Abstract
The use of social media sites for political expression has added a new layer to the study of political discussion. In this type of user-generated content applications, interpersonal and mass-mediated types of communication converge in one medium. Therefore, in this study, we bring together insights from interpersonal communication theories and mass communication models that predict online expression, adding explanatory power to this form of political participation. Specifically, we contribute to the understanding of social media political expression by examining the influence of communication competence and expected outcomes on expressive behaviors. Results of a representative survey of the adult urban population of Colombian Facebook (n = 598) and Twitter (n = 185) users, suggest that social media communication competence and expected status, novelty, and social outcomes influence political expression on social media. Furthermore, users’ social media political expression expected outcomes through Facebook and Twitter differ. This study illustrates the important role that interpersonal communication factors have for social media political discussion, and the difference between users’ expectations when they express their political views through different social media platforms.
Introduction
New forms of political engagement have emerged with the dissemination of social media sites and applications. These emerging modes of participation might be as exceptional as the traditional offline modes of participation (Bekafigo & McBride, 2013; Rainie, Smith, Schlozman, Brady, & Verba, 2012), but they are an important resort for politically active social media users to coordinate political action (Earl, Hurwitz, Mesinas, Tolan, & Arlotti, 2013; Segerberg & Bennett, 2011), mobilize (Yamamoto, Kushin, & Dalisay, 2015), and to express political views (Himelboim et al., 2016; Larsson & Moe, 2012). Political expression on social media has been found to positively influence other political behaviors such as traditional types of political participation (de Zúñiga, Bachmann, Hsu, & Brundidge, 2013), including political mobilization (Rojas & Puig-i-Abril, 2009) and electoral campaigning (Dimitrova, Shehata, Strömbäck, & Nord, 2014).
Furthermore, research has found that online settings might be more politically diverse than face-to-face settings (Baek, Wojcieszak, & Carpini, 2012; Bakshy, Messing, & Adamic, 2015) and may offer a political outlet for marginalized groups (Bekafigo & McBride, 2013). This evidence emphasizes both the importance of social media political expression for the health of democracies and the relevance of studying the antecedents of political expression in social media.
The use of social media platforms for political expression has added a new layer to the study of political discussion. Some researchers have characterized social media sites, such as Facebook and Twitter, more akin to large group communication platforms. These studies have described the different communication processes and dynamics that take place on these social media sites as similar to larger online communities and groups (i.e., Karnik, Oakley, Venkatanathan, Spiliotopoulos, & Nisi, 2013; Park, Kee, & Valenzuela, 2009). Although this characterization can provide an understanding of the communicative processes that take place on social media, we find an important difference in that, within this type of user-generated content applications, interpersonal and mass-mediated types of communication converge in one medium, enabling mass/personal forms of communication (Carr & Hayes, 2015) in a scale that was not possible before. Such conceptualization of social media fits our theoretical explanation of social media political expression. This study brings together interpersonal and mass communication models and concepts to predict social media political expression. Such an approach adds explanatory power to this emerging type of political participation. In this study, we specifically examine the influence of expected outcomes of social media political expression and communication competence on social media political expression.
We follow Carr and Hayes’ (2015) conceptualization of social media as “Internet-based channels that allow users to opportunistically interact and selectively self-present, either in real-time or asynchronously, with both broad and narrow audiences who derive value from user-generated content and the perception of interaction with others” (p. 50). In our analyses, we distinguish political expression among different social media platforms, assuming that an individuals’ derivation of value from content generation and consumption, as well as their perceptions of interactions with others, differ between diverse social media applications. We argue that just as social media have a set of properties, individuals who use them have their own characteristics and needs (Leonardi, 2013). The utility of those social media characteristics for individuals (i.e., their affordances) is placed in relation with their needs and expectations (Hutchby, 2001). Therefore, some social media platforms will meet individuals’ expectations better than others.
We also highlight the potential role of communication competence and motives of use as critical antecedents of political expression on social media. Communication competence is defined as the perceived effectiveness and appropriateness people believe they have as communicators, and it is related with a need to fulfill interpersonal objectives (Canary & Spitzberg, 1987). In the context of digital media, perceptions of competence may be different as the communication practices of individuals change (Spitzberg, 2006). We conceptualize perceived social media communication competence (SMCC) as the degree to which people perceive that they effectively and appropriately achieve their interpersonal communication objectives when they interact with others through social media.
Research focused on motives for media use have a long tradition. A more recent approach to the study of media attendance and use motives is that advanced by LaRose and colleagues (LaRose & Eastin, 2004; LaRose, Mastro, & Eastin, 2001), based on Social Cognitive Theory (SCT; Bandura, 1986, 1991, 1997). This study uses the same approach to media use motives and conceptualizes them in the logic of expected outcomes. According to SCT, individuals act partly based on the expected outcomes they believe they will experience once they perform a particular behavior. In other words, expected outcomes act as incentives, as individuals foresee desired futures from particular behaviors. Moreover, a SCT approach of expected outcomes ensures the inclusion of types of incentives that other motivation-based approaches of media use do not usually include (e.g., uses and gratifications research does not usually include status incentives; LaRose et al., 2001).
We selected Colombia as the context to study the influence of these factors on social media political expression. Connectivity levels, mobile and smartphones use, as well as the popularity of social media have continued to rise steadily in recent years in the country (MinTIC, 2015). To understand the magnitude of this phenomena, consider that in 2012, Colombia ranked 12 among the Top 20 countries in the world with the most Twitter accounts (Bennett, 2012), and 43% of its population on Facebook, making it 17th in the world in terms of total users (Allin1Social, 2015).
In sum, this study examines how SMCC and expected novelty, status, and social outcomes influence political expression through Facebook and Twitter, two of the most popular social media applications in Colombia.
Literature Review
SMCC
Political expression influences other forms of political behaviors (Kwak, Williams, Wang, & Lee, 2005) and cognitions (Eveland, Hayes, Shah, & Kwak, 2005). Findings from previous research suggest that since childhood, political talk shapes subsequent political attitudes and behaviors (McIntosh, Hart, & Youniss, 2007). The importance of political expression is backed up by a plethora of studies and theories developed to understand the different factors that might influence individuals’ willingness to express their political positions openly.
Previous studies have shown that an individual’s general communication predispositions, and not only the content of the communication under discussion, influence political expression. In addition to fear of isolation (Noelle-Neumann, 1974), research has also uncovered how opinion climate, opinion congruency, and individual traits, such as communication apprehension and self-censorship, influence an individual’s level of outspokenness (Donsbach, Salmon, & Tsfati, 2014; Ho & McLeod, 2008; Matthes et al., 2012; Sheehan, 2015; Willnat, Lee, & Detenber, 2002). Findings by Willnat et al. (2002) and by Ho and McLeod (2008) suggest that some people tend to avoid situations in which they have to speak publicly and this phenomenon might actually be stronger on social media (Hampton et al., 2014).
We conceptualize political expression as communications that express a specific opinion on current events or political processes or that disseminate information relevant to the interpretation of those events or processes. Political expression is a political behavior that has interpersonal communication factors and dynamics also associated with it. In this sense, individuals’ perceptions of how competent they are in their interpersonal communication interactions are also relevant in the study of opinion expression and opinion expression through social media.
The concept of communication competence has been uncovered as a key factor for individuals’ interpersonal interactions. The notion of communication competence allows for an understanding of the way in which people direct their communication and achieve their relational goals. Communication competence is defined as an individual’s disposition for effective and appropriate interaction given a particular interpersonal relational context (Spitzberg, 1983). Although defined as encompassing skills that go beyond interpersonal communication, and extend to media use and news consumption, communication competence has been previously found to be positively related with activism (Shah, McLeod, & Lee, 2009). The notion of competence is located in the theoretical framework of self-determination theory and refers to the need to master challenging tasks (Ryan & Deci, 2000). Although competence and self-efficacy are commonly used interchangeably, they differ from each other empirically and conceptually. Both concepts are based on different theoretical assumptions. Self-efficacy does not come from the need to master but is defined as belief in one’s capabilities (Rodgers, Markland, Selzler, Murray, & Wilson, 2014).
The concept of communication competence has been previously studied in the context of interpersonal mediated communication. Spitzberg (2006) proposed the concept of Computer-Mediated Communication competence (CMCC), while Bakke (2010) proposed the concept of mobile communication competence (MCC). These studies have found that CMCC and MCC were associated with relational and communication outcomes. Moreover, although interpersonal communication competence has been found to directly affect mediated communication competence (Hwang, 2011), evidence also suggests that individuals tend to choose a communication channel, including face-to-face, depending on their perceived communication competence for each channel (Keaten & Kelly, 2008).
In this study, we propose the concept of SMCC, defined as individuals’ perceptions that they effectively and appropriately use social media to conduct interpersonal interactions. Social media are characterized by their selective self-presentation affordances and by the derivation of value through content generation. Users perform their identities through the construction of the profiles and the contribution of content. Moreover, through the selective contribution of specific types of content, they express themselves and construct an identity of themselves for their audience (Marwick & boyd, 2011). We argue that users, to take advantage of these affordances and to actually achieve these interpersonal goals, need a distinct set of competencies specific to the characteristics of social media interactions that are different from other mediated communication competence concepts proposed previously.
We hypothesize that in the same way previous mediated communication competence concepts have been found to influence communication behaviors (Bakke, 2010; Spitzberg, 2006), SMCC should also be related with political expression through social media applications, specifically, Twitter and Facebook. Thus, we pose the following hypotheses:
H1a. Communication competence will be positively associated with Facebook political expression.
H1b. Communication competence will be positively associated with Twitter political expression.
Expected Outcomes of Political Expression on Social Media
Following Carr and Hayes (2015), a constitutive element of social media is the value users derive from content generation, consumption, and the perception of interaction with others. Social media are also characterized by the possibility of combining mass and interpersonal communication channels, affording mass-personal forms of communication. Despite social media sharing these characteristics, the value these affordances bring is set in relation to users’ expectations (Hutchby, 2001). Therefore, each social media application meets individuals’ expectations in a different manner, depending on its specific features.
Communication researchers have explored how individuals select what media to use from a motivational theory perspective, especially in terms of uses and gratifications (Palmgreen & Rayburn, 1982; Palmgreen, Wenner, & Rosengren, 1985). Here, we adopt a SCT perspective of the UG approach (LaRose, 2009; LaRose & Eastin, 2004). From a SCT perspective, the gratifications sought-gratifications obtained dynamic (Palmgreen & Rayburn, 1982) is interpreted in terms of the successful enactive experience mechanism of SCT (see LaRose & Eastin, 2004; Velasquez & LaRose, 2015). LaRose et al. (2001) explain that previous media experiences determine future media exposure because those experiences influence the perceptions of the likely outcomes of that future media use. In this sense, individuals exercise forethought and imagine the predicted consequences of their behaviors. Outcome expectations are translated into cognitively based motivations through forethought. The notion of outcome expectations is also found in expectancy-value theory (Ajzen & Fishbein, 2008; Atkinson & Reitman, 1956), but in the form of behavior’s salient anticipated outcomes. The strength of the motivation is determined by the belief that certain actions will bring about specific outcomes and by how much those outcomes are valued. These motivations may be categorized under sensory, social, monetary, activity, and status incentives (Bandura, 1986, p. 232). Therefore, social media users may participate on social media sites in different ways depending on the outcome they expect that interacting, providing content, or consuming others’ content will bring to them.
In this study, we focus on the influence of social, status, and novelty incentives on political expression through social media. Social motivations are defined as the value people give to the reactions others have to their different behaviors (Bandura, 1986). Individuals act based on social motivations when their behavioral decisions are partly determined by considerations of the expressions of interest and approval of others or of acceptance into a group.
Status incentives, although similar to social incentives, differ in that the driver of the behavior is not necessarily acceptance per se, but rather that a certain level of competence or knowledge will secure a rank among others that comes with various benefits (Bandura, 1986). Individuals act based on status incentives when they believe the outcome of their skilled behaviors will be highly valued in a group and will enable them to reach a certain rank in that social organization.
Novelty incentives, on the other hand, emerge from the sensory feedback that different activities may provide. The notion of novelty incentives is based on the assumption that change and novel experiences increase the quality and effectiveness of sensory feedback. When novelty emerges, it creates a need to resolve the conflict between already existing knowledge or expectations and current experiences. Potential incongruity motivates exploration to resolve the conflict between previous experiences or beliefs and current circumstances. The sensory effect this is expected to produce motivates individuals to seek novelty through new experiences, new ideas, or new information (Bandura, 1986).
In this study, we argue that users’ expectations of what they will get from expressing their political views through Facebook and through Twitter will differ. Although both are social media applications, they are different in important ways. Facebook as a social network site offers users the possibility of building profiles within a system, of connecting with other users in this system, and of viewing and going over the connections and the information about other users with whom a user has connected (boyd & Ellison, 2007). Twitter, in contrast, is a microblogging application that allows users to share, comment, or express themselves in no more than 140 characters and to follow others based on their own interests. Twitter is less a site for making or maintaining social connections, and more for building less formal relationships and weak connections with others (Chen, 2011) and sharing political information; therefore, it is more suitable for expressing and learning what others have to say about different issues of interest to the follower. Although there is the option for sending private messages, Twitter follows more of a broadcast model as people can follow others independent of the relationship they have.
In contrast, on Facebook connections are necessarily reciprocal. One cannot be friends with someone who does not consider him/her their friend. Facebook facilitates a network structure of strong and weak ties, while the asymmetric design of Twitter makes it more suited for weak connections (Valenzuela, Arriagada, & Scherman, 2014). These might allow for differences in the value users derive from the expression of their political views. Therefore, we examine the different outcomes individuals foresee from expressing their political views in each of these social media applications.
Social media sites constantly collapse multiple social contexts, potentially resulting in self-censorship behaviors, as individuals selectively self-present through these platforms. Additionally, individuals usually enact their offline social networks on Facebook and perform their identities through this social networking site, making this performance part of their identities (Cheung, Chiu, & Lee, 2011; Gangadharbatla, 2008). Therefore, individuals will tend to refrain from expressing their political views on Facebook when they believe this type of behavior might bring to them negative social consequences. Therefore, we pose the following hypothesis:
H2. Social expected outcomes are negatively related to political expression on Facebook.
The mass personal characteristics of social media sites (Carr & Hayes, 2015) also encourage micro-celebrity practices (Marwick & boyd, 2011). These practices are rooted on the expectation of acquiring a particular status among Facebook friends and Twitter followers. Through their social media political engagement, individuals present their political selves and garner the privilege of their audience’s attention (Tufekci, 2013), with the objective of influencing the political views of their friends and followers. Status incentives refer to the rank one can acquire within a particular social group. Although reference groups might vary between Facebook and Twitter, the expectation of receiving the recognition, respect, and admiration of those in that social group should act as a predictor of political expression on each of these applications. In this sense, we expect that Facebook and Twitter users would post political views based on the status their posts might bring to them.
H3. Status expected outcomes are positively related to political expression on Facebook.
H4. Status expected outcomes are positively related to political expression on Twitter.
Finally, as Twitter is a social media application where users might have access to a wider and more diverse set of opinions, we expect that novelty expected outcomes would be related with political expression through this social media site.
H5. Novelty expected outcomes are positively related to political expression on Twitter.
Colombian Context
Colombia’s political system has been characterized as that of a flawed democracy. In 2015, the nation ranked 62 in The Economist’s (2016) Democracy Index and received high marks for its electoral process and civil liberties, but lower marks for political participation overall. In terms of its recent political history, over the course of the 21st century a conservative–liberal party divide evolved into a multiparty system, with certain parties representing conservative positions, others representing the political center, and others representing the political left. The press in Colombia tends to be closely tied to big business interests and has been described as a market-based press with a “weak legacy of media pluralism” (Waisbord, 2008, p. 3). In Colombia, the Internet has allowed more ideologically oriented political expression, with a vibrant sphere of political communication emerging on Twitter, where political leaders routinely express their views to millions of followers.
In 2002, Alvaro Uribe, a right-wing politician who promised that guerrillas would be defeated militarily, was elected president. Uribe was subsequently re-elected for a second 4-year term in 2006. In 2010, Uribe’s former defense minister and Partido de la U candidate, Juan Manuel Santos, was elected president. Once elected, Santos distanced himself from Uribe, moving closer to the center and initiating a peace process with FARC. Uribe and some other members of his party left and created a new coalition. Under the new banner of Centro Democratico, they tried to stop Santo’s reelection bid in 2014.
This recent election was one of the most polarizing in Colombian history, with campaigns engaging in negative campaigning and, allegedly, illegal surveillance, defamation, and electoral fraud—claims that are being investigated by the Colombian judiciary. Ultimately, Santos, with a broader coalition that included center and leftist parties, was able to prevail and continue negotiations with FARC that have now culminated in a historic peace agreement that puts an end to a violent political conflict of more than 50 years.
Although this initiative has generated polarization in Colombian public opinion, it seems like Colombian society has reached a moment for political inclusion, dialogue, and discussion. Social media has become an additional channel for the exchange of political views to happen. Freedom House (2015) considers the Internet in Colombia to be “partially free” with no social media or political content blocked in the country, in which “prosecution for online expression are rare,” but were illegal surveillance remains a concern.
This context—characterized by heightened political tension, sharp division in terms of electoral politics and policy positions, and increased reliance on social media to communicate political ideas—seems an important scenario to explore the questions of motivations and political expression.
Methods
Sample and Data Collection
This study relies on national survey data collected from 28 June to 29 July 2014 in 10 cities in Colombia by the Universities of Wisconsin and Externado de Colombia as part of their biennial study of communication and political attitudes in Colombia. The sample was designed to represent Colombia’s adult urban population—76% of Colombia’s 47.6 million inhabitants live in urban areas (DANE, 2014).
Survey respondents were selected using a multi-stage stratified random sample procedure that selected households randomly based on city size and census data. Once the number of households was allocated for a given city, a number of city blocks were selected randomly within socioeconomic strata as established by the national census. Then, individual households were randomly selected within each block. Finally, the study used the “adult in the household who most recently celebrated a birthday” technique to identify an individual respondent at random. Up to three visits to each household were made (if needed) to increase participation in the survey. A local professional polling firm, Deproyectos Limitada, collected the data and 1,102 face-to-face completed responses were obtained for a response rate of 55.5%. 1 All surveys were conducted in Spanish using back-translation survey techniques for questions originally developed in English.
Sample Characteristics
All participants were asked if they had either a Twitter or a Facebook account (in Colombia, there are no local social networking sites; Twitter and Facebook in the Spanish language are the dominant social media). From the 1,102 survey respondents, 598 (54.3%) stated that they had a Facebook account, while only 185 (16.8%) reported that they had a Twitter account. Of those that reported that they had at least one social media application account, 57.6% identified as females, and 42.4% identified as males. The average age of social media users was 33.79 years (standard deviation [SD] = 12.31). Most participants reported that they had completed at least some years of college education (SD = 1.61). The average monthly income was within a range equivalent to US$500–US$1,000 (SD = 1.46).
Analysis
All subsequent analyses were performed only for those respondents who reported having a Facebook or Twitter account. Mean values were imputed for missing data. In no case did missing data exceed 5% of the sample. An exploratory factor analyses was performed to determine if the items measuring each of the three expected outcomes reflected a common pattern. The analyses were performed with three fixed factors, as we expected that the items would load into three factors. Items loading with at least .6 on the primary factor and less than .4 on all other items were retained (Hair, Tatham, Anderson, & Black, 1998).
The principle components factor analysis with Varimax rotation showed that the items that loaded on Factor 1 correspond to novelty outcomes, items that loaded on Factor 2 correspond to status expected outcomes, while items that loaded on Factor 3 correspond to social expected outcomes (see Table 1).
Rotated Factor Analysis Solution of Social, Status, and Novelty Expected Outcomes.
Varimax rotation with Eigenvalues > 1 specified, three factors extracted explaining 89.24% of the variance. Bold-face values represent the factor in which that item loaded.
Measures
The value for the dependent, independent, and control variables were calculated by averaging the items that comprised any given scale.
Dependent Variables
Social media political expression through Facebook (M = 2.31, SD = 1.43, Spearman-Brown = .79) was measured as an additive index of two items. The preface asked participants how frequently they performed the following activities on Facebook (0 = never, 5 = frequently): (1) express your opinion about current topics in Facebook and (2) share news with your contacts in Facebook.
Social media political expression through Twitter (M = 2.52, SD = 1.52, Spearman-Brown = .83) was measured as an additive index of two items. The preface of the questions asked participants about how frequently (0 = never, 5 = frequently) they performed the following activities: (1) express your opinion about current topics in Twitter and (2) share news with your followers in Twitter.
Independent Variables
Social, status, and novelty outcomes comprised two items each. Items were adapted from previous research (Lampe, Wash, Velasquez, & Ozkaya, 2010; Wohn, Velasquez, Bjornrud, & Lampe, 2012). Respondents were asked to express how much they agreed with the statements expressed in each item (0 = not likely at all, 6 = very likely). The preface was “When using social media sites to discuss about politics, how likely are you to . . .”
Social expected outcome (M = 1.76, SD = 1.53, Spearman-Brown = .87) was measured with the following: (1) feel you belong to a group and (2) feel closer to other people.
Status outcome (M = 1.52, SD = 1.47, Spearman-Brown = .9) was determined by asking respondents about the following: (1) feel important and (2) impress your contacts.
Novelty outcome (M = 2.60, SD = 1.59, Spearman-Brown = .86) was measured with the following: (1) get new information and (2) learn new things.
Scales employed in previous studies (Bakke, 2010; Spitzberg, 2006) were adapted to measure the SMCC variable (M = 2.86, SD = 1.15, Cronbach’s = .88). The index comprised six items, and the preface to the questions asked participants to indicate their level of agreement with a set of statements. The variable was calculated by averaging the answer for each of the six items (0 = strongly disagree, 5 = strongly agree). Statements included the following: (1) I generally get what I want out of social media interactions, (2) I consistently achieve my goals in social media interactions, (3) I get my ideas across clearly in conversations with others in social media, (4) I feel understood when I interact with others in social media, (5) When I publish content in a social network site, I pay attention to both what I say and how I say it, and (6) In social media sites, I am careful to make my comments and behaviors appropriate to the situation.
Control Variables
Besides controlling for the sociodemographic variables described above, other control variables included were internal political efficacy (M = 2.44, SD = 1.78, Spearman-Brown = .85), as people who feel they can make more of a difference in the political system are prone to expressing their views. Internal political efficacy was measured by asking respondents about their level of agreement (0 = totally disagree, 5 = totally agree), with the following: (1) “Sometimes government and politics are so complex that someone like me can’t really understand what is going on” (reversed) and (2) “People like me can influence what local government does.” Another control variable included was political interest (M = 2.19, SD = 1.55, Spearman-Brown = .93), measured with two items that asked participants their level of interest (0 = not at all, 5 = a lot) in (1) local politics and (2) national politics.
General social media use was also controlled for. Participants were asked to report how frequently (0 = never, 5 = frequently) they (1) sent pictures or videos to other contacts and (2) stay in touch with family and friends, both through Facebook (M = 3.09, SD = 1.23, Spearman-Brown = .62) and Twitter (M = 2.14, SD = 1.38, Spearman-Brown = .71).
Results
Prior to starting with the analysis of our hypotheses, tests were performed to identify possible influential outliers, but none were identified. This study was set out to examine the influence of expected social, status, and novelty outcomes and SMCC on social media political expression. Table 2 shows the correlations among the variables. All the assumptions of multiple regression analysis were examined. No indication of multicollinearity among the independent variables was present, according to the variance inflation factors (VIF) calculated. All of them were below 2.5. To examine the hypothesized relationships, two hierarchical regression models were calculated, one examining these relationships for Facebook and the other for Twitter political expression. For both regressions, the first block included sociodemographic variables, internal efficacy, political interest, and general Facebook and Twitter use. In the second block, we introduced our SMCC and in a third block, social, status, and novelty expected outcomes.
Pearson Product Moment Correlation Matrix, n = 598.
Correlation is significant at the .05 level.
Correlation is significant at the .01 level.
Facebook Political Expression
In the initial model for Facebook political expression, R2 = .042, F(7, 585) = 60.361, p < .001, the level of education (β = .108, p < .05), general Facebook use (β = .544, p < .001), internal political efficacy (β = .106, p < .01), and political interest (β = .226, p < .001), were significantly related with the dependent variable (see Table 2). Controlling for other factors, the higher the level of education, the higher the political interest and the more individuals use Facebook in general, the more they tended to express their political views through this social media site.
Hypothesis 1A stated that communication competence would be positively associated with Facebook political expression. This hypothesis was supported. When SMCC was added, the R2 had a significant increase, R2 = .020, F(1, 584) = 20.509. Results of the second model, R2 = .044, F(8, 584) = 57.141, suggested that the significant predictors of Facebook political expression were education (β = .101, p < .01), general Facebook use (β = .494, p < .001), internal efficacy (β = .083, p < .05), political interest (β = .197, p < .001), and SMCC (β = .159, p < .001). Results suggest that the more competent individuals feel to communicate through social media, the more they will tend to express their political views through Facebook.
The second hypothesis stated that social expected outcomes would be negatively related to political expression, while the third hypothesis stated that status would be positively related to political expression through Facebook. However, only the hypothesis regarding the relationship between status and social media political expression was supported. In the third model, social, status, and novelty outcomes were included, increasing significantly the R2, R2 = .014, F(3, 581) = 4.995. In this model, R2 = .045, F(11, 581) = 43.772, education (β = .092, p < .01), general Facebook use (β = .478, p < .001), internal political efficacy (β = .075, p < .05), political interest (β = .169, p < .001), SMCC (β = .095, p < .05), status expected outcomes (β = .099, p < .05), and novelty expected outcomes (β = .099, p < .05) were significantly related with Facebook political expression.
These results indicate that the more individuals feel that they will get the respect and admiration of their Facebook contacts because of their political expression and the more they expect to acquire new information and learn new things when participants use social media sites to discuss about politics, the more they will tend to express their political views through this social media site.
Twitter Political Expression
In the initial model for Twitter political expression, R2 = .040, F(7, 175) = 16.692, p < .001, age (β = −.187, p < .05), income (β = .152, p < .05), and general Twitter use (β = .513, p < .001) were significantly related with the dependent variable (see Table 3). Controlling for other factors, the higher the younger the individual, the higher their income and the more individuals use Twitter in general, the more they tended to express their political views through this social media site.
Hierarchical Regression Analyses Predicting Facebook Political Expression, n = 598.
p < .05; **p < .01; and *** p < .001.
Hypothesis 1B stated that communication competence would have a positive association with Twitter political expression. Results supported the hypothesis. When SMCC was added, the R2 had a significant increase, R2 = .078, F(1, 174) = 25.937. Results of the second model, R2 = .047, F(8, 174) = 19.929, suggested that the significant predictors of Twitter political expression were age (β = −.156, p < .01), general Twitter use (β = .493, p < .001), internal efficacy (β = .083, p < .05), political interest (β = .197, p < .001), and SMCC (β = .318, p < .001). As in Facebook political expression, the higher individuals’ communication competence perceptions through social media, the more they will tend to express their political views through Twitter. Therefore, results supported hypothesis 1B (Table 4).
Hierarchical Regression Analyses Predicting Twitter Political Expression, n = 185.
p < .05; **p < .01; and ***p < .001.
Hypothesis 4 and 5 examined the positive relationship between status and novelty expected outcomes and political expression on Twitter. Only hypothesis 5 was supported. In the third model, social, status, and novelty outcomes were included, increasing significantly the variance explained, R2 = .026, F(3, 171) = 2.99. In this model, R2 = .050, F(11, 171) = 15.811; age (β = −.148, p < .01); general Twitter use (β = .521, p < .001); SMCC (β = .281, p < .001); and novelty expected outcomes (β = .241, p < .01) were significant predictors.
According to these results, the younger an individual is, the more he/she uses Twitter, the higher his/her SMCC and the more he/she perceives he/she will learn new information from his/her political interactions through Twitter, the more he/she will tend to engage in political expressive behaviors through this social media site.
Discussion
This study sets out to examine the relationship between interpersonal and mass communication factors and political expression on social media. Based on the different nature of social media platforms, we explored the different role that social media expected outcomes and SMCC played in explaining political expression through Facebook and Twitter.
As hypothesized, SMCC was related with political expression in both platforms. This finding suggests that interpersonal mediated communication factors are related with mediated political expression. This highlights the interpersonal communication nature of social media applications and the influential role it plays in political expression. The way in which individuals perceive the effectiveness and appropriateness of their communication when they interact with others on social media influences the degree in which they express their political views on both Facebook and Twitter.
The communication dynamics specific to social media, such as the difficulty in establishing an audience and the potential collapse of the social context in which the communication process takes place (Marwick & boyd, 2011) probably require a set of communication competencies specific to social media. Those competencies include the knowledge of using different strategies for recognizing the potential audience, adjusting the content posted, using or modifying privacy settings, or using direct messaging features (e.g., the use of @ on Twitter) to make sure that communication effectively and appropriately achieves users’ interpersonal objectives when they interact with others on social media. These findings point to the possibility that users employ a distinct set of competencies that are specific to the communication affordances of social media, and that these competencies, as they are related with all interpersonal communication objectives on social media, also influence the likelihood that individuals will express their political views on this type of communication technology.
While SMCC refers to a self-assessment related to how effective people perceive they are when they interact with others, the self-efficacy concept refers to individuals’ perceived capability in a particular realm of action. Therefore, higher levels of SMCC might result from increased levels of communication efficacy through social media. An integration of these two concepts will contribute to strengthen a SCT perspective of social media political expression and contribute to an understanding of the cognitive and communicative processes that influence social media political expression. Studies might also benefit from an exploration of the difference between perceived and actual competence, and the different outcomes of both in terms of social media political expression.
The expectation of learning new information, in other words, novelty expected outcomes, was associated with Twitter political expression. This makes sense, as one of the most common uses of Twitter is for getting news (Pew Research Center, 2015). Interestingly, this relationship was also significant on Facebook. This means that users of both social media sites expect to see and learn about new political information when they engage in political expression through these sites. Findings in previous studies illustrate how individuals follow others with similar political views and are exposed selectively to content posted on Twitter (Himelboim, Mccreery, & Smith, 2013; Himelboim, Smith, & Shneiderman, 2013). However, other studies (Bakshy et al., 2015) suggest that on social media sites such as Facebook, individuals are exposed to ideologically discordant content through friends who have different political affiliations. This increased heterogeneity of network ties (Ugander, Karrer, Backstrom, & Marlow, 2011) is facilitated by the social rather than political origin of these networks, which include more weak ties (Bakshy, Rosenn, Marlow, & Adamic, 2012), thus reducing the overall density of the network. While it is possible that under conditions of political polarization, Facebook users could cull their networks based on political difference, so far the emerging evidence suggests that social networks such as Facebook, in which weak ties prevail, tend to expose people to more, not less diverse opinions.
Moreover, during electoral cycles, content that is discrepant with an individual’s opinion is not avoided (Knobloch-Westerwick & Kleinman, 2011). Since data for this study were collected short after a presidential election in Colombia took place, it is not clear if individuals were motivated by new content that would fit their current political views, or new content that would be counter-attitudinal but would provide useful information for decision-making (Knobloch-Westerwick & Kleinman, 2011). It is even possible that both options were the case. For individuals who had already made their mind about whom to vote for, novelty motivations were related with an interest in new information to find more arguments for their decision, and that for those who have not decided yet, novelty motivations were associated with a motivation to explore to resolve potential incongruences between new information and previous beliefs. In this sense, novelty expected outcomes might be explained by different mechanisms depending on the stage of the voting decision-making process. This issue would require further exploration in future studies.
Another possible explanation for these findings could be related with the question of whether the acquisition of new information is generally consistent, or people turn to a particular network to get a specific type of information (Bandura, 1986, p. 153). Our findings regarding the relationship between Facebook political expression and novelty expected outcomes might indicate that individuals consider this type of network structure also relevant for new political information and expression. It might be that new ideas and information are diffused not so much depending on the nature of the tie in the social network but rather on the nature of the transaction that takes place in that social relationship (Bandura, 1986, p. 152). Therefore, as SCT suggests, research should also take into account the psychosocial factors that determine the diffusion of novelties despite the structural characteristics of social networks. Furthermore, our conceptualization of Facebook implies interpersonal dynamics among strong ties, while the Facebook pages feature of this social media platform might resemble weak ties types of relationships. Facebook pages are also a possible venue for political expression. Other studies might benefit by including this form of social media participation.
Status outcomes refer to the expectation individuals have that a certain behavior will be highly valued within their social group, and that this recognition will allow them to acquire certain rank and privileges in said group. The significant relationship between status expected outcomes and political expression through Facebook can be explained in terms of the way in which users engage with this platform and its main characteristics. Facebook friends usually resemble individuals’ weak and strong tie connections in the offline context. Furthermore, the algorithm that defines whose content people are exposed to on Facebook is in part determined by their own usage patterns. This is somewhat constant, allowing users to imagine a particular and specific audience for their posts (Litt, 2012). Although most probably different from the real audience, users perceive that this clearly defined and bounded audience will confer them with a higher status as the audience consumes the political content posted by users.
Unexpectedly, the hypothesized relationship between status expected outcomes and political expression on Twitter was not supported. It could be argued that Twitter users see the beneficial outcomes more in terms of weak connections (Chen, 2011), and not as a platform that enables the constitution of a group of friends with the power of granting a status to users. However, while Twitter connections do not necessarily represent a fraction of the offline social network of individuals, and there is no social expectation that connections will be reciprocal, Twitter users also post content with a specific audience in mind (Marwick & boyd, 2011). The attention they receive from their followers is expressed through retweets, mentions, and favorited tweets. This attention is the expression of a status akin to that of celebrities (Greenwood, 2013), whose status is acquired through the attention they get from people that do not represent strong relationships.
In this sense, the lack of evidence supporting a positive relationship between status expected outcomes and Twitter political expression can be explained in terms of what Tufekci (2013) suggests are the characteristics of activist micro-celebrities. Micro-celebrity activists get the attention they seek both through social media and through being featured on mass media, and their success is more notorious in authoritarian regimes. The presence of activist micro-celebrities is rare in Colombian traditional mass media. Moreover, mass media journalists and traditional politicians have mostly dominated Twitter political expression in Colombia, which makes activist micro-celebrities rare cases. It might be that this sample was not big enough to include these Twitter users and capture such a relationship or that in the Colombian context, with a limited political culture of participation, status orientations are not as closely linked to participatory acts.
Also, the status of Facebook users emerges from a network structure that differs from the weak connections that Twitter facilitates (Chen, 2011). It might be that our operationalization of status reflects more the definition of the status that can be obtained by the type of connections that Facebook facilitates. New research should explore the meaning that activist micro-celebrities give to the status they get from their influence on the public opinion through Twitter. It is also plausible that interpersonal impression management, for example, making followers think I’m well informed, versus political objectives, such as influencing the direction of people’s vote, are also at play during a presidential election. These differing motivations to engage in expression also need to be considered by the future research in this area.
The hypothesized negative relationship between social expected outcomes and Facebook political expression was not supported either. Although the coefficient of the regression was negative, it was not statistically significant. Before advancing any conclusions, further studies should strengthen the measurement of this variable, as only two items comprised the scale. Also, it might be the case that the relationship between negative outcomes of expressing political views and Facebook political expression might be moderated by a general predisposition of using Facebook for social reasons.
The future studies should also take into account activity incentives, defined in terms of the enjoyment and pleasure that some types of activities provide (Bandura, 1986, p. 236). These also account for another set of expected outcomes that motivate behavior; and, they might influence social media political expression. However, political expression on social media might not be determined only by expected outcomes. According to SCT, self-efficacy perceptions regarding the behavior in question act as predictors (Bandura, 1997, p. 49). The self-efficacy construct is at the core of SCT as it not only influences behavior but also shapes what outcomes are expected from the behavior (Bandura, 1986, p. 231). Furthermore, in the context of political uses of social media, other studies (Velasquez & LaRose, 2015) have found the important role of social media political efficacy perceptions and of enactive experience.
Future studies should also account for possible moderator and/or mediator variables in the relationship between political expression, expected outcomes, and social media political expression. For instance, individuals’ frequency of political use of social media, social network diversity, exposure to diverse opinions, and even issue salience of the topic under discussion could play a moderating role. Furthermore, online interpersonal factors, such as selective self-presentation, might affect not only willingness to express political views but also might moderate the role played by social expected outcomes. In addition, panel designs that are more suited to test for causality could control for general social media use, with more robust reliable measures than the ones we were able to obtain, while being able to examine how variations in expected outcomes affect political expression. Another potential limitation of our study is the use of never-to-frequently scales to measure expressive activities, which might mean different things to different respondents, despite their common use. A final limitation of our study is that our expression questions refer to sharing news and opinions about current events and not specifically about political news, or political opinions. The future research should inquire more specifically about political news and political opinions, since some respondents might be thinking about events that are less political in nature, which could be reducing the magnitude of the impact of our independent variables on our criterion variables.
Despite what the future research should establish, our findings are of import as we illustrate the role played by an interpersonal communication factor in social media political expression, and how individuals tend to express on Facebook and Twitter differently, depending on the outcomes they expect from their political expression.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was possible by a Grant from The Wisconsin Alumni Research Foundation (WARF).