, we find that of the platforms examined here, Twitter appears to offer the greatest possibility of fostering a large network, that is, a network composed of weak ties (Valenzuela et al., 2018). We observe distinct patterns around the size of one’s network on Twitter and subsequent expression. The positive relationship between network size and political expression is much larger than for other platforms. This finding is replicated across all four countries. Research on online political expression needs to take specific affordances and architectures of social media platforms into account (Bossetta, 2018). We add contextual nuance by demonstrating that the relationship between privacy concerns and political expression on social media only holds for some platforms when looking at country-specific samples. In the UK, privacy concerns are negatively related to political expression on Instagram and Twitter. In France and Canada, we find this negative relationship for Facebook. In the US, we find no significant relationship between privacy concerns and political posting. Hoffmann and Boulianne 17 Beyond our core variables, we find consistency across the three platforms: political efficacy, exposure to political information, being male, young, and right-wing positively relate to political expression. Education is negatively related to expression on Facebook (cf., Hoffmann & Lutz, 2017; Kelm et al., 2023; Mitchelstein et al., 2021). This study is subject to some limitations. We rely on cross-sectional data, which allows for correlational analyses rather than causal interpretations. We account for two variables shown in previous research to relate to political expression on social media: political efficacy and exposure to online political information. Still, other variables, such as personality or status orientation, may help further untangle the relationships explored here. That said, the models have good R-squared statistics, meaning the set of variables that we include explains a lot of the variation in posting (45% for Instagram, 40% for Twitter, and 35% for Facebook). We present data collected in 2019. In 2021, Meta announced a downrating of political content and has recently announced a reversal of these changes for the US market, Twitter/X has undergone significant changes in content moderation policies (Kaplan, 2025). Some of these changes may asymmetrically affect left- or right-wing users’ willingness to engage in political expression. We account for political orientation in our models, but future studies should examine potential changes in user behavior. We base our argument on the privacy calculus theory but do not test specific instances of privacy calculations. Future studies may both qualitatively explore the distinct benefits and risks associated with political expression, depending on political interest, and may test the role of political interest in the context of specific political online behaviors. While we have a single measure of political expression, our research offers an important platform focus, which is missing from existing scholarship (Lane et al., 2022). Had we combined posting activities across platforms (Facebook, Twitter, and Instagram), we would have replicated our key findings using this multi-item measure: privacy concerns and networks significantly relate to posting of political content. Political expression on social media is contingent upon both institutional and social privacy concerns. In this study, we apply the privacy calculus theory to focus on two predictors that previous studies have highlighted as related to online political expression in complex, even unexpected ways: institutional privacy concerns and social network size. Based on the privacy calculus theory, we show how political interest plays a moderating role here across distinct social media platforms. We argue that politically interested users disregard privacy concerns to derive enjoyment from politically engaging large online audiences. Our cross-national study examines these relationships comparatively for three social media platforms: Facebook, Twitter, and Instagram. Political expression on social media is rare, but it is important to understand the factors influencing people’s decisions to post, particularly when these posts are used as measures of public opinion and platforms serve as virtual public spheres. ORCID iDs 0 0 Christian Pieter Hoffmann https://orcid.org/0000-0002-5282-6950 Shelley Boulianne https://orcid.org/0000-0002-8951-1098 18 Communication Research 00(0) Ethical Considerations The survey received human-subject ethics approval prior to data collection (File No. 101662), according to Canada’s Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (TCPS) Funding The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is funded by the Social Sciences and Humanities Research Council of Canada (Grant No. 435-2019-04-94). Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Data Availability Statement DOI 10.17605/OSF.IO/KSAUD Anonymous link: https://osf.io/ksaud/?view_only=27b69cf9a 12e4044a440228a99a1b7a1 Open Practices DOI 10.17605/OSF.IO/KSAUD References Ahmed, S., & Lee, S. (2025). The inhibition effect: Privacy concerns disrupt the positive effects of social media use on online political participation. New Media & Society, 27, 203–224. Ahmed, S., & Masood, M. (2025). Assessing the effects of privacy concerns on social mediadriven political discussions and participation: A comparative study of Asian and Western contexts. Social Science Computer Review, 43, 67–88. Barnidge, M., Huber, B., de Zúñiga, H. G., & Liu, J. H. (2018). Social media as a sphere for “risky” political expression: A twenty-country multilevel comparative analysis. The International Journal of Press/Politics, 23(2), 161–182. Best, S. J., & Krueger, B. S. (2011). Government monitoring and political participation in the United States: The distinct roles of anger and anxiety. American Politics Research, 39(1), 85–117. Bimber, B., Cunill, M. C., Copeland, L., & Gibson, R. (2015). Digital media and political participation: The moderating role of political interest across acts and over time. Social Science Computer Review, 33(1), 21–42. Blank, G. (2013). Who creates content? Stratification and content creation on the Internet. Information Communication & Society, 16(4), 590–612. Bor, A., & Petersen, M. B. (2022). The psychology of online political hostility: A comprehensive, cross-national test of the mismatch hypothesis. American Political Science Review, 116(1), 1–18. Hoffmann and Boulianne 19 Bossetta, M. (2018). The digital architectures of social media: Comparing political campaigning on Facebook, Twitter, Instagram, and Snapchat in the 2016 U.S. election. Journalism & Mass Communication Quarterly, 95(2), 471–496. Boulianne, S., & Hoffmann, C. P. (2022). The social, civic, and political uses of Instagram in four countries. Journal of Quantitative Description: Digital Media, 2, 30. https://doi. org/10.51685/jqd.2022.001 Boulianne, S., Hoffmann, C. P., & Bossetta, M. (2024). Social media platforms for politics: A comparison of Facebook, Instagram, Twitter, YouTube, reddit, snapchat, and WhatsApp. New Media & Society. Advance online publication. https://doi.org/10.1177/14614448241262415 Boulianne, S., & Koc-Michalska, K. (2022). The role of personality in political talk and likeminded discussion. The International Journal of Press/Politics, 27(1), 285–310. Boulianne, S., Oser, J., & Hoffmann, C. P. (2023). Powerless in the digital age? A systematic review and meta-analysis of political efficacy and digital media use. New Media & Society, 25(9), 2512–2536. Boyd, D. (2006). Friends, friendsters, and top 8: Writing community into being on social network sites. First Monday, 11(12). Boyd, D. M., & Ellison, N. B. (2007). Social network sites: Definition, history, and scholarship. Journal of Computer-Mediated Communication, 13(1), 210–230. Brady, W. J., McLoughlin, K., Doan, T. N., & Crockett, M. J. (2021). How social learning amplifies moral outrage expression in online social networks. Science Advances, 7(33), eabe5641. Business Insider. (2022). User trust in social platforms is falling, according to our new study. https://www.insiderintelligence.com/content/user-trust-social-platforms-falling-accordingour-new-study Dal, A. (2024). Firewalls have ears: How horizontal privacy regulation influences online political expression in Russia. Journal of International Communication, 18, 2731–2752. Davis, N. T., & Dunaway, J. L. (2016). Party polarization, media choice, and mass partisanideological sorting. Public Opinion Quarterly, 80(S1), 272–297. Dencik, L., & Cable, J. (2017). The advent of surveillance realism: Public opinion and activist responses to the Snowden leaks. Journal of International Communication, 11, 763–781. Denny, K., & Doyle, O. (2008). Political interest, cognitive ability and personality: Determinants of voter turnout in britain. British Journal of Political Science, 38(2), 291–310. Dienlin, T., Masur, P. K., & Trepte, S. (2023). A longitudinal analysis of the privacy paradox. New Media & Society, 25(5), 1043–1064. Dienlin, T., & Metzger, M. J. (2016). An extended privacy calculus model for SNSs: Analyzing self-disclosure and self-withdrawal in a representative US sample. Journal of ComputerMediated Communication, 21(5), 368–383. Dienlin, T., & Trepte, S. (2015). Is the privacy paradox a relic of the past? An in-depth analysis of privacy attitudes and privacy behaviors. European Journal of Social Psychology, 45(3), 285–297. Dinev, T., & Hart, P. (2004). Internet privacy concerns and their antecedents - measurement validity and a regression model. Behaviour & Information Technology, 23(6), 413–422. Dinev, T., & Hart, P. (2006). An extended privacy calculus model for e-commerce transactions. Information Systems Research, 17(1), 61–80. European Data Protection Board. (2022, September 15). Record fine for Instagram following EDPB intervention. https://www.edpb.europa.eu/news/news/2022/record-fine-instagramfollowing-edpb-intervention_en 20 Communication Research 00(0) Evans, S. K., Pearce, K. E., Vitak, J., & Treem, J. W. (2017). Explicating affordances: A conceptual framework for understanding affordances in communication research. Journal of Computer-Mediated Communication, 22(1), 35–52. Eveland, W. P., & Hively, M. H. (2009). Political discussion frequency, network size, and “heterogeneity” of discussion as predictors of political knowledge and participation. Journal of Communication, 59(2), 205–224. Eveland, W. P., Hutchens, M. J., & Morey, A. C. (2013). Political network size and its antecedents and consequences. Political Communication, 30(3), 371–394. Gil de Zúñiga, H., & Valenzuela, S. (2011). The mediating path to a stronger citizenship: Online and offline networks, weak ties, and civic engagement. Communication Research, 38(3), 397–421. Grubbs, J. B., Warmke, B., Tosi, J., James, A. S., & Campbell, W. K. (2019). Moral grandstanding in public discourse: Status-seeking motives as a potential explanatory mechanism in predicting conflict. PLoS One, 14(10), e0223749. Guess, A., Munger, K., Nagler, J., & Tucker, J. (2019). How accurate are survey responses on social media and politics? Political Communication, 36(2), 241–258. Hampton, K. N., Shin, I., & Lu, W. (2017). Social media and political discussion: when online presence silences offline conversation. Information, Communication & Society, 20(7), 1090–1107. Hoffman, D. L., Novak, T. P., & Peralta, M. A. (1999). Information privacy in the marketspace: Implications for the commercial uses of anonymity on the Web. Information Society, 15(2), 129–139. Hoffmann, C. P., & Lutz, C. (2017, July). Spiral of silence 2.0: Political self-censorship among young Facebook users [Conference session]. Proceedings of the 8th International Conference on Social Media & Society. Hoffmann, C. P., & Lutz, C. (2023). The contextual role of privacy concerns in online political participation. European Journal of Communication, 38(4), 363–379. Hoffmann, C. P., Lutz, C., & Meckel, M. (2015). Content creation on the Internet: A social cognitive perspective on the participation divide. Information Communication & Society, 18(6), 696–716. Holbert, R. L., & Park, E. (2020). Conceptualizing, organizing, and positing moderation in communication research. Communication Theory, 30(3), 227–246. John, N. A., & Dvir-Gvirsman, S. (2015). “I don’t like you any more”: Facebook unfriending by Israelis during the Israel-Gaza Conflict of 2014. Journal of Communication, 65(6), 953–974. Jost, J. T., Federico, C. M., & Napier, J. L. (2009). Political ideology: Its structure, functions, and elective affinities. Annual Review of Psychology, 60, 307–337. Kakavand, A. E. (2024). Far-right social media communication in the light of technology affordances: A systematic literature review. Annals of the International Communication Association, 48(1), 37–56. Kaplan, J. (2025). More speech and fewer mistakes. https://about.fb.com/news/2025/01/metamore-speech-fewer-mistakes/ Kelm, O., Phillips, G., Frieß, D., & Ziegele, M. (2023). Politische Informationen und Diskussionen in Sozialen Medien: Neun Plattformen im Vergleich [Political information and discussion on social media: A comparison of nine platforms]. Media Perspektiven, 10, 1–16. Hoffmann and Boulianne 21 Kezer, M., Dienlin, T., & Baruh, L. (2022). Getting the privacy calculus right: Analyzing the relations between privacy concerns, expected benefits, and self-disclosure using response surface analysis. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 16(4). Kim, J. W., Guess, A., Nyhan, B., & Reifler, J. (2021). The distorting prism of social media: How self-selection and exposure to incivility fuel online comment toxicity. Journal of Communication, 71(6), 922–946. Krasnova, H., Veltri, N. F., & Günther, O. (2012). Self-disclosure and privacy calculus on social networking sites: The role of culture. Business & Information Systems Engineering, 4, 127–135. Lane, D. S., Do, K., & Molina-Rogers, N. (2022). What is political expression on social media anyway? A systematic review. Journal of Information Technology & Politics, 19(3), 331– 345. Lee, Y. H., & Yuan, C. W. (2020). The privacy calculus of “friending” across multiple social media platforms. Social Media + Society, 6(2). Lokot, T. (2020). Articulating networked citizenship on the Russian internet: A case for competing affordances. Social Media + Society, 6(4), 1–12. Lutz, C., & Ranzini, G. (2017). Where dating meets data: Investigating social and institutional privacy concerns on Tinder. Social Media + Society, 3(1), 2056305117697735. Lu, Y., Lee, J. K., & Kim, E. (2018). Network characteristics matter in politics on Facebook: Evidence from a US national survey. Online Information Review, 42(3), 372–386. Mak, M. K., Koo, A. Z. X., & Rojas, H. (2024). Social media engagement against fear of restrictions and surveillance: The mediating role of privacy management. New Media & Society, 26(4), 1984–2005. Malhotra, N. K., Kim, S. S., & Agarwal, J. (2004). Internet users’ information privacy concerns (IUIPC): The construct, the scale, and a causal model. Information Systems Research, 15(4), 336–355. Marwick, A. E., & Boyd, D. (2011). I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience. New Media & Society, 13(1), 114–133. Masur, P. K., Epstein, D., Quinn, K., Wilhelm, C., Baruh, L., & Lutz, C. (2025). Comparative privacy research: Literature review, framework, and research agenda. Information Society, 41(2), 69–90. https://doi.org/10.1080/01972243.2025.2451863 Matthes, J., Knoll, J., & von Sikorski, C. (2018). The “spiral of silence” revisited: A metaanalysis on the relationship between perceptions of opinion support and political opinion expression. Communication Research, 45(1), 3–33. McClain, C., Anderson, M., & Gelles-Watnick, R. (2024). How Americans Navigate Politics on TikTok, X, Facebook and Instagram. https://www.pewresearch.org/internet/2024/06/12/ how-americans-navigate-politics-on-tiktok-x-facebook-and-instagram/ McKnight, D. H., Choudhury, V., & Kacmar, C. (2002). Developing and validating trust measures for e-commerce: An integrative typology. Information Systems Research, 13(3), 334–359. Miltgen, C. L., & Peyrat-Guillard, D. (2014). Cultural and generational influences on privacy concerns: A qualitative study in seven European countries. European Journal of Information Systems, 23(2), 103–125. Mitchelstein, E., Boczkowski, P., & Giuliano, C. (2021). Platform matters: Political opinion expression on social media. Weizenbaum Journal of the Digital Society, 1(1), 1–26. 22 Communication Research 00(0) Morrow, E. (2022). Priming privacy: The effect of privacy news consumption on privacy attitudes, beliefs, and Knowledge. Journal of Broadcasting & Electronic Media, 66(5), 772–793. Newman, N., Fletcher, R., Robertson, C. T., Arguedas, A. R., & Nielsen, R. K. (2024). Digital news report 2024. https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2024 Pearce, K. E., Vitak, J., & Barta, K. (2018). Socially mediated visibility: Friendship and dissent in authoritarian Azerbaijan. Journal of International Communication, 12, 1310–1331. Prior, M. (2010). You’ve either got it or you don’t? The stability of political interest over the life cycle. Journal of Politics, 72(3), 747–766. Quinn, K., Epstein, D., & Moon, B. (2019). We care about different things: Non-elite conceptualizations of social media privacy. Social Media + Society, 5(3), 2056305119866008. Rogowski, J. C., & Sutherland, J. L. (2016). How ideology fuels affective polarization. Political Behavior, 38(2), 485–508. Ruess, C., Hoffmann, C. P., Boulianne, S., & Heger, K. (2023). Online political participation: The evolution of a concept. Information Communication & Society, 26(8), 1495–1512. Russo, S., & Stattin, H. (2017). Stability and change in youths’ political interest. Social Indicators Research, 132(2), 643–658. Schulz, W. S. (2025). What would You Say? Estimating causal effects of social context on political expression. Political Analysis, 33(2), 122–139. Settle, J. E., & Carlson, T. N. (2019). Opting out of political discussions. Political Communication, 36(3), 476–496. Sheehan, K. (2015). A change in the climate: Online social capital and the spiral of silence. First Monday, 20(5). Skoric, M. M., Zhu, Q., & Lin, J. H. T. (2018). What predicts selective avoidance on social media? A study of political unfriending in Hong Kong and Taiwan. American Behavioral Scientist, 62(8), 1097–1115. Stoycheff, E. (2016). Under surveillance: Examining Facebook’s spiral of silence effects in the wake of NSA Internet monitoring. Journalism & Mass Communication Quarterly, 93(2), 296–311. Trepte, S., Reinecke, L., Ellison, N. B., Quiring, O., Yao, M. Z., & Ziegele, M. (2017). A crosscultural perspective on the privacy calculus. Social Media + Society, 3(1). Turow, J., Hennessy, M., & Draper, N. (2018). Persistent misperceptions: Americans’ misplaced confidence in privacy policies, 2003–2015. Journal of Broadcasting & Electronic Media, 62(3), 461–478. Valenzuela, S., Correa, T., & Gil de Zúñiga, H. (2018). Ties, likes, and tweets: Using strong and weak ties to explain differences in protest participation across Facebook and Twitter use. Political Communication, 35(1), 117–134. Van Deth, J. W. (1990). Interest in politics. In M. K. Jennings, J. W. van Deth, S. H. Barnes, D. Fuchs, F. J. Heunks, R. Inglehart, M. Kaase, H. D. Klingemann & J. J. A. Thomassen (Eds.), Continuities in political action: A longitudinal study of political orientations in three Western democracies (pp. 275–312). De Gruyter. Vitak, J. (2012). The impact of context collapse and privacy on Social Network Site Disclosures. Journal of Broadcasting & Electronic Media, 56(4), 451–470. Vogels, E. A., Auxier, B., & Anderson, M. (2021). Partisan differences in social media use show up for some platforms, but not Facebook. Pew Research Center. https://www.pewresearch.org/short-reads/2021/04/07/partisan-differences-in-social-media-use-show-upforsome-platforms-but-not-facebook/ Hoffmann and Boulianne 23 Weeks, B. E., Halversen, A., & Neubaum, G. (2024). Too scared to share? Fear of social sanctions for political expression on social media. Journal of Computer-Mediated Communication, 29(1), zmad041. Wojcieszak, M. E., & Mutz, D. C. (2009). Online groups and political discourse: Do online discussion spaces facilitate exposure to political disagreement? Journal of Communication, 59(1), 40–56. Yang, J., Barnidge, M., & Rojas, H. (2017). The politics of “unfriending”: User filtration in response to political disagreement on social media. Computers in Human Behavior, 70, 22–29. Young, A. L., & Quan-Haase, A. (2013). Privacy protection strategies on Facebook: The Internet privacy paradox revisited. Information Communication & Society, 16(4), 479–500. Zerback, T., & Fawzi, N. (2017). Can online exemplars trigger a spiral of silence? Examining the effects of exemplar opinions on perceptions of public opinion and speaking out. New Media & Society, 19(7), 1034–1051. Zhang, J., Cooks, E., Kim, B., Viehouser, M., & Barnidge, M. (2024). Who says what? Social networks and digital inequalities in online political expression and content creation. Information Society, 40(5), 361–375. Author Biographies Christian Pieter Hoffmann is professor of communication management and political communication at the Institute of Communication and Media Studies and the Institute of Political Science, University of Leipzig. His research is focused on online participation, trust, self-disclosure, and privacy protection in social media. Shelley Boulianne is the R. Klein Chair in Communication Studies at Mount Royal University (Canada). She has held professor positions at the MacEwan University (Canada), Université Catholique de Lille (France) and the University of Southampton (UK). Her research examines the global dynamics of digital media use for citizens’ engagement in civic and political life. 24 −0.104 0.056 .062 −0.011 0.025 .660 −0.010 0.002 <.001 0.027 0.070 .702 0.329 0.062 <.001 0.282 0.049 <.001 0.309 0.026 <.001 0.104 0.037 .004 0.076 0.022 .001 −0.044 0.026 .099 R-squared = .349, n = 988 −0.274 0.051 <.001 −0.064 0.024 .007 −0.010 0.002 <.001 0.081 0.069 .242 0.324 0.059 <.001 0.275 0.046 <.001 0.232 0.027 <.001 0.172 0.035 <.001 0.124 0.020 <.001 −0.009 0.026 .742 R-squared = .378, n = 1164 p-Value −0.070 0.075 .351 −0.043 0.033 .191 −0.013 0.003 <.001 −0.126 0.092 .169 0.378 0.085 <.001 0.325 0.064 <.001 0.191 0.036 <.001 0.131 0.050 .010 0.197 0.029 <.001 −0.077 0.037 .039 R-squared = .380, n = 590 −0.187 0.069 .007 0.027 0.032 .405 −0.006 0.002 .007 −0.004 0.091 .968 0.418 0.082 <.001 0.362 0.061 <.001 0.191 0.034 <.001 0.101 0.048 .035 0.260 0.027 <.001 −0.001 0.038 .982 R-squared = .429, n = 675 SE *The reference groups for the analysis are males and those in the center/moderate/neither left nor right in terms of political ideology. Females Education Age Left-wing Right-wing Political efficacy Exposure to political information on the platform Political interest Network size on the platform Privacy concerns Model information UK Females* Education Age Left-wing* Right-wing* Political efficacy Exposure to political information on the platform Political interest Network size on the platform Privacy concerns Model information USA b p-Value b SE Twitter Facebook SE p-Value −0.157 0.070 .026 −0.018 0.030 .544 −0.008 0.003 .003 −0.189 0.085 .026 0.374 0.078 <.001 0.188 0.061 .002 0.494 0.038 <.001 0.049 0.044 .269 0.046 0.028 .102 −0.135 0.034 <.001 R-squared = .447, n = 519 −0.230 0.064 <.001 −0.008 0.029 .772 −0.007 0.002 .004 −0.256 0.084 .003 0.283 0.077 <.001 0.275 0.055 <.001 0.382 0.034 <.001 0.082 0.044 .064 0.124 0.025 <.001 −0.009 0.035 .788 R-squared = .473, n = 665 b Instagram Appendix Table A1. Ordinary Least Squares Regression Model of Political Expression on Each Platform (USA and UK). Appendix 25 −0.091 0.050 .069 0.018 0.027 .500 −0.009 0.002 <.001 −0.119 0.065 .067 0.316 0.064 <.001 0.211 0.048 <.001 0.249 0.025 <.001 0.150 0.034 <.001 0.059 0.021 .005 −0.068 0.024 .005 R-squared = .293, n = 1075 −0.120 0.051 .019 −0.044 0.024 .069 −0.009 0.002 <.001 0.007 0.063 .913 0.210 0.061 .001 0.285 0.043 <.001 0.304 0.027 <.001 0.070 0.033 .034 0.119 0.023 <.001 −0.073 0.025 .003 R-squared = .373, n = 1019 p-Value −0.151 0.071 .033 0.035 0.036 .326 −0.007 0.002 .002 −0.091 0.089 .310 0.346 0.087 <.001 0.205 0.065 .002 0.217 0.034 <.001 0.127 0.049 .009 0.230 0.031 <.001 −0.046 0.037 .212 R-squared = .385, n = 534 −0.065 0.088 .458 −0.125 0.039 .002 −0.007 0.003 .020 −0.044 0.110 .691 0.223 0.102 .030 0.353 0.072 <.001 0.315 0.045 <.001 0.030 0.057 .595 0.239 0.040 <.001 −0.092 0.048 .053 R-squared = .391, n = 427 SE *The reference groups for the analysis are males. and those in the center/moderate/neither left nor right in terms of political ideology. Females Education Age Left-wing Right-wing Political efficacy Exposure to political information on the platform Political interest Network size on the platform Privacy concerns Model information Canada Females* Education Age Left-wing* Right-wing* Political efficacy Exposure to political information on the platform Political interest Network size on the platform Privacy concerns Model information France b p-Value b SE Twitter Facebook SE p-Value −0.169 0.064 .008 −0.008 0.032 .808 −0.007 0.002 .003 −0.152 0.078 .053 0.281 0.076 <.001 0.171 0.058 .003 0.380 0.036 <.001 0.050 0.044 .265 0.055 0.028 .049 −0.055 0.033 .093 R-squared = .397, n = 567 −0.114 0.073 .118 −0.031 0.033 .345 −0.006 0.003 .015 −0.012 0.092 .896 0.208 0.083 .013 0.251 0.062 <.001 0.509 0.040 <.001 −0.017 0.045 .714 0.071 0.032 .028 −0.075 0.038 .051 R-squared = .454, n = 502 b Instagram Appendix Table A2. Ordinary Least Squares Regression Model of Political Expression on Each Platform (France and Canada). 26 France* United Kingdom* Canada* Females* Education Age Left-wing* Right-wing* France* United Kingdom* Canada* Females* Education Age Left-wing* Right-wing* Political efficacy Exposure to political information on the platform Political interest Network size on the platform Privacy concerns Privacy concerns × Political interest 0.036 0.036 0.035 0.026 0.012 0.001 0.033 0.030 0.023 0.013 SE 0.308 0.045 −0.017 −0.106 −0.018 −0.009 −0.086 0.344 0.320 0.216 <.001 .047 .648 <.001 .016 <.001 .775 <.001 <.001 <.001 0.053 0.049 0.049 0.037 0.017 0.001 0.047 0.044 0.032 0.018 SE <.001 .349 .724 .004 .291 <.001 .067 <.001 <.001 <.001 p-Value −0.166 0.069 .017 0.234 0.015 <.001 −0.259 0.055 <.001 0.077 0.019 <.001 R-squared = .402, n = 2.229 0.296 0.053 <.001 0.035 0.048 .471 −0.023 0.049 .634 −0.108 0.036 .003 −0.022 0.017 .193 −0.008 0.001 <.001 −0.087 0.047 .062 0.334 0.044 <.001 b Twitter p-Value −0.047 0.045 .298 0.095 0.011 <.001 −0.175 0.033 <.001 0.051 0.012 <.001 R-squared = .347, n = 4.250 0.212 0.036 <.001 0.083 0.035 .019 0.021 0.034 .549 −0.142 0.025 <.001 −0.032 0.012 .009 −0.010 0.001 <.001 −0.011 0.033 .739 0.286 0.030 <.001 0.221 0.071 0.016 −0.152 −0.030 −0.010 −0.010 0.289 0.266 0.272 b Facebook Appendix Table A3. Full Models of Interaction Effects. 0.046 0.046 0.044 0.033 0.015 0.001 0.042 0.039 0.029 0.018 SE <.001 <.001 .474 <.001 .144 <.001 <.001 <.001 <.001 <.001 p-Value (continued) −0.145 0.061 .018 0.075 0.014 <.001 −0.204 0.046 <.001 0.053 0.016 .001 R-squared = .453, n = 2.256 0.297 0.046 <.001 0.166 0.045 <.001 0.021 0.044 .629 −0.159 0.033 <.001 −0.025 0.015 .101 −0.007 0.001 <.001 −0.164 0.042 <.001 0.280 0.039 <.001 0.312 0.175 0.032 −0.163 −0.022 −0.007 −0.169 0.289 0.231 0.440 b Instagram 27 0.259 0.270 0.023 0.013 SE <.001 <.001 p-Value −0.102 0.031 .001 −0.146 0.028 <.001 −0.046 0.013 <.001 0.089 0.010 <.001 R-squared = .360, n = 4249 b Facebook 0.317 0.221 0.032 0.018 SE <.001 <.001 p-Value −0.101 0.040 .011 −0.047 0.046 .310 −0.058 0.019 .003 0.098 0.015 <.001 R-squared = .409, n = 2229 b Twitter 0.230 0.430 0.029 0.018 SE <.001 <.001 p-Value −0.132 0.036 <.001 −0.135 0.038 <.001 −0.070 0.017 <.001 0.077 0.013 <.001 R-squared = .460, n = 2257 b Instagram *The reference groups for the analysis are US respondents for the series of country variables. males. and those in the center/moderate/neither left nor right in terms of political ideology. Political efficacy Exposure to political information on the platform Political interest Network size on the platform Privacy concerns Network size × Political interest Appendix Table A3. (continued) 28 France* United Kingdom* Canada* Females* Education Age Left-wing* Right-wing* Political efficacy Exposure to political information on the platform Network size on the platform Privacy concerns 1 = Not at all interested 0.064 0.067 0.067 0.050 0.025 0.002 0.074 0.072 0.039 0.024 .151 .841 .868 .079 .634 .787 .218 .027 .718 <.001 0.030 0.019 .113 −0.057 0.019 .002 R-squared = .119, n = 545 0.092 0.013 0.011 −0.087 0.012 <0.001 −0.091 0.159 0.014 0.157 0.145 0.133 0.142 0.109 0.052 0.004 0.140 0.158 0.078 0.048 SE .056 .483 .612 .286 .831 .245 .551 .002 .056 .001 p-Value 0.085 0.040 .035 −0.082 0.043 .059 R-squared = .220, n = 199 0.279 0.094 −0.072 −0.116 0.011 −0.004 −0.084 0.495 0.150 0.156 b p-Value b SE Twitter Facebook Appendix Table A4. Key Variables by Political Interest (Not At All and Not Very Interested). 0.094 0.092 0.095 0.074 0.033 0.003 0.093 0.105 0.051 0.040 SE .133 .340 .802 .019 .668 .665 .833 .033 .197 <.001 p-Value (continued) 0.027 0.026 .316 −0.067 0.027 .015 R-squared = .235, n = 258 0.141 0.088 −0.024 −0.173 −0.014 −0.001 −0.020 0.225 0.066 0.248 b Instagram 29 0.063 0.069 0.066 0.044 0.023 0.001 0.060 0.054 0.040 0.023 <.001 .036 .250 <.001 .427 <.001 .517 .002 <.001 <.001 0.073 0.019 <.001 −0.062 0.022 .006 R-squared = .184, n = 1167 0.287 0.144 0.076 −0.219 −0.018 −0.007 −0.039 0.165 0.178 0.213 0.099 0.102 0.102 0.068 0.034 0.002 0.092 0.086 0.062 0.036 SE <.001 .266 .527 <.001 .393 .027 .011 .004 <.001 <.001 p-Value 0.180 0.031 <.001 −0.135 0.038 <.001 R-squared = .284, n = 513 0.405 0.114 0.064 −0.269 0.029 −0.005 −0.234 0.247 0.236 0.203 b p-Value b SE Twitter Facebook 0.080 0.086 0.084 0.059 0.028 0.002 0.073 0.067 0.049 0.033 SE <.001 .086 .671 <.001 .329 .045 .005 .023 .041 <.001 p-Value 0.037 0.025 .140 −0.144 0.031 <.001 R-squared = .394, n = 579 0.345 0.148 0.036 −0.247 −0.027 −0.004 −0.206 0.153 0.101 0.444 b Instagram *The reference groups for the analysis are US respondents for the series of country variables. males. and those in the center/moderate/neither left nor right in terms of political ideology. France* United Kingdom* Canada* Females* Education Age Left-wing* Right-wing* Political efficacy Exposure to political information on the platform Network size on the platform Privacy concerns 2 = Not very interested Appendix Table A4.  (continued) 30 France* United Kingdom* Canada* Females* Education Age Left-wing* Right-wing* Political efficacy Exposure to political information on the platform Network size on the platform Privacy concerns 3 = Fairly interested 0.063 0.055 0.054 0.041 0.020 0.001 0.052 0.047 0.039 0.021 .001 .385 .410 .017 .025 <.001 .848 <.001 <.001 <.001 0.070 0.017 <.001 −0.069 0.022 .002 R-squared = .309, n = 1644 0.203 0.048 0.044 −0.097 −0.044 −0.012 −0.010 0.285 0.348 0.281 0.086 0.072 0.075 0.056 0.026 0.002 0.072 0.065 0.051 0.027 SE <.001 .974 .373 .567 .045 <.001 .421 <.001 <.001 <.001 p-Value 0.240 0.024 <.001 −0.075 0.032 .019 R-squared = .356, n = 905 0.357 0.002 0.067 −0.032 −0.052 −0.010 −0.058 0.310 0.329 0.205 b p-Value b SE Twitter Facebook Appendix Table A5. Key Variables by Political Interest (Fairly and Very Interested). 0.083 0.073 0.072 0.055 0.025 0.002 0.071 0.063 0.053 0.030 SE <.001 .002 .178 .008 .657 <.001 .030 <.001 <.001 <.001 p-Value (continued) 0.067 0.024 .005 −0.084 0.031 .007 R-squared = .412, n = 891 0.343 0.221 0.098 −0.145 −0.011 −0.008 −0.154 0.286 0.315 0.417 b Instagram 31 0.091 0.091 0.083 0.066 0.029 0.081 0.076 0.059 0.032 S.E. .025 .020 .788 .060 .295 .116 <.001 <.001 <.001 p-value 0.176 0.026 <.001 0.019 0.034 .568 R-squared = .417, n = 890 0.204 0.212 0.022 −0.123 −0.030 0.127 0.464 0.376 0.401 b Facebook 0.114 0.104 0.101 0.080 0.035 0.100 0.094 0.070 0.040 S.E. .111 .237 .396 .513 .782 .871 <.001 <.001 <.001 p-value 0.300 0.030 <.001 0.028 0.043 .512 R-squared = .423, n = 609 0.182 0.123 −0.086 −0.052 −0.010 −0.016 0.453 0.443 0.297 b Twitter 0.106 0.110 0.096 0.075 0.033 0.097 0.092 0.069 0.041 S.E. .046 .187 .560 .352 .178 .124 <.001 <.001 <.001 p-value 0.151 0.032 <.001 0.030 0.042 .478 R-squared = .531, n = 525 0.211 0.146 −0.056 −0.070 −0.045 −0.149 0.373 0.330 0.465 b Instagram *The reference groups for the analysis are US respondents for the series of country variables. males. and those in the center/moderate/neither left nor right in terms of political ideology. France* United Kingdom* Canada* Females* Education Left-wing* Right-wing* Political efficacy Exposure to political information on the platform Network size on the platform Privacy concerns 4 = Very interested Appendix Table A5. (continued) 32 1. Facebook posting of political r 1.000 content p 2. Twitter posting of political content r 0.751 p 0.000 3. Instagram posting of political r 0.735 content p 0.000 4. Political interest r 0.333 p 0.000 5. Privacy concerns r 0.110 p 0.000 6. Facebook network size r 0.249 p 0.000 7. Twitter network size r 0.319 p 0.000 8. Instagram network size r 0.282 p 0.000 9. Facebook exposure to political r 0.387 information p 0.000 10. Twitter exposure to political r 0.272 information p 0.000 11. Instagram exposure to political r 0.493 information p 0.000 12. Political Efficacy r 0.404 p 0.000 13. Females r −0.138 p 0.000 1 0.746 0.000 0.340 0.000 0.144 0.000 0.202 0.000 0.418 0.000 0.286 0.000 0.281 0.000 0.389 0.000 0.503 0.000 0.416 0.000 −0.180 0.000 0.712 0.000 1.000 2 0.223 0.000 0.059 0.020 0.216 0.000 0.382 0.000 0.355 0.000 0.188 0.000 0.191 0.000 0.605 0.000 0.364 0.000 −0.112 0.000 0.700 0.000 0.706 0.000 1.000 3 0.317 0.000 0.089 0.000 0.176 0.000 0.063 0.013 0.299 0.000 0.338 0.000 0.262 0.000 0.567 0.000 −0.190 0.000 0.284 0.000 0.295 0.000 0.201 0.000 1.000 4 0.076 0.003 0.119 0.000 0.083 0.001 0.276 0.000 0.301 0.000 0.162 0.000 0.289 0.000 −0.033 0.192 0.087 0.001 0.128 0.000 0.054 0.033 0.273 0.000 1.000 5 0.249 0.000 0.347 0.000 0.298 0.000 0.149 0.000 0.119 0.000 0.430 0.000 1.000 7 0.487 0.000 0.589 0.615 0.000 0.000 0.237 0.159 0.000 0.000 0.150 0.257 0.000 0.000 0.258 0.323 0.000 0.000 0.140 0.220 0.000 0.000 0.027 −0.076 0.294 0.003 0.197 0.000 0.131 0.000 0.128 0.000 0.081 0.001 0.076 0.003 1.000 6 0.372 0.000 0.256 0.000 0.165 0.000 0.270 0.000 0.276 0.000 0.217 0.000 0.128 0.000 0.160 0.000 1.000 9 0.178 0.000 0.160 0.549 0.000 0.000 0.385 0.439 0.000 0.000 0.168 0.289 0.000 0.000 0.073 −0.010 0.004 0.686 0.217 0.000 0.205 0.000 0.248 0.000 0.091 0.000 0.083 0.001 0.530 0.000 0.569 0.000 1.000 8 Appendix Table A6. Zero-order Pearson Correlation (Below Diagonal) Versus Partials (Above Diagonal). 0.435 0.000 0.439 0.000 0.547 0.000 0.243 0.000 0.162 0.000 0.171 0.000 0.233 0.000 0.288 0.000 0.429 0.000 0.411 0.000 1.000 11 −0.119 0.000 0.345 0.000 0.358 0.000 0.322 0.000 0.509 0.000 0.289 0.000 0.097 0.000 0.163 0.000 0.143 0.000 0.252 0.000 0.263 0.000 0.333 0.000 1.000 12 (continued) 0.429 0.000 0.315 0.377 0.000 0.000 −0.094 −0.044 0.000 0.088 0.236 0.000 0.360 0.000 0.160 0.000 0.290 0.000 0.301 0.000 0.114 0.000 0.218 0.000 0.133 0.000 0.533 0.000 1.000 10 33 17. Right-wing 16. Left-wing 15. Age 14. Education 1 r 0.091 p 0.000 r −0.147 p 0.000 r −0.057 p 0.026 r 0.309 p 0.000 Appendix Table A6. (continued) 0.115 0.000 −0.189 0.000 −0.055 0.032 0.290 0.000 2 0.088 0.001 −0.259 0.000 −0.144 0.000 0.318 0.000 3 0.214 0.000 0.108 0.000 0.106 0.000 0.211 0.000 4 0.111 0.000 0.062 0.015 0.092 0.000 0.079 0.002 5 7 0.164 0.159 0.000 0.000 −0.269 −0.259 0.000 0.000 −0.019 −0.021 0.461 0.419 0.110 0.159 0.000 0.000 6 9 0.120 0.098 0.000 0.000 −0.429 −0.050 0.000 0.051 −0.061 0.085 0.016 0.001 0.159 0.094 0.000 0.000 8 11 0.148 0.161 0.000 0.000 −0.073 −0.247 0.004 0.000 0.139 −0.055 0.000 0.030 0.074 0.223 0.004 0.000 10 0.223 0.000 −0.012 0.651 0.053 0.038 0.240 0.000 12