Social media and expressive citizenship: Understanding the relationships between social and entertainment expression on Facebook and political participation

Social media and expressive citizenship: Understanding the relationships between social and entertainment expression on Facebook and political participation

Accepted Manuscript Social media and expressive citizenship: Understanding the relationships between social and entertainment expression on Facebook a...

931KB Sizes 1 Downloads 45 Views

Accepted Manuscript Social media and expressive citizenship: Understanding the relationships between social and entertainment expression on Facebook and political participation Rebecca Ping Yu, Yu Won Oh PII: DOI: Reference:

S0736-5853(18)30460-X https://doi.org/10.1016/j.tele.2018.09.010 TELE 1174

To appear in:

Telematics and Informatics

Received Date: Revised Date: Accepted Date:

27 April 2018 23 September 2018 25 September 2018

Please cite this article as: Yu, R.P., Oh, Y.W., Social media and expressive citizenship: Understanding the relationships between social and entertainment expression on Facebook and political participation, Telematics and Informatics (2018), doi: https://doi.org/10.1016/j.tele.2018.09.010

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Social media and expressive citizenship: Understanding the relationships between social and entertainment expression on Facebook and political participation

Author name and affiliation *Rebecca Ping Yu Assistant Professor Department of Communication and Technology No.1, Sec. 1, Liujia 5th Rd., HK219B Zhubei City, Hsinchu County, Taiwan (R.O.C.) 302 886-3-5712121 ext. 58725 [email protected] Yu Won Oh Assistant Professor Department of Media and Communication Sejong University 209 Neungdong-ro (Sejong-gwan), Room 207, Gwangjin-gu, Seoul, 05006, Republic of Korea. 82-10-2126-4280 [email protected] *indicates the corresponding author Funding The work was supported by the Department of Communication Studies, University of Michigan, Rackham Graduate Student Research Grant, University of Michigan, and the Ministry of Science and Technology of the Republic of China (MOST 105-2410-H-009-065–SSS)

Abstract Substantial research has explored the political significance of social media use in the context of collective actions. Yet much remains unknown about whether common, day-to-day, nonpolitically oriented activities on social media relate to political engagement. Focusing on Facebook, the primary social media platform for most

1

Americans, this study investigates whether and how social and entertainment expression on the site are associated with political participation among a diverse online sample of American adults. Results show that social and entertainment expressive Facebook use are indirectly associated with political participation through political communication in the form of interactive political expression on the site. In addition, findings demonstrate that social expressive use is also conducive to political participation via offline political talk, but entertainment expressive use is not significantly related to political talk in offline settings. Further analysis shows that the interactions between political interest and each of the expressive uses are largely insignificantly related to political communication and participation. Overall, the study’s findings help to clarify the distinctions between the two types of nonpolitical Facebook use and the underlying process by which these uses contribute to political participation. Keywords: Social expressive use, entertainment expressive use, interactive political expression, political talk offline, political participation

2

1. Introduction Peer socializing and entertainment activities on platforms like Facebook have become deeply woven into the fabric of social media users’ daily lives. According to a 2018 Pew survey, 68% of Americans use Facebook, 74% of whom visit the site at least once a day (Smith & Anderson, 2018). While many studies have explored the political significance of social media use in the context of collective actions, relatively little is known about whether or how the common, day-to-day, nonpolitical activities on the platform relate to political engagement. In particular, existing research regarding the link between seemingly nonpolitical social media practices and pro-democratic behaviors remains inconclusive. While some research shows that social- or entertainment-oriented media use distracts users from democratic engagement (e.g., Y. Kim, Chen, & Gil de Zúñiga, 2013; Shah, Kwak, & Holbert, 2001), other studies demonstrate that nonpolitically oriented social media uses can promote democratic attitudes or behaviors (e.g., Diehl, Weeks, & Gil de Zúñiga, 2015; Gil de Zúñiga, Molyneux, & Zheng, 2014; Penney, 2014; Yu, 2016). Drawing on the emergent expressive model of citizenship, which emphasizes the importance of expressing oneself in a social- or entertainment-based networked environment, this study investigates whether and how nonpolitically oriented but expressive social media uses relate to political participation. With a focus on Facebook, the primary social media platform for Americans, we propose a theoretical model that expects that social expressive Facebook use (e.g., sharing content about one’s personal life) and entertainment expressive Facebook use (e.g., sharing content about one’s entertainment interests) are both positively associated with political communication, both on the site (i.e., interactive political expression) and offline (i.e., political talk offline); we further expect this Facebook-related political expression to be positively related to traditional forms of offline political participation. Moreover, due to the longstanding concerns about unequal 3

participation between those who are more politically interested and those who are less so (Bimber, Cunill, Copeland, & Gibson, 2015; Xenos & Moy, 2007; Xenos, Vromen, & Loader, 2014), we test the moderating role of political interest in the associations between social and entertainment expressive uses and political communication and participation. The findings of our study contribute to prior research in three important ways. First, in contrast to the traditional citizenship model, which identifies being informed about public affairs as promoting democratic engagement, this study demonstrates that social and entertainment expressive Facebook use are both conducive to interactive political expression on the same site, lending support to the alternative expressive-citizenship model. Second, the differential results between social and entertainment expressive Facebook use and political talk offline highlights the distinctions between the two forms of expression, affirming the importance of differentiating social media use, rather than treating it as a uniform activity. Third, by testing the mediating role of political communication on Facebook and offline in the links between social and entertainment expressive uses and political participation, and the moderating role of political interest in the associations between social and entertainment expressive uses and political outcomes, this study helps explain the process and mechanism by which social and entertainment expression on Facebook are transformed into political actions. 2. Literature review 2.1 From informed citizenship to expressive citizenship: Theorizing expressive practices on social media With the ascent of social media, increasing scholarship has contended that democratic citizenship models have shifted from informed citizenship to expressive citizenship. The traditional citizenship model, which emphasizes the importance of being informed about current affairs and voting accordingly, argues that

4

people are less politically engaged than in the past because they increasingly turn away from news consumption while embracing social- and entertainment-related activities (e.g., Prior, 2007; Stoycheff, Nisbet, & Epstein, 2016). In contrast to this decline narrative, a number of scholars have proposed an newly emergent citizenship model, positing that what we see should not be viewed as a decline but rather as a change in the form of political engagement, one that values expressing oneself and connecting to others based on loosely connected social relationships and entertainment interests (for a review, see Kligler-Vilenchik, 2017). For example, Bennett and his colleagues (Bennett, 2008; Bennett, Wells, & Rank, 2009) argue that whereas traditional concepts of “dutiful citizenship” look to news consumption from mainstream media sources as the primary means through which individuals gain civic awareness and engage in politics, a new model of “actualizing citizenship” suggests that expression related to social and personal interests within peer-to-peer networks is likely to trigger political behaviors. Zuckerman’s (2014) participatory civics makes a similar claim, arguing that those who use digital media to share perspectives on a daily basis (e.g., sharing a photo about lunch) are more likely to employ the same tools to engage in civic activities (e.g., sharing photos about collective actions). Likewise, Jenkins’s (2006) idea of participatory culture emphasizes the increased opportunities offered by digital media for people to express interests and cultivate social relations, thus laying the foundation for expressive forms of civic and political participation. Together, these perspectives frame political change as something that can be promoted not only through established political institutions like political parties, but also through forms of social and cultural expression that only appear to be nonpolitical and unconnected to democratic behaviors. In this study, we give more systematic consideration to the emergent expressive-citizenship model by

5

empirically testing whether and how seemingly nonpolitical expressive acts on Facebook are associated with political participation. Based on prior work that emphasizes the mediating role of political communication to transform media use into political action (Ito et al., 2015; Kligler-Vilenchik, 2013), we unpack the expressive-citizenship model by first examining the extent to which nonpolitically oriented expression on Facebook encourages political communication both on the site (i.e., interactive political expression) and offline (i.e., face-to-face political talk). In the following section, we review literature that examines how forms of political communication are associated with social and entertainment expressive Facebook use. 2.2 Types of nonpolitical expressive use on Facebook and forms of political communication To shed light on how nonpolitical expressive acts on Facebook relate to political communication on the site and offline, this study differentiates two forms of nonpolitical expressive use: social and entertainment. Prior work has researched social and entertainment uses for various kinds of media, including the Internet (December, 1996; Eighmey & McCord, 1998; Papacharissi & Rubin, 2000), mobile phones (Campbell & Kwak, 2010), and social media (Chen, 2015; Park, Kee, & Valenzuela, 2009; Skoric, Zhu, Goh, & Pang, 2015). Social expressive use involves disclosure of everyday personal experiences related to school, work, family, and relationships (Vraga, Bode, Smithson, & Troller-Renfree, 2016), whereas entertainment-related expressive use typically centers on expression of personal interests, such as gaming, sports, and music (Delli Carpini, 2012; Stoycheff et al., 2016). Given their topical differences, social and entertainment expression may each trigger different interaction patterns. Because sharing one’s personal life on Facebook is more appropriate and relevant to one’s offline contacts, social expressive use may facilitate both online and offline interactions with existing friends and family members. By contrast, entertainment expressive use tends to foster interactions with those

6

who share similar interests but are not necessarily offline contacts (boyd, 2014). 2.3 Social and entertainment expressive use and interactive political expression on Facebook We first examine how social- and entertainment-oriented expression on Facebook are associated with political communication on the site, which we define here as interactive political expression, such as sharing content about political and social issues. It is important to note that because social media like Facebook is characterized by mass personal communication in which the boundary between interpersonal and mass communication is blurred (Carr & Hayes, 2015; Marwick & boyd, 2011), people tend to be cautious about voicing their political views on social media because of the difficulty of limiting one’s potential audience and the possibility of misinterpretation (boyd, 2011; Thorson, 2014). At the same time, the supporting networked environment that is cultivated through social and entertainment expressive acts may inspire users to express their political views on the site. Indeed, prior work illustrates that expressive acts on Facebook, such as “responding to friends’ requests,” are essential mechanisms through which users maintain their relationships (Ellison, Vitak, Gray, & Lampe, 2014). Through this process, norms of reciprocity and trust are likely to be established within users’ social media networks (Ellison et al., 2014), which may encourage users to express their political views in front of social media connections when opportunities arise. Similarly, Jenkins, Itō, and boyd (2016) argue that entertainment expressive acts, such as producing music, comics, and videos online, are building blocks of political engagement, because such acts help users learn skills of expression and become part of “latent publics” or networks that can later be activated for political causes. Indeed, there is evidence that those who actively generate or share content like music, videos, and blog posts online are more likely to express their political views in online spaces (Ekström & Östman, 2013). We therefore expect that users who frequently

7

engage in social and entertainment expressive activities on Facebook will also be more likely to engage in interactive political expression on the site. H1: Social expressive Facebook use is positively associated with interactive political expression on Facebook. H2: Entertainment expressive Facebook use is positively associated with interactive political expression on Facebook. 2.4 Social and entertainment expressive use and political talk offline In contrast to interactive political expression on Facebook, which may be seen by a larger and wider array of individuals in the network, political talk offline is often bounded within a closed network of family and friends (e.g., Conover, Searing, & Crewe, 2002; Eliasoph, 1998; Gerber, Huber, Doherty, & Dowling, 2012). Social media use may still relate to political expression offline, however. Because social expressive activities on social media like Facebook create topics for conversation and cultivate relationships with existing friends and family offline (e.g., Cornejo, Tentori, & Favela, 2013; Erickson, 2011), social expressive acts on such a platform may increase opportunities for political conversations in offline settings as well. Indeed, recent research suggests that Facebook use is conducive to offline political participation by facilitating interaction with those with whom one already has strong relationships (Halpern, Valenzuela, & Katz, 2017; Valenzuela, Correa, & Gil de Zúñiga, 2018). We therefore hypothesize the following: H3: Social expressive Facebook use is positively associated with political talk offline. By contrast, the link between entertainment expressive Facebook use and offline political talk is less clear. It may be that entertainment expressive use of social media, like social expressive use, may facilitate offline interactions and create opportunities for political talk offline. Yet other research finds that entertainment

8

expressive activity online has insignificant effects on political talk offline (Ekström & Östman, 2013), perhaps because entertainment expressive use fosters interactions with people who share one’s interests or hobbies but are not necessarily contacts offline (boyd, 2014). Accordingly, entertainment expressive Facebook use may contribute little to political talk offline with existing friends and family. Given our limited knowledge of the effects of entertainment expressive use, we pose the following research question: RQ1: What is the relationship between entertainment expressive Facebook use and political talk offline? 2.5 Forms of political communication and political participation Substantial research has identified the important mediating role of political communication—both interactive political expression on social media and political talk in face-to-face settings—in connecting news media use and democratic participation (Gil de Zúñiga et al., 2014; Lee, Shah, & McLeod, 2012; Shah, Cho, Eveland, & Kwak, 2005; Shah et al., 2007). This area of study sees these forms of communication as a key conduit to democratic citizenship because engaging in political communication provides opportunities for individuals to develop, reflect upon, and affirm their political views (Carpini, Cook, & Jacobs, 2004; J. Kim & Kim, 2008; Tetlock, Skitka, & Boettger, 1989). As suggested by the expressive-citizenship model, political communication practices are also likely to strengthen the confidence and core skills essential for offline political participation, transforming individuals from political observers to political participants. Much work has established face-to-face political talk as a precursor of civic or political actions (Huckfeldt & Sprague, 1995; Shah et al., 2005; Wyatt, Katz, & Kim, 2000). Such conversations, whether casual or serious, enable participants to exchange information, elaborate on important issues facing a society (Gastil & Dillard, 1999), and learn about opportunities for democratic participation (Klofstad, 2007). Political

9

communication, in short, constitutes the “soul of democracy” (J. Kim, Wyatt, & Katz, 1999, p. 392) and functions as an important mechanism through which political opinions and attitudes can be transformed into democratic actions. The democratic importance of interactive political expression on social media is somewhat contested. Some dismiss interactive political expression online as “slacktivism,” arguing that it only distracts users from more consequential political participation (Gladwell, 2010; Morozov, 2011). However, empirical studies have suggested otherwise. Gil de Zúñiga et al. (2014) demonstrate a positive association between political expression on social media and political participation. Vaccari et al. (2015) also found that those who frequently post something about politics are more likely to attend offline political events after receiving online invitations. Other research shows that political expression on social media reinforces users’ preexisting political preferences (Cho, Ahmed, Keum, Choi, & Lee, 2018), suggesting that expressing oneself politically on Facebook would be likely to increase, rather than decrease, offline participation. Overall, empirical evidence suggests that the engagment in low-threshold interactive political expression on social media may pave the way for high-threshold political participation. Based on the above literature, we hypothesize the following: H4: Political talk offline is positively associated with political participation. H5: Interactive political expression on Facebook is positively associated with political participation. Thus far, we have proposed relationships between social and entertainment expressive use and forms of political communication and between forms of political communication and political participation. Implicitly suggested by the proposed associations, we expect that social and entertainment expressive use are indirectly

10

related to political participation through interactive political expression on Facebook and political talk offline. Therefore, we propose the following: H6: Social expressive Facebook use is indirectly related to political participation through (a) interactive political expression on the site and (b) political talk offline. H7: Entertainment expressive Facebook use is indirectly related to political participation through interactive political expression on the site. 2.6 Political interest as a moderator Given the associations outlined above, we now turn to the possibility that the proposed associations between social and entertainment expression and political outcomes might be stronger or weaker depending on political interest. We focused on political interest because it is a central theoretical question in the study of social media and politics to understand whether social and entertainment expressive uses tend to widen or narrow the participation gaps between those who are more politically interested and those who are less so (Bimber et al., 2015). Two perspectives—the reinforcement hypothesis and mobilization hypothesis—may lead to divergent expectations. The reinforcement hypothesis suggests that because those who are more politically interested are intrinsically motivated (Prior, 2010), they may seek out more news, learn more about politics, and participate in civic and political activities at higher rates than the less politically interested (Xenos & Moy, 2007). In this regard, political interest may facilitate the transition from social and entertainment expression to political engagement, widening the existing gaps in political participation. By contrast, the mobilization hypothesis posits that those who are less politically interested are as likely or more likely to incidentally encounter political

11

information or even dissonant political content in social- and entertainment-oriented online spaces compared to those who are more politically interested (Valeriani & Vaccari, 2015; Wojcieszak & Mutz, 2009). Thus, participation in communities that center on social and recreational interests may inadvertently mobilize the less politically interested to engage in the political process, softening the existing participatory gaps by political interest (Jenkins et al., 2016). Given the contrary expectations, the following research questions are proposed: RQ2: Is the interaction between political interest and social expressive Facebook use associated with (a) interactive political expression on the site, (b) political talk offline and (c) political participation? RQ3: Is the interaction between political interest and entertainment expressive Facebook use associated with (a) interactive political expression on the site, (b) political talk offline and (c) political participation? 3. Method 3.1 Sample The current study used data from an online survey conducted through Qualtrics (an online survey panel provider) in August 2014. Adult Americans (aged 18 years and older) who were part of the Qualtrics online panel were invited via methods such as web banners, website referrals, and email invitation to complete the survey in exchange for payment. Since the current study focused on Facebook users, 727 individuals who self-identified as Facebook users were included. Comparing this group with a nationally representative sample from the Pew Research Center’s Internet and American Life Project of US adults who used Facebook in 2014 (Duggan, Ellison, Lampe, Lenhart, & Madden, 2015), the age of the sample in the study was slightly younger (MCurrent study = 46.06; MPew = 48.97), the proportion of males was lower (MaleCurrent study = 36%; MalePew = 47%), but income and education were comparable to the Pew sample

(income—MdnCurrent study = $50,000 to $69,999;

MdnPew = $50,000 to under $75,000; education— MdnCurrent study = some college; MdnPew = some college). As 12

several large-scale online and offline social movements in the United States, such as marriage equality and the Ferguson protests, were associated with use of social media platforms like Facebook (e.g., Clark, 2016; Penney, 2014), we use this sample to shed light on how the everyday Facebook expressive practices that are seemingly nonpolitical in nature are associated with political engagement online and offline. 3.2 Measures 3.2.1 Predictor variables Types of nonpolitical expressive Facebook use. Different types of expressive Facebook use were captured by asking how often respondents engage in three activities on a 5-point scale ranging from 1 (never) to 5 (very frequently), including “liking,” “commenting on,” and “posting or sharing,” in loops in reference to two topical domains: personal life (e.g., work, school, relationships, or family) and entertainment interests (e.g., sports, movies, food, or music). Based on the question structure, social and entertainment expressive uses were created by combining the three modes of use related to personal life (3 items:  = .88, M = 8.85, SD = 2.91) and entertainment interests (3 items:  = .92, M = 8.08, SD = 3.14), respectively. To determine whether the data supported separating social and entertainment expressive uses, confirmatory factor analysis (CFA) was conducted to compare the one-factor structure with the two-factor structure. The results of CFA show that the two-factor structure (χ2(8) = 90.58, p < .001, CFI = .98, NFI = .97, RMSEA = .12, AIC = 128.58) fit the data better than did the one-factor model (χ2(9) = 632.23, p < .001, CFI = .81, NFI = .81, RMSEA = .31, AIC = 668.23). Although the fit for the data could be better given the marginal RMSEA, the goodness of fit statistics provide support for the structure of the two variables. Interactive political expression on Facebook. Three items asked how often respondents expressed

13

opinions about politically related topics (e.g., elections, government, human rights, and economics) on Facebook. Three modes of political expression, including “liking,” “commenting on,” and “posting or sharing” were asked in loops in reference to political and social issues. The response scale ranged from 1 (never) to 5 (very frequently). A composite measure was created by summing three types of political expression (3 items:  = .94, M = 7.07, SD = 3.30). Political talk offline. Using a 5-point scale (1 = never to 5 = very frequently), respondents provided an estimate of how often they talk about political and social issues with “friends” and “family” in person. A composite measure was formed by adding the two items (2 items:  = .83, M = 5.56, SD = 2.08). 3.2.3 Outcome variable Political participation. Respondents were asked to indicate “whether they have engaged in the activities during the past two years,” including “attended a political meeting, rally, or speech,” “contributed money to a political campaign or cause,” “wore/displayed a campaign button, bumper sticker, or yard sign,” “signed a petition for a candidate or cause,” and “worked for a political party or candidate.” Response options were “yes” (coded as 1) and “no” (coded as 0). These items were combined into an additive index (5 items: M = .60, SD = 1.08). 3.2.3 Control variables Demographics, including age, gender, race, education, and income, were included in the model to control for potential confounding effects, as prior work shows that these demographic variables are predictive of political engagement (Brady, Verba, & Schlozman, 1995). In addition, news media use and a number of politically related variables were included as controls based on previous studies.

14

Newspaper, television, and web news use. Different forms of news media use were included as controls, because news media use is positively associated with political and civic outcomes (e.g., Shah et al., 2001). On an 8-point scale ranging from 0 (0 days) to 7 (7 days), respondents were asked how many days in the past week they consumed news from newspaper(s) (i.e., reading a print copy of a local newspaper; 1 item: M = 2.39, SD = 2.65), television (i.e., watching national nightly news on CBS, ABC, or NBC and watching local TV news; 2 items: M = 6.67, SD = 5.05), website (i.e., visiting national newspaper websites, visiting local newspaper website, visiting TV news websites; 3 items: M = 4.57, SD = 5.52). Political view. Political view was included as a control to account for the possibility that one’s ideological inclination may influence his or her political expression, as prior work shows that ideological inclination is likely to shape a set of value assumptions, such as expressing political views (e.g., Converse, 1969). This variable was measured on a 5-point scale ranging from 1 (very liberal) to 5 (very conservative;1 item: M = 3.03, SD = 1.07). Political interest. This study controls for the influence of political interest because this construct is predictive of political participation. To assess political interest, respondents were asked, “How much interest do you have in politics,” on a 5-point scale ranging from 1 (not at all to 5 (a great deal) (1 item: M = 3.09, SD = 1.17). Political efficacy. Political efficacy was included as a control because previous research has consistently demonstrated a positive link between political efficacy and political participation (e.g., Kenski & Stroud, 2006). This variable was captured by asking respondents, “How much do you think that people like you can influence decisions made by government officials?” on a 5-point scale ranging from 1 (not at all) to 5 (a great deal) (1

15

item: M = 2.53, SD = 1.11). Table 1 summarizes zero-ordered correlations among all the variables included in the study. [Insert Table 1 about here] 3.3. Analysis strategy To test the proposed direct and indirect associations, structural equation modeling (SEM) was used. The analysis was conducted in Mplus 8.1. The first structural model includes the five latent variables: social expressive Facebook use, entertainment expressive Facebook use, interactive political expression on Facebook, political discussion offline, and political participation. In the first structural model, weighted least squares with means and variance adjusted estimation (WLSMV) was used because of binary indicators of political participation. In addition to the five latent factors, the effects of demographics (age, gender, race, education, and income), news media use (newspaper, television, and web-news use), and the politically related variables (political interest, political view, and political efficacy) on the five endogenous latent variables were controlled. Using the latent moderated structural equations (LMS) method, the second structural model tested whether political interest moderates the associations between social and entertainment expressive uses and political communication and participation. Compared to the regression-based approach that assumes no measurement errors, LMS corrects for measurement errors when estimating latent interaction effects (Cheung & Lau, 2017; Sardeshmukh & Vandenberg, 2017). The interaction between political interest and each of the two forms of nonpolitically oriented expression on Facebook and political participation were defined using the XWITH statement, with maximum likelihood (ML) and 500 bootstrap draws. Results

16

Prior to estimating structural models to test our hypotheses and research questions, we evaluated the measurement model of five latent variables: social expressive use, entertainment expressive use, interactive political expression on Facebook, political talk offline, and political participation. To evaluate the model fit, prior work shows that nonsignificant χ2 test (p > .05), CFI (comparative fit index), and TLI (Tucker-Lewis index) values greater than .95, and RMSEA (root mean square error of approximation) values below .08 generally constitute good fit (Kline, 2005). While a χ2 statistic is often reported, other fit indices, such as CFI, TLI, and RMSEA, are typically relied on more heavily in evaluation of the model fit because χ2 is highly sensitive to sample size (Kline, 2005; Little, 2013). From this perspective, the measurement model of this study demonstrates acceptable goodness-of-fit values (χ2(94) = 143.35, p < .01, CFI = .98, TLI = .97, RMSEA = .03), suggesting that the five-factor model fit the data well. We compare the current model with two plausible alternative models. Results indicate that the presumed model has a better fit than other models. Complete information of the model comparisons is described in detail in Supplementary Appendix A. The estimated factor loadings of the indicators on their respective latent variables are reported in Table 2. All the indicators load relatively strongly on their latent variables (the standardized factor loadings range from .68 to .92), and each is statistically significant (p < .001), indicating convergent validity (Brown, 2006). The correlations among latent variables range from .33 (the correlation between family Facebook use and offline political talk) to .73 (the correlation between social and entertainment Facebook use), less than the .85 value that indicates problematically poor discriminant validity (Brown, 2006). Turning to the structural equation models, the first structural equation model exhibits a good fit with the data with regard to CFI (.97), TLI (.96), and RMSEA (.03) but not with the χ2 test (χ2(215) = 319.32, p < .001).

17

The values of CFI, TLI, and RMSEA suggest that the structural model fits the data well, explaining 57% of the variances for political expression on Facebook, 41% of the variance in political talk offline, and 37% of the variance in political participation. In the second structural model that includes interaction terms, because the LMS approach does not report typical fit indices, a baseline model that does not include interaction terms is compared with the model with interaction terms. The interaction model is a better-fitting model than the baseline model because the former has a smaller AIC (AICinteraction = 22764.76; AICbaseline = 22777.39), and the difference in the AIC of the two models is greater than 10 (Burnham & Anderson, 2002). The full results of the structural equation models are shown in Supplementary Appendix B. [Insert Table 2 about here] In the first and second hypotheses, we predicted that social (H1) and entertainment (H2) expressive Facebook use were positively associated with interactive political expression on Facebook, respectively. As Figure 1 shows, both social (ß = .31, p < .001) and entertainment expressive Facebook use (ß = .31, p < .001) are positive predictors of interactive political expression on Facebook, supporting H1 and H2. [Insert Figure 1 about here] The third hypothesis and a research question explored the relationships between social (H3) and entertainment expressive Facebook use (RQ1) and political talk offline. As Figure 1 shows, social expressive Facebook use is positively associated with political talk offline (ß = .22, p < .001), thus we find support for H3. Regarding RQ1, Figure 1 reveals that the relationship between entertainment expressive use and political talk offline is not statistically significant (ß = .05, p = .314). The fourth and fifth hypotheses predicted that interactive political expression on Facebook and political

18

talk offline were positively associated with political participation, respectively. Figure 1 shows that both interactive political expression on Facebook (ß = .21, p < .01) and political talk offline (ß = .18, p < .01) are significant and positive predictors of political participation. Thus, both H4 and H5 are supported. Turning to indirect relationships, we predicted indirect relationships between social expressive use and political participation via interactive political expression on Facebook (H6a) and political talk offline (H6b). As Table 3 shows, the results support H6a, revealing that the indirect effect of social expressive Facebook use on political participation through interactive political expression on Facebook is statistically significant (ß = .06, p < .01). Similarly, the results support H6b, confirming that social expressive Facebook use significantly facilitates political participation indirectly via political talk offline (ß = .04, p < .05). Together, these results indicate that social expressive acts on Facebook enhance political participation indirectly through both forms of political communication. As expected in H7, further analysis shows that entertainment expressive use is positively associated with political participation indirectly through interactive political expression on Facebook (ß = .06, p < .01; see Table 3). These results suggest that entertainment expressive activities foster political participation primarily and indirectly through interactive political expression on the site. [Insert Table 3 about here] Finally, we evaluate whether the interactions between political interest and social and entertainment expressive uses are associated with political outcomes (RQ2 to RQ3). As Table 4 shows, most of the interaction terms are not significantly associated with political outcomes (RQ2b, RQ2c, RQ3a, RQ3b, and RQ3c). One exception is that political interest significantly interacts with social expressive use to promote interactive political

19

expression on Facebook (b = .10, p < .05). This finding suggests that those who are high in political interest and are more involved in social expressive use are more likely to express themselves politically on the platform (RQ2a). Thus, while political interest moderates the association between social expressive use and interactive political expression on Facebook, political interest plays little moderating role in other associations. [Insert Table 4 about here] 4. Discussion Drawing on the emergent citizenship models that suggests civic behaviors may be cultivated through self-expressive acts in the networked social media environment, this study investigates the associations between social and entertainment expressive Facebook use, forms of political communication, and political participation. Results show that seemingly nonpolitically oriented expressive actions create pathways to political participation indirectly through political expression on the platform and through political talk offline. In other words, nonpolitically oriented expressive activities only seem nonpolitical; in fact, they do have political significance. Several implications can be drawn from these findings. First, the study lends support to emergent expressive-citizenship models. In particular, we found that Facebook users who frequently share about their personal life and entertainment interests online are also more likely to engage in interactive political expression on the site. These findings are contrary to prior work suggesting that the use of digital media for social or recreational purposes may distract citizens from public engagement (e.g., Shah et al., 2001; Stoycheff et al., 2016). One possible explanation for the conflicting findings is that the results of earlier work, such as Shah et al. (2001), may not be generalizable over time because in the early years of Internet use, those who had Internet access may have had few friends and family members with similar access, which would have limited the

20

opportunities for meaningful interactions (Kraut et al., 2002). Moreover, the measures of entertainment use in previous studies tend not to explicitly consider social or entertainment-related expressive practices. For instance, according to Stoycheff et al.(2016), entertainment use is measured by the frequency of “playing video games and watching videos, movies and TV shows” (p. 9). Indeed, with a focus on social and entertainment expressive acts on Facebook, our study shows that these practices are beneficial for political expression on the platforms, perhaps because users may acquire and practice skills that are essential for political expression (Jenkins, 2006) or may be exposed to social or public affairs incidentally (Valeriani & Vaccari, 2016; Wojcieszak & Mutz, 2009) through these expressions. This, in turn, may motivate users to express themselves politically when appropriate in order to seek approval from others (Lilleker & Koc-Michalska, 2017). From this perspective, expression of social and entertainment interest in a networked social media environment is an important step toward expressive citizenship. Our findings also shed light on the process and mechanism through which social and entertainment expressive activities facilitate offline political participation. It is worth noting that while findings show that the direct associations between both expressive activities and political participation are not statistically significant, social and entertainment expressive Facebook use only indirectly fosters political participation via political expression on the site or political talk offline. The pattern of results is consistent with prior work that suggests that rather than spurring political actions directly, nonpolitically oriented social media uses may trigger political participation through a more transformative process (Ito et al., 2015; Kligler-Vilenchik, 2013). When individuals are encouraged to discuss current social and political issues in a supportive environment—either online or offline—that is organized around nonpolitically oriented expressive activities, they may have more

21

confidence, core skills, and external motivations to take subsquent political actions (Jenkins, 2006; Kligler-Vilenchik, 2013; Lilleker & Koc-Michalska, 2017). Further analysis shows that those who are more politically interested and are more involved in social expressive use are more likely to express themselves politically on the site. Aside from this significant result, for the most part, political interest plays a relatively small moderating role in the associations between social and entertainment expressive use and political outcomes. This finding is consistent with some prior work (e.g., Xenos et al., 2014) that suggests that these patterns of social media use may help soften unequal participatory gaps between the more interested and the less interested. Together, the findings suggest that interactive political expression on Facebook and political talk offline are both important mechanisms that help transform social and/or entertainment expressive activities into offline political involvement, and such transformations are largely consistent across the more and less politically interested. Importantly, our findings do show different patterns of associations between the two forms of expressive acts and political talk offline. Specifically, it appears that social and entertainment expressive uses may lead to different political outcomes: social expressive use is positively associated with political talk offline, whereas entertainment expressive use is not a significant predictor of this outcome. A possible explanation for this is that Facebook users who frequently engage in social or entertainment expressive uses may have different interaction patterns with their Facebook networks, which in turn may lead to different trends of political talk offline. Prior work has demonstrated that social media users may have different privacy attitudes and self-disclosure preferences based on different network structures and interaction patterns (Choi & Bazarova, 2014). In light of these findings, those who frequently engage in social expressive use (e.g., sharing something about family) may

22

be more likely to connect with their existing offline contacts on the site, which may result in more opportunities to interact and discuss politics with these contacts in offline settings. By contrast, for those who frequently engage in entertainment expressive use (e.g., sharing something about sports), possibly through a Facebook page or group, their interaction targets tend to be others who share similar interests or hobbies but may not be among a user’s offline contacts. As a result, enhanced entertainment expressive use may not increase political discussion with existing offline friends and family. Taken together, the inconsistent findings across the two nonpolitically oriented Facebook uses—social and entertainment—highlight the important differences between the two practices. Future research should continue to probe how different uses of social media may contribute to political outcomes in different ways. Our findings also suggest important directions for civic education. Social and entertainment expressive Facebook use only indirectly facilitate political participation via political expression on the site or political talk offline; therefore, it is important for civic educators to relate social media users’ social and entertainment experiences to current political issues and encourage them to express themselves politically. Even those who are less politically interested may start to think more about politics and see themselves as politically engaged through this process (Jenkins et al., 2016), thereby laying the foundation for subsequent political participation. Several potential limitations of this study should be noted, as they also suggest directions for future research. First, given that only cross-sectional data were used, the results of this study cannot indicate the directionality of the associations. Although the speculated directions of the associations in the study are consistent with prior work in this area (Boulianne, 2015), the use of longitudinal panel data is needed to clarify the causal relationships. Such research using longitudinal approaches would also help to explicate the causal

23

mechanisms behind the associations presented in this study. Second, while this study considers important confounding variables, it cannot control for every potential confounding variable. One potential factor for future work to consider is personality traits. Prior work shows that different personality traits predict political engagement (Gerber, Huber, Doherty, & Dowling, 2011). For instance, extraverted individuals are more likely to talk about politics at clubs, church, and the workplace (Hibbing, Ritchie, & Anderson, 2011). Future research can further explore the extent to which the direct and indirect associations examined in the study differ by various personality traits. Third, although the measure of political interest is widely used, it is a one-item measure. As results of the study show that nonpolitically oriented expressive activities on Facebook are indirectly associated with political participation via political expression on the site, the forms of political participation may have become more versatile. In such a context, a more comprehensive measure may be needed to better capture interest in politics. Another limitation of this study is the use of self-report data rather than actual measures of behaviors, which may bias the estimation of social media behaviors from events such as recall errors. While we have no reasons to expect systematically biased reporting in the measures, future research should use other modes of data collection, such as server-level behavioral measures, to better capture what users do on social media. Finally, though Facebook is the most widely used social media site in the United States and many other countries at the time of writing, future research should consider the affordances of different social media sites and further examine whether the findings are consistent across different sites and countries. Despite these limitations, this study is an important step in investigating the relationships between social

24

and entertainment expressive use, political communication, and political participation. Largely consistent with the emergent-citizenship models, our results suggest that social- and entertainment-oriented expressions are associated with political communication online and offline, which in turn triggers political participation. Further, political interest plays a relatively small role in moderating the associations between social and entertainment expressive uses and political outcomes. Taken together, results of the study shed light on the emergent expressive-citizenship models, while highlighting the importance of considering the varied forms of social media use and forms of political communication.

Appendices Appendix A: Results of Measurement Models To evaluate the dimensionality of the construct, we compare the original presumed model with two plausible alternative models. Alternative model 1 assumes that the indicators of social and entertainment use load on the same factor (see Table A1, column 2). As Table A2 shows, fit indices of alternative model 1 suggest a reasonably good fit (χ2 (98) = 216.67, p < .001, CFI =.95, TLI = .94, RMSEA = .04). Building on alternative model 1, alternative model 2 not only combines the social and entertainment related uses, but also assumes that the political expression on Facebook and offline political talk load on the same factor (see Table A1, column 3). Alternative model 2 does not fit the data well (χ2 (101) = 399.50, p < .001, CFI =.87, TLI = .85, RMSEA = .06). To further determine whether the original model fits significantly better or worse than the alternative models, χ2 difference tests and ΔCFI were used. In Table A2, the results of the χ2 difference tests show that the original model fit the data significantly better than alternative model 1 (Δχ2 (4) = 50.56, p < .001) and 2 (Δχ2 (7) =

25

160.26, p < .001). However, as a χ2 difference test is sensitive to sample size, it is recommended to consult ΔCFI that is relatively independent of sample size. Table A2 shows that compared to the original model, ΔCFI

of alternative models 1 and 2 are greater than .01 (.03 and .11, respectively), suggesting that the original model has a better fit. Thus, compared to the alternative models, the original model fits the data better.

26

Table A1. Factor loadings for measurement models Original model Parameter estimate Social FB use Social-related like Social-related comment Social-related post

Alternative model 1 Std. Parameter estimate

0.88 ***

0.91 ***

Social-related comment Social-related post

Interactive political

expression on FB

expression on FB

Post

0.92 ***

0.92 ***

0.92 ***

Political discussion offline Friends Family

0.78** *

0.77** *

Std.

0.92**

Like

*

0.92**

Comment

*

0.92**

Post

0.91

*

***

0.77 ***

*

0.77**

Family

*

Political participation 0.90 ***

Social-related post

27

*

0.77** *

Entertainment-related 0.79** * like Entertainment-related 0.83** * comment Entertainment-related 0.83** * post

Like Comment Post

Friends Family

0.89** *

0.89** *

0.88** *

0.57** *

0.49** *

Political participation **

Attended meeting

0.78**

0.91**

Friends

Social-related comment

Political communication

Political discussion offline

Political participation Attended meeting

Parameter estimate

Entertainment-related 0.79** * like Entertainment-related 0.83** * comment Entertainment-related 0.84** * post

Interactive political

Comment

Std.

FB social and entertainment FB social and entertainment related uses related uses ** 0.77 0.68 0.69** Social-related like Social-related like *** * *

Entertainment FB use Entertainment-related 0.86 *** like Entertainment-related 0.91 *** comment Entertainment-related 0.91 *** post

Like

Alternative model 2

0.90 *

Attended meeting

0.90** *

Contributed money Wore a campaign button Signed a petition

0.81 ***

0.87 ***

0.68 ***

Worked for a political 0.89 *** party

Contributed money Wore a campaign button Signed a petition

0.81** *

0.87** *

0.68** *

Worked for a political 0.89** * party

Note. Std. = standardized regression coefficients.

28

Contributed money Wore a campaign button Signed a petition

0.81** *

0.87** *

0.68** *

Worked for a political 0.90** * party

Table A2. Model fit statistics for measurement models 2

χ (df)

2

p of χ (df) CFI TLI RMSEA[90%CI]

Original model 143.35(94) < .01 Alternative model 1 216.67(98) < .001 Alternative model 2 399.50(101) < .001

.98 .97 .03[.02, .04] .95 .94 .04[.03, .05] .87 .85 .06[.06, .07]

Δχ2(df)

p of ΔCFI Δχ2(df)





50.56(4) < .001 160.26(7) < .001



0.03 0.11

Notes. df = degrees of freedom; CFI = comparative fit index; RMSEA = root mean square error of approximation; CI = confidence interval. A dash indicates no data.

Appendix B: Full Results of the Structural Equation Model Table B1: Coefficients from structural equation model 1, all variables Standardized coefficient

p-value

0.15

0.006

Age

0.15

0.006

Gender (Male)

0.06

0.153

Race (White)

-0.02

0.615

Education

0.05

0.286

Income

-0.08

0.089

Newspaper

0.03

0.598

TV news

0.04

0.482

News website

0.13

0.010

Political view

-0.07

0.108

Political interest

0.25

0.000

Political efficacy

0.03

0.554

Political talk offline

0.18

0.002

-0.15 0.01 0.21

0.051 0.870 0.001

Outcome: Political participation Predictors

Entertainment expressive FB use Social expressive FB use Political expression on FB Outcome: Political talk offline Predictors 29

Standardized coefficient

p-value

Age

-0.09

0.033

Gender (Male)

0.03

0.444

Race (White)

-0.07

0.059

Education

0.07

0.044

Income

0.04

0.260

Newspaper

0.05

0.149

TV news

0.03

0.482

News website

0.02

0.693

Political view

0.03

0.309

Political interest

0.46

0.000

Political efficacy

0.05

0.088

0.05 0.22

0.314 0.000

0.01

0.784

Gender (Male)

0.08

0.006

Race (White)

-0.02

0.576

Education

-0.05

0.086

Income

-0.03

0.323

Newspaper

-0.01

0.769

TV news

-0.03

0.308

News website

0.11

0.000

Political view

-0.01

0.699

Political interest

0.26

0.000

Political efficacy

0.09

0.001

0.31 0.31

0.000 0.000

-0.28

0.000

-0.14

0.000

Entertainment expressive FB use Social expressive FB use Outcome: Political expression on FB Predictors Age

Entertainment expressive FB use Social expressive FB use Outcome: Social expressive FB use Predictors Age Gender (Male) 30

Standardized coefficient

p-value

Race (White)

0.07

0.057

Education

-0.11

0.005

Income

0.00

0.915

Newspaper

0.06

0.196

TV news

0.02

0.658

News website

0.23

0.000

Political view

0.00

0.995

Political interest

0.05

0.225

Political efficacy

0.14

0.000

Age

-0.37

0.000

Gender (Male)

0.06

0.104

Race (White)

0.03

0.379

Education

-0.11

0.005

Income

0.00

0.909

Newspaper

0.12

0.002

TV news

0.10

0.009

News website

0.23

0.000

Political view

-0.04

0.217

Political interest

0.04

0.294

Political efficacy

0.13

0.000

Outcome: Entertainment expressive FB use Predictors

31

Table B2: Coefficients from structural equation model 2, all variables Uns.

SE

p-value

Baseline Outcome: Political participation Predictors Age Gender (Male) Race (White) Education

Uns.

SE

p-value

Interaction

0.03 0.35 -0.08 0.14

0.01 0.32 0.36 0.17

0.014 0.280 0.827 0.419

0.03 0.37 -0.10 0.13

0.01 0.32 0.36 0.18

0.017 0.239 0.783 0.474

Income

-0.16

0.12

0.159

-0.15

0.12

0.192

Newspaper TV news News website Political view Political interest Political efficacy Political talk offline Interactive political expression on FB

0.02 0.03 0.08 -0.25 0.66 0.11 0.70 0.63

0.06 0.03 0.04 0.14 0.20 0.13 0.23 0.26

0.683 0.420 0.028 0.076 0.001 0.425 0.003 0.015

0.02 0.03 0.08 -0.25 0.49 0.12 0.71 0.63

0.06 0.03 0.04 0.14 0.22 0.13 0.24 0.26

0.678 0.363 0.024 0.077 0.026 0.373 0.003 0.017

0.01 -0.53 -

0.29 0.29 -

0.961 0.063 -

0.03 -0.47 -0.28 -0.28

0.31 0.30 0.26 0.30

0.916 0.118 0.293 0.812

Age Gender (Male) Race (White)

-0.01 0.07 -0.14

0.00 0.08 0.08

0.032 0.392 0.082

-0.01 0.06 -0.14

0.00 0.08 0.08

0.035 0.440 0.071

Education Income Newspaper TV news News website Political view Political interest

0.08 0.03 0.02 0.00 0.00 0.03 0.39

0.04 0.02 0.02 0.01 0.01 0.03 0.04

0.052 0.225 0.254 0.607 0.651 0.325 0.000

0.08 0.03 0.02 0.01 0.00 0.03 0.41

0.04 0.02 0.02 0.01 0.01 0.03 0.05

0.041 0.266 0.268 0.595 0.777 0.314 0.000

Social expressive FB use Entertainment expressive FB use Social X Political Interest Entertainment X Political Interest Outcome: Political talk offline Predictors

32

Uns.

SE

p-value

Uns.

SE

p-value

0.05 0.22 0.06 -

0.04 0.07 0.06 -

0.211 0.001 0.360 -

0.05 0.21 0.07 -0.06 0.07

0.04 0.07 0.07 0.05 0.05

0.215 0.002 0.268 0.240 0.141

0.00 0.17

0.00 0.06

0.998 0.007

0.00 0.16

0.00 0.06

0.934 0.008

Race (White)

-0.03

0.07

0.641

-0.01

0.07

0.841

Education Income Newspaper TV news News website Political view Political interest Political efficacy

-0.05 -0.02 0.00 -0.01 0.02 -0.01 0.23 0.09

0.03 0.02 0.01 0.01 0.01 0.03 0.03 0.03

0.051 0.326 0.798 0.369 0.000 0.777 0.000 0.006

-0.05 -0.02 0.00 -0.01 0.02 0.00 0.28 0.08

0.03 0.02 0.01 0.01 0.01 0.03 0.04 0.03

0.057 0.326 0.938 0.175 0.001 0.895 0.000 0.012

0.32 0.34 -

0.06 0.07 -

0.000 0.000 -

0.33 0.33 0.10 0.00

0.06 0.07 0.05 0.05

0.000 0.000 0.029 0.995

-0.01 -0.28 0.14

0.00 0.07 0.09

0.000 0.000 0.093

-0.01 -0.27 0.15

0.00 0.07 0.09

0.000 0.000 0.077

-0.11 0.01 0.02 0.00 0.04 0.00 0.10

0.04 0.03 0.02 0.01 0.01 0.03 0.04

0.009 0.846 0.161 0.693 0.000 0.948 0.004

-0.10 0.01 0.02 0.00 0.04 0.00 0.11

0.04 0.03 0.02 0.01 0.01 0.03 0.03

0.012 0.840 0.151 0.635 0.000 0.929 0.001

Political efficacy Social expressive FB use Entertainment expressive FB use Social X Political Interest Entertainment X Political Interest Outcome: Political expression on FB Predictors Age Gender (Male)

Social expressive FB use Entertainment expressive FB use Social X Political Interest Entertainment X Political Interest Outcome: Social expressive FB use Predictors Age Gender (Male) Race (White) Education Income Newspaper TV news News website Political view Political efficacy

33

Uns.

SE

p-value

Uns.

SE

p-value

-0.02 0.12 0.07 -0.11 0.00 0.05 0.02

0.00 0.07 0.08 0.04 0.03 0.02 0.01

0.000 0.101 0.374 0.006 0.874 0.004 0.022

-0.02 0.13 0.07 -0.11 0.01 0.05 0.02

0.00 0.07 0.08 0.04 0.03 0.02 0.01

0.000 0.083 0.355 0.007 0.859 0.004 0.018

News website

0.04

0.01

0.000

0.04

0.01

0.000

Political view Political efficacy

-0.04 0.12

0.03 0.04

0.202 0.001

-0.04 0.13

0.03 0.03

0.190 0.000

Outcome: Entertainment expressive FB use Predictors Age Gender (Male) Race (White) Education Income Newspaper TV news

Notes. Unst. = unstandardized factor loading; SE = standard error of the unstandardized factor loadings; A dash indicates no data.

References Bennett, W. L. (2008). Changing citizenship in the digital age. In W. L. Bennett (Ed.), Civic life online: Learning how digital media can engage youth (Vol. 1, pp. 1-24). Cambridge, MA.: MIT Press. Bennett, W. L., Wells, C., & Rank, A. (2009). Young citizens and civic learning: Two paradigms of citizenship in the digital age. Citizenship Studies, 13(2), 105-120. doi:10.1080/13621020902731116 Bimber, B., Cunill, M. C., Copeland, L., & Gibson, L. (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. doi:10.1177/0894439314526559 Boulianne, S. (2015). Social media use and participation: A meta-analysis of current research. Information, Communication & Society, 18(5), 524-538. doi:10.1080/1369118X.2015.1008542 boyd, d. (2011). Social network sites as networked publics: Affordances, dynamics, and implications. In Z. Papacharissi (Ed.), A networked self: Identity, community, and culture on social network sites. (pp. 39-58). New York: Routledge. 34

boyd, d. (2014). It's complicated: the social lives of networked teens. New Haven: Yale University Press. Brady, H. E., Verba, S., & Schlozman, K. L. (1995). Beyond SES: A resource model of political participation. American Political Science Review, 89(02), 271-294. doi:doi:10.2307/2082425 Burnham, K. P., & Anderson, D. R. (2002). Model selection and multimodel inference: a practical information-theoretic approach. New York: Springer. Campbell, S. W., & Kwak, N. (2010). Mobile communication and civic life: Linking patterns of use to civic and political engagement. Journal of Communication, 60(3), 536-555. doi:10.1111/j.1460-2466.2010.01496.x Carpini, M. X. D., Cook, F. L., & Jacobs, L. R. (2004). Public deliberation, discursive, participation and citizen engagement: A review of the empirical literature. Annual Review of Political Science, 7(1), 315-344. doi:10.1146/annurev.polisci.7.121003.091630 Carr, C. T., & Hayes, R. A. (2015). Social media: Defining, developing, and divining. Atlantic Journal of Communication, 23(1), 46-65. doi:10.1080/15456870.2015.972282 Chen, G. M. (2015). Why do women bloggers use social media? Recreation and information motivations outweigh engagement motivations. New Media & Society, 17(1), 24-40. doi:10.1177/1461444813504269 Cheung, G. W., & Lau, R. S. (2017). Accuracy of parameter estimates and confidence intervals in moderated mediation models: A comparison of regression and latent moderated structural equations. Organizational Research Methods, 20(4), 746-769. doi:10.1177/1094428115595869 Cho, J., Ahmed, S., Keum, H., Choi, Y. J., & Lee, J. H. (2018). Influencing myself: Self-reinforcement through online political expression. Communication Research, 45(1), 83-111. doi:10.1177/0093650216644020 Choi, Y. H., & Bazarova, N. N. (2014). Self-disclosure characteristics and motivations in social media: Extending the functional model to multiple social network sites. Human Communication Research, 41(4), 480-500. doi:10.1111/hcre.12053 Clark, L. S. (2016). Participants on the margins:# BlackLivesMatter and the role that shared artifacts of engagement played among minoritized political newcomers on Snapchat, Facebook, and Twitter. International Journal of Communication, 10(26), 235-253. Conover, P. J., Searing, D. D., & Crewe, I. M. (2002). The deliberative potential of political discussion. British Journal of Political Science, 32(01), 21-62. doi:http://dx.doi.org.proxy.lib.umich.edu/10.1017/S0007123402000029 Converse, P. E. (1969). Of time and partisan stability. Comparative Political Studies, 2(2), 139-171. 35

Cornejo, R., Tentori, M., & Favela, J. (2013). Ambient awareness to strengthen the family social network of older adults. Computer Supported Cooperative Work (CSCW), 22(2–3), 309–344. doi:10.1007/s10606-012-9166-2 December, J. (1996). Units of analysis for Internet communication. Journal of Computer-Mediated Communication, 1(4), 0-0. doi:10.1111/j.1083-6101.1996.tb00173.x Delli Carpini, M. X. (2012). Entertainment media and the political engagement of citizens. In H. A. Semetko & M. Scammell (Eds.), The SAGE handbook of political communication (pp. 9-21). London: Sage. Diehl, T., Weeks, B. E., & Gil de Zúñiga, H. (2015). Political persuasion on social media: Tracing direct and indirect effects of news use and social interaction. New Media & Society. doi:10.1177/1461444815616224 Duggan, M., Ellison, N. B., Lampe, C., Lenhart, A., & Madden, M. (2015). Social Media Update 2014. Retrieved from Washington, DC: http://www.pewinternet.org/files/2015/01/PI_SocialMediaUpdate20144.pdf Eighmey, J., & McCord, L. (1998). Adding value in the information age: Uses and gratifications of sites on the World Wide Web. Journal of Business Research, 41(3), 187-194. doi:https://doi.org/10.1016/S0148-2963(97)00061-1 Ekström, M., & Östman, J. (2013). Information, interaction, and creative production: The effects of three forms of Internet use on youth democratic engagement. Communication Research. doi:10.1177/0093650213476295 Eliasoph, N. X. (1998). Avoiding politics: How Americans produce apathy in everyday life. Cambridge, New York: Cambridge University Press. Ellison, N. B., Vitak, J., Gray, R., & Lampe, C. (2014). Cultivating social resources on social network sites: Facebook relationship maintenance behaviors and their role in social capital processes. Journal of Computer-Mediated Communication, 19(4), 855-870. doi:10.1111/jcc4.12078 Erickson, L. B. (2011). Social media, social capital, and seniors: The impact of Facebook on bonding and bridging social capital of individuals over 65. Paper presented at the The Americas Conference on Information Systems (AMCIS), Detroit, MI, USA. Gastil, J., & Dillard, J. P. (1999). Increasing political sophistication through public deliberation. Political Communication, 16(1), 3-23. doi:10.1080/105846099198749 Gerber, A. S., Huber, G. A., Doherty, D., & Dowling, C. M. (2011). The big five personality traits in the political arena. Annual Review of Political Science, 14(1), 265-287. doi:doi:10.1146/annurev-polisci-051010-111659 Gerber, A. S., Huber, G. A., Doherty, D., & Dowling, C. M. (2012). Disagreement and the avoidance of political discussion: Aggregate relationships and differences across 36

personality traits. American Journal of Political Science, 56(4), 849-874. doi:10.1111/j.1540-5907.2011.00571.x Gil de Zúñiga, H., Molyneux, L., & Zheng, P. (2014). Social media, political expression, and political participation: Panel analysis of lagged and concurrent relationships. Journal of Communication, 64(4), 612-634. doi:10.1111/jcom.12103 Gladwell, M. (2010). Small change. The New Yorker, pp. 42-49. Halpern, D., Valenzuela, S., & Katz, J. E. (2017). We face, I tweet: How different social media influence political participation through collective and internal efficacy. Journal of Computer-Mediated Communication, 22(6), 320-336. doi:10.1111/jcc4.12198 Hibbing, M. V., Ritchie, M., & Anderson, M. R. (2011). Personality and political discussion. Political Behavior, 33(4), 601-624. doi:10.1007/s11109-010-9147-4 Huckfeldt, R., & Sprague, J. D. (1995). Citizens, politics, and social communication: information and influence in an election campaign. Cambridge [England] ; New York: Cambridge University Press. Ito, M., Soep, E., Kligler-Vilenchik, N., Shresthova, S., Gamber-Thompson, L., & Zimmerman, A. (2015). Learning connected civics: Narratives, practices, infrastructures. Curriculum Inquiry, 45(1), 10-29. doi:10.1080/03626784.2014.995063 Jenkins, H. (2006). Convergence culture: Where old and new media collide. New York: New York University Press. Jenkins, H., Itō, M., & boyd, d. (2016). Participatory culture in a networked era: A conversation on youth, learning, commerce, and politics. Cambridge, UK ; Malden, MA: Polity Press. Kenski, K., & Stroud, N. J. (2006). Connections between Internet use and political efficacy, knowledge, and participation. Journal of Broadcasting & Electronic Media, 50(2), 173-192. doi:10.1207/s15506878jobem5002_1 Kim, J., & Kim, E. J. (2008). Theorizing dialogic deliberation: Everyday political talk as communicative action and dialogue. Communication Theory, 18(1), 51-70. doi:10.1111/j.1468-2885.2007.00313.x Kim, J., Wyatt, R. O., & Katz, E. (1999). News, talk, opinion, participation: The part played by conversation in deliberative democracy. Political Communication, 16(4), 361-385. doi:10.1080/105846099198541 Kim, Y., Chen, H.-T., & Gil de Zúñiga, H. (2013). Stumbling upon news on the Internet: Effects of incidental news exposure and relative entertainment use on political engagement. Computers in Human Behavior, 29(6), 2607-2614. doi:http://dx.doi.org/10.1016/j.chb.2013.06.005 Kligler-Vilenchik, N. (2013). Decreasing world suck: Fan communities, mechanisms of translation, and participatory politics. Los Angeles: MacArthur Foundation. 37

Kligler-Vilenchik, N. (2017). Alternative citizenship models: Contextualizing new media and the new “good citizen”. New Media & Society, 19(11), 1887-1903. doi:10.1177/1461444817713742 Kline, R. B. (2005). Principles and practice of structural equation modeling. New York: Guilford Press. Klofstad, C. A. (2007). Talk leads to recruitment. Political Research Quarterly, 60(2), 180-191. doi:10.1177/1065912907301708 Kraut, R., Kiesler, S., Boneva, B., Cummings, J., Helgeson, V., & Crawford, A. (2002). Internet paradox revisited. Journal of Social Issues, 58(1), 49-74. doi:doi:10.1111/1540-4560.00248 Lee, N.-J., Shah, D. V., & McLeod, J. M. (2012). Processes of political socialization: A communication mediation approach to youth civic engagement. Communication Research, 40(5), 669-697. doi:10.1177/0093650212436712 Lilleker, D. G., & Koc-Michalska, K. (2017). What drives political participation? Motivations and mobilization in a digital age. Political Communication, 34(1), 21-43. doi:10.1080/10584609.2016.1225235 Little, T. D. (2013). Longitudinal structural equation modeling. New York: Guilford Press. 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. doi:10.1177/1461444810365313 Morozov, E. (2011). The net delusion: How not to liberate the world. The New Yorker, pp. 42-49. Papacharissi, Z., & Rubin, A. M. (2000). Predictors of Internet use. Journal of Broadcasting & Electronic Media, 44(2), 175-196. doi:10.1207/s15506878jobem4402_2 Park, N., Kee, K. F., & Valenzuela, S. (2009). Being immersed in social networking environment: Facebook groups, uses and gratifications, and social outcomes. CyberPsychology & Behavior, 12(6), 729-733. doi:10.1089/cpb.2009.0003 Penney, J. (2014). Social media and symbolic action: Exploring participation in the Facebook red equal sign profile picture campaign. Journal of Computer-Mediated Communication. doi:10.1111/jcc4.12092 Prior, M. (2007). Post-broadcast democracy: How media choice increases inequality in political involvement and polarizes elections. New York: Cambridge University Press. Prior, M. (2010). You’ve either got it or you don’t? The stability of political interest over the life cycle. The Journal of Politics, 72(03), 747-766.

38

Sardeshmukh, S. R., & Vandenberg, R. J. (2017). Integrating moderation and mediation: A structural equation modeling approach. Organizational Research Methods, 20(4), 721-745. doi:10.1177/1094428115621609 Shah, D. V., Cho, J., Eveland, W. P., & Kwak, N. (2005). Information and expression in a digital age: Modeling Internet effects on civic participation. Communication Research, 32(5), 531-565. doi:10.1177/0093650205279209 Shah, D. V., Cho, J., Nah, S., Gotlieb, M. R., Hwang, H., Lee, N.-J., . . . McLeod, D. M. (2007). Campaign ads, online messaging, and participation: Extending the communication mediation model. Journal of Communication, 57(4), 676-703. doi:10.1111/j.1460-2466.2007.00363.x Shah, D. V., Kwak, N., & Holbert, R. L. (2001). "Connecting" and "disconnecting" with civic life: Patterns of Internet use and the production of social capital. Political Communication, 18(2), 141-162. doi:10.1080/105846001750322952 Skoric, M. M., Zhu, Q., Goh, D., & Pang, N. (2015). Social media and citizen engagement: A meta-analytic review. New Media & Society. doi:10.1177/1461444815616221 Smith, A., & Anderson, M. (2018). Social media use in 2018. Retrieved from http://assets.pewresearch.org/wp-content/uploads/sites/14/2018/03/01105133/PI_2018.03 .01_Social-Media_FINAL.pdf Stoycheff, E., Nisbet, E., C. , & Epstein, D. (2016). Differential effects of capital-enhancing and recreational Internet use on citizens’ demand for democracy. Communication Research. doi:10.1177/0093650216644645 Tetlock, P. E., Skitka, L., & Boettger, R. (1989). Social and cognitive strategies for coping with accountability: Conformity, complexity, and bolstering. Journal of personality and social psychology, 57(4), 632-640. doi:10.1037/0022-3514.57.4.632 Thorson, K. (2014). Facing an uncertain reception: Young citizens and political interaction on Facebook. Information, Communication & Society, 17(2), 203-216. doi:10.1080/1369118X.2013.862563 Vaccari, C., Valeriani, A., Barberá, P., Bonneau, R., Jost, J. T., Nagler, J., & Tucker, J. A. (2015). Political expression and action on social media: Exploring the relationship between lower- and higher-threshold political activities among Twitter users in Italy. Journal of Computer-Mediated Communication, 20(2), 221-239. doi:10.1111/jcc4.12108 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. doi:10.1080/10584609.2017.1334726

39

Valeriani, A., & Vaccari, C. (2015). Accidental exposure to politics on social media as online participation equalizer in Germany, Italy, and the United Kingdom. New Media & Society, 18(9), 1857-1874. doi:10.1177/1461444815616223 Valeriani, A., & Vaccari, C. (2016). Accidental exposure to politics on social media as online participation equalizer in Germany, Italy, and the United Kingdom. New Media & Society, 18(9), 1857-1874. doi:10.1177/1461444815616223 Vraga, E. K., Bode, L., Smithson, A.-B., & Troller-Renfree, S. (2016). Blurred lines: Defining social, news, and political posts on Facebook. Journal of Information Technology & Politics, 13(3), 272-294. doi:10.1080/19331681.2016.1160265 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. doi:10.1111/j.1460-2466.2008.01403.x Wyatt, R. O., Katz, E., & Kim, J. (2000). Bridging the spheres: political and personal conversation in public and private spaces. Journal of Communication, 50(1), 71-92. doi:10.1111/j.1460-2466.2000.tb02834.x Xenos, M., & Moy, P. (2007). Direct and differential effects of the Internet on political and civic engagement. Journal of Communication, 57(4), 704-718. doi:10.1111/j.1460-2466.2007.00364.x Xenos, M., Vromen, A., & Loader, B. D. (2014). The great equalizer? Patterns of social media use and youth political engagement in three advanced democracies. Information, Communication & Society, 17(2), 151-167. doi:10.1080/1369118X.2013.871318 Yu, R. P. (2016). The relationship between passive and active non-political social media use and political expression on Facebook and Twitter. Computers in Human Behavior, 58, 413-420. doi:http://dx.doi.org/10.1016/j.chb.2016.01.019 Zuckerman, E. (2014). New media, new civics? Policy & Internet, 6(2), 151-168. doi:10.1002/1944-2866

40

Table 1. Zero-order correlations 1

Age

2 3 4 5 6 7 8 9 10

Gender (Male) Race (White) Education Income Newspaper TV news News website Political view Political interest

11 12 13 14 15 16

Political efficacy Political talk offline Entertainment expressive use Social expressive use Political expression on FB Political participation

1

2

3

4

5

6

7

8

-.06 .22*** .06 .05 .32*** .36*** -.08* .17*** .19***

-.05 .05 .10** .09* .10** .16*** .02 .19***

-.01 .00 .05 -.05 -.15*** .15 .01

.39*** .18*** .00 .19*** -.08* .18***

.26*** .12** .25*** .04 .17***

.37*** .29*** .00 .26***

.29*** -.01 .28***

-.05 .33***

***

***

***

-.04 -.04 -.31*** -.24*** -.16*** .09

.06 .12** .14*** -.07 .16*** .13***

**

-.10 -.09* -.10** -.02 -.08* -.06

*

.08 .15*** -.04 -.06 -.02 .11**

.07 .17*** .06 .02 .06 .07

.13 .19*** .12** .04 .13*** .18***

.14 .17*** .11** .02 .13** .18***

9

.24 .30*** .34*** .25*** .38*** .29***

10

1

.34*** .50*** .14*** .09* .38*** .35***

.2 .2 .1 .3 .2

-.04 -.12** -.04 -.12** -.06 -.09* -.05

Notes: Baseline N = 727. Cell entries are Pearson’s R correlation coefficients. Gender and race are dichotomous, and Pearson’s point-biserial correlations are used. *p < .05, **p < . 01, ***p < .001.

41

Table 2. Standardized coefficients and significant levels of the measurement model Parameter estimate

Standardized coefficients

Social FB use Like Comment Post Entertainment FB use Like Comment Post Interactive political expression on FB Like

0.77*** 0.88*** 0.91***

Comment Post Political discussion offline Friends Family Political participation Attended meeting, rally, or speech Contributed money to a political campaign or cause Wore/displayed a campaign button, bumper sticker, or yard

0.92*** 0.92***

0.86*** 0.91*** 0.91*** 0.92***

sign Signed a petition for a candidate or cause Worked for a political party or candidate

0.91*** 0.77*** 0.90*** 0.81*** 0.87*** 0.68*** 0.89***

Notes: N = 727, *p < .05, **p < . 01, ***p < .001. The model fit of the measurement model is χ2 (94) = 143.35, p < .01, CFI =.98, TLI = .97, RMSEA = .03.

42

Figure 1. Structural equation model of social and entertainment expressive Facebook use, interactive political expression on Facebook, political talk offline, and political participation.

.01

.31***

Social expressive FB use

Interactive political expression on FB .21**

.22

R2 = .57

***

Political participation

.68***

R2 = .37

.31*** Entertainment expressive FB use

Political talk offline

.05

.18**

R2 = .41

-.15 Notes: Path entries are standardized coefficients. N = 727, *p < .05, **p < .01, ***p < .001. The paths from exogenous variables (age, race, gender, education, income, newspaper use, television news use, web news use, political view, political interest, and political efficacy) to the endogenous variables were estimated but are not displayed. Model goodness fit: χ2 (215) = 319.32, p < .001, CFI = .97, TLI = .96, RMSEA = .03.

43

Table 3. Indirect effects of social and entertainment expressive Facebook use on political participation through interactive political expression on Facebook and political talk offline Standardized coefficient Social expressive FB use  Interactive political expression on FB  Political participation Entertainment expressive FB use  Interactive political expression on FB  Political participation Social expressive FB use  Political talk offline  Political participation Entertainment expressive FB use  Political talk offline  Political participation Notes: N = 727, *p < .05, **p < .01, ***p < .001

44

0.06** 0.06** 0.04* 0.01

Table 4. Structural equation model of interaction between social and entertainment expressive Facebook use, interactive political expression on Facebook, political talk offline, and political participation. Baseline

Interaction

Unstandardized coefficient

SE

0.32*** 0.34***

Predictor Social expressive FB use Entertainment expressive FB use Social use  Political interest Entertainment use  Political interest Outcome: Political participation Predictor Social expressive FB use Entertainment expressive FB use

Outcome: Interactive political expression on FB Predictor Social expressive FB use Entertainment expressive FB use

Unstandardized coefficient

SE

0.06 0.07

0.33*** 0.33*** 0.10* 0.00

0.06 0.07 0.05 0.05

0.22** 0.06

0.07 0.06

0.21** 0.07 -0.06 0.07

0.07 0.07 0.05 0.05

0.01 -0.53

0.29 0.29

0.03 -0.47

0.31 0.30

0.63* 0.70**

0.26 0.23

0.63* 0.71** -0.28 -0.28 -11267.40 22764.76 23292.48

0.26 0.24 0.26 0.30

Social use  Political interest Entertainment use  Political interest Outcome: Political discussion offline

Interactive political expression on FB Political discussion offline Social use  Political interest Entertainment use  Political interest Log likelihood Akaike information criterion (AIC) Bayesian information criterion (BIC)

-11278.70 22777.39 23282.17

Notes: N = 727, *p < .05, **p < .01, ***p < .001; SE = standard error of the unstandardized coefficient.

45

Highlights:  Social and entertainment expressive Facebook use are positively related to interactive political expression on the site.   

Social expressive Facebook use is positively related to offline political talk, but entertainment expressive use is not. Political expression on Facebook and offline political talk are mediators that link social and entertainment expression on Facebook with political participation. The interactions between political interest and each of the expressive uses are largely insignificantly related to political communication and participation.

46