When thinking about an average Trump supporter, the Make America Great Again red baseball cap quickly comes to mind. More than a simple piece of political apparel, the hat came to symbolize the group identity of the 45th’s President electorate and the strength of their shared consciousness. Research shows that this identity was mostly centered around being White (Sides et al., 2017). The aim of this paper is to understand how this group identity was fostered and in which ways online and offline social capital played a role in that process.
The Trump electorate has been widely studied and trying to understand the reasons behind this vote choice has been the source of much scholarly work. Up to this point, the average Trump voter drawn by research and exit polls is one of a poorly educated and rather wealthy White man (Cook et al., 2017). In his book, Alienated America, Tim Carney describes the Trump voter as poor in terms of social capital, something he broadly defines as social connections and the sense of belonging to a community. Carney describes an America that has lost the sense of belonging to a Nation coupled with a slowly disintegrating feeling of group consciousness, especially among White Americans who feel threatened by changing demographics (Filindra et al., 2020). Scholarly accounts of group voting among White Americans have shown a positive correlation between White identity and a vote for President Trump (Sides et al., 2017). Other scholars have focused on the impact of social media on the 2016 campaign and the growing distrust in the traditional so-called mainstream media within Trump’s voter base (Gerbaudo, 2018). Gerbaudo explains how the internet has provided populists a new political agora where ‘lonely crowds’ could become ‘an online crowd of partisan supporters’ (Gerbaudo, 2018: 750).
This broad picture of the Trump voter touches upon different concepts and theories of voting behavior and potentially put them in conflict.
Concepts such as group voting and group consciousness have demonstrated that a strong feeling of belonging to a group (i.e. ethnic or cultural) is a mobilizing factor leading to increased political participation (Huddy, 2013). However, a strong identity can not be formed in a vacuum, it requires contextual factors as well as a community to foster the sense of belonging. The kind of social networks that foster group belonging can be found in theories of social capital (Putnam, Bourdieu…). Holding high levels of social capital is positively associated with a strong sense of group belonging and increased political participation (Brehm & Rahn, 1997). The new networks and relationships formed on social media have led certain scholars to develop the concept of online social capital (Zúñiga et al., 2012). They argue that the ties made online can foster and maintain group identities. The use of social media has also been found to be a mobilizing force when it comes to elections (Ellison et al., 2007).
The 2016 American Presidential election seems to be an interesting case study. While Trump voters have been characterized as a population deprived of social capital (Carney, 2019 ; Giuliano & Wacziarg, 2020) they have also been described as very active on social media (Gerbaudo, 2018). Moreover, it seems that identity and group consciousness were central to the Trump vote. This observation indicates that Trump voters, despite displaying low levels of social capital, did engage in politics and boasted a pretty strong sense of group belonging. This leads me to further investigate the relationship between group identity/consciousness and offline social capital and online social capital. The research question then, revolves around the assumption that in the 2016 election, online social capital might have trumped the negative effects of the lack of offline social capital for the Trump electorate in the mobilization of their White identity.
Studying this question is valuable for different reasons. First, it allows one to better understand the potential impacts of social capital, whether it is formed online or offline, on electoral participation and more specifically as it relates to the activation of group identity/consciousness. Second, it is a way to give more depth to the already well-developed literature on the Trump electorate. Finally, and from a broader perspective, studying this topic sheds light on broader concepts such as populism, polarization or the impact of social media on contemporary politics.
This paper relies on ANES data from the 2016 Post-election Time Series Study, to draw a precise picture of the Trump electorate, its offline and online social capital levels as well as the salience of its White identity.
The next section provides a more comprehensive review of the literature on the topic together with a presentation of key theoretical concepts used to develop my hypotheses.
Theory
Social capital
The concept of social capital was popularized by Robert Putnam in a 1993 article in which he broadly defines it as “features of social organization, such as networks, norms, and trust, that facilitate coordination and cooperation for mutual benefit” (Putnam, 1993). Later on, in his famous book Bowling Alone, Putnam argues that declining social capital among a large share of the American population is responsible for decreasing trust and, as a consequence, decreasing political participation. His definition, although widely cited, has been criticized. Woolcock argues that Putnam focuses on the consequences of social capital and not on the processes that explain its formation (Woolcock, 1998). A lesser known definition of social capital, at least in the American literature, was developed by French sociologist Pierre Bourdieu. His account of social capital is more focused on its sources. He sees it as a pool of relational resources that are developed and then used to build stronger, more resilient communities (Siisiainen, 2003). From a more concrete standpoint, social capital can be understood as a web of relationships developed through membership in an organization, a church or simply involvement in one's community.
Social capital has been positively linked with increased political participation (Brehm & Rahn, 1997). The reasoning goes as follows: relationships build trust, this trust gives confidence in the idea that solutions can be found in collective action and that the common good is reached through consensus rather than by the pursuit of individualistic ideals. This dynamic then turns into a virtuous circle, with trust in others expanding into trust in institutions.
New tools such as social media have sparked the interest of researchers interested in social capital. Social media sites such as Facebook, allow users to transfer their offline relationships online. But it also allows them to meet new people through groups for example. These groups can be centered around shared hobbies, interests or ideologies. This new form of building and maintaining trust has led some researchers to develop the term “social media social capital” (Zúñiga et al., 2012). The use of social media has also been positively linked to increased political participation (Ellison et al., 2007). Moreover, social media seems to be particularly important for populists electorates, and in our case, the Trump electorate (Gerbaudo, 2018). Social media platforms were used as much as information outlets as well as virtual gathering spaces for Trump supporters (Gerbaudo, 2018).
The Trump electorate of 2016 has been described as having low levels of social capital. Low levels of social capital are usually attributed to those who choose not to vote. From a theoretical standpoint this is somewhat surprising and puzzling. How were Trump voters mobilized even if they had low levels of social capital? One possible answer could be found in their usage of social media, it can be hypothesized that their low levels of offline social capital were counterbalanced by high levels of online social capital. Therefore I formulate the following hypotheses:
H1: Trump voters held lower levels of offline social capital than the average 2016 voter
H1a: Trump voters held higher levels of online social capital than the average 2016 voter
Group identity and social capital
The term White anxiety has been used a lot to try and explain the reasons that pushed so many Americans to vote for Donald Trump in 2016. The appropriate theoretical framework to analyze this claim can be found in the rich literature on group identity and group consciousness. Leonie Huddy reviews this literature in a very comprehensive way and outlines a few of the most important findings related to group identity and voting (Huddy, 2013). There are different kinds of identities, some are conscious while others are not. Some are politicized, others aren’t. Being White in America can be considered as being a social identity that has gained political content, and in some context, has become politicized (Miller et al., 1981). If it is the case, group identities can lead to group voting, meaning that a group will vote to defend its status (Michelson and Valenzuela, 2016). The context plays an important role in the mobilization of these identities, identity saliency can be triggered by a context, a feeling of status loss can for example fuel participation (Bobo, 1983). Some candidates can also capitalize on these situations and strategically appeal to these feelings (Simien and Hampson, 2017; Stokes-Brown, 2006).
Social capital and group identity and consciousness are related. In his seminal work Putnam explains that social capital negatively impacted participation because of the loss of group belonging, which can be a substitute for group identity. A lack of social capital can then explain decreased consciousness of group identity. However, high levels of social capital can increase group identity. Being part of a community, sharing common struggles and finding solutions for a community reinforces the sense of group belonging. Francis Fukuyama discussed the topic and argued that social groups have boundaries and that social capital plays a real role within these boundaries. He also points out that: “Many groups achieve internal cohesion at the expense of outsiders, who can be treated with suspicion, hostility or outright hatred” (Fukuyama, 2001: 8). If internal cohesion is linked to a certain identity we can assume that high levels of social capital lead to a strong sense of group identity, built in part in the rejection of an out-group.
The same logic can be applied to online social capital. Social media platforms have provided a space for groups to express their identity and increase their consciousness as it allowed them to find people with similar identities. Gerbaudo states the following: “social media discussions have provided gathering spaces where the ‘lonely crowds’ produced by the hyperindividualism of neoliberal society could coalesce, where the atoms of the dispersed social network could be re-forged into a new political community, into an ‘online crowd’ of partisan supporters” (Gerbaudo, 2018: 750).
The dynamics outlined here are very resonant with the 2016 presidential election and especially with the Trump electorate. Studies have repeatedly found that the Trump electorate was overwhelmingly White and that defending White identity was at the core of a vote for Donald Trump (Sides et al., 2017). Other works have found that this identity was activated by a feeling of status loss and fears about the changing demography of America (Major et al., 2018). Candidate Trump also strategically appealed to those fears. My assumptions drawn from the literature about the activation of group identity by high levels of online and offline social capital lead me to formulate the following hypotheses:
H2: Trump voters with higher levels of offline social capital saw being White as more central to their identity than those with lower levels
H2a: Trump voters with higher levels of online social capital saw being White as more central to their identity than those with lower levels
Data & Methods
For this analysis I relied on ANES data from the 2016 Time Series Study post-election survey. The data was treated via the statistical analysis software Stata. I used simple frequency tables to measure online and offline social capital as well as White identity. In the next subsections I briefly explain how I operationalized those measures.
Social capital
There is not a consensus on the way that social capital should be measured. Francis Fukuyama sees two main approaches: the first one consists of going at the source of social capital, meaning looking at group membership, the second one is more concerned with the outcome and measures trust levels in different groups (Fukuyama, 2001). Here, I will use the former. Therefore my measure of social capital relies on four variables from the survey. The first one asks if the respondent is a member of an organization. The second one asks if the respondent has done community work in the past year. The third one asks whether or not the respondent has been to a meeting about a community issue in the past year. Finally, the last one asks the respondent if he or she has done any volunteer work in the past year.
Online social capital
My operationalization of online social capital follows the same logic as that of social capital. There is, however, limited data on social media use in this survey. This limitation will be discussed in the final section of this paper. I measure online social capital using two variables, the first one asking the respondent if he or she has a Facebook account and the second one asking if the respondent has used that account to discuss politics.
White Identity
My measure of White identity is pretty straightforward, as there is a question in the survey that is meant to measure just that. The question asks the respondent if being White is an important part of his/her identity. As this is a non-binary measure, I had to make a choice to decide which respondents had a strong sense of group identity or not. Therefore I consider a respondent as having a strong White identity if he or she has responded “extremely important” or “very important” to that question.
In addition to that, I add another measure which is derived from Fukuyama’s idea that a strong identity can be fostered at the expense of an out-group. Therefore I also look at the racial attitudes of Trump voters with a question from the survey which asks respondents to react to the stereotype that blacks are violent. Respondents were asked to answer this question by placing themselves on a scale ranging from 1 (peaceful) to 7 (violent). I consider a respondent as agreeing to that stereotype if he or she is above 5.
Results
In this section, I present the results of the analysis of the data as they relate to my hypotheses. I then briefly analyze them. A more thorough discussion about the implications of these results is given in the final section of this paper.
Out of the four hypotheses formulated, only one is supported by the data. This might be surprising but some theoretical arguments as well as methodological limitations can be put forward to explain why.
My first hypothesis is supported by the data. Table 1 shows that overall, Trump voters held lower levels of offline social capital than the average 2016 voter. This is consistent with the findings of Giuliano and Wacziarg (2020). For example, only 29% of Trump voters had attended a meeting about a community issue, while 34% of all respondents and 38% of Clinton voters had done so. On this specific measure, Trump voters and abstentionists are similar. This finding shows that low levels of offline social capital are positively linked to abstention or a vote for Donald Trump. Two arguments can be given to explain this phenomenon. For abstentionism, Putnam’s original argument seems to be true: no social capital leads to a disaffection of politics (Putnam, 2000). However, how can we explain that despite displaying low levels of offline social capital, some voters decided to vote for Donald Trump? In their paper, Giuliano and Wacziarg argue that the low levels of social capital made those who voted for Trump ill prepared for financial hardship and decreased their trust in politics which in turn made them blame elites (Giuliano and Wacziarg, 2020). These were both central themes of the Trump campaign.
Another explanation could be that Trump voters, instead of being mobilized thanks to high levels of offline social capital, were boosted by high levels of online social capital. This is what H1a was assuming. However, H1a is not completely supported by the data. As table 2 shows, Trump supporters did not consistently hold higher levels of online social capital than the average 2016 voter. The most surprising finding is linked to the assumption that there was an “electoral affinity” between populism and social media (Gerbaudo). In fact, according to these figures Clinton voters seemed to hold higher levels of online social capital. Overall, these findings show that high levels of online social capital are associated with higher participation in politics, but the “electoral affinity” between populism and social media cannot be entirely supported. Possible limitations to the test of this hypothesis include, very scarce data on social media use as well as possible biases because of the lack of control variables such as age or education levels.
The next set of hypotheses was interested in the relationship between levels of offline and social capital and White identity. My expectation was that higher levels of both capitals would strengthen White identity, thinking that having more social connections would strengthen group consciousness. This reasoning however, is not supported by the data as both hypotheses are unvalidated.
H2 assumed that high levels of offline social capital would strengthen White identity among Trump voters. As table 3 shows, Trump voters who were less involved in their community were more likely to see being White as an important part of their identity and slightly held more racist beliefs. This means that the opposite of H2 is true: Trump voters with lower levels of offline social capital had a stronger White identity and held more negative attitudes towards Blacks.
H2a held the same assumption as H2 but with online social capital. Table 4 displays similar results as table 3, although less significant. Trump voters who were less active on social media generally saw being White as more important to their identity than those who were more connected online.
Again, several arguments can be used to explain these results. First, and as stated earlier, it should be pointed out that from a methodological standpoint, the measure used for online social media (and therefore social media use) is somewhat weak. Moreover, these tests are not as precise as they could be. Control variables such as age, place of residence as well as education levels would be a great start to make these results more significant. Secondly, these results suggest the following: in this particular case low levels of online and offline social capital are positively linked to increased White identity and increased negative feelings towards Blacks. A way of explaining this might be that, as Putnam suggests, more social capital leads to increased belief in the solutioning of problems collectively. This would mean that instead of reassuring oneself of his/her identity in a web of relationships in a homogenous bubble of similar peers, social capital (of both forms) allows people to open up and accept differences. The next section develops that train of thought and gives ideas for future research on the topic.
Discussion
The initial aim of this paper was to understand how, despite displaying low levels of offline social capital, Trump voters were mobilized as well as how social capital might have been an important stimulator of the saliency of White identity. It was hypothesized that high levels of online social capital trumped the negative effects of the lack of offline social capital. It was also expected that a lot of social capital was positively linked with a stronger sense of White identity. Both of these assumptions did not stand the test of the data, as evidence could not be found. Although not validating my hypotheses, these results spark interesting thoughts.
First, what can be said for sure is that Putnam’s original idea that low levels of social capital are linked with lower participation, is true. But what these results show is that social capital is not only linked to participation or abstention, it also seems to be ideologically loaded. Consistently, results show that Clinton voters boasted higher levels of both capitals than Trump voters. This indicates that, in 2016 at least, higher levels of social capital were positively linked with a democratic vote. This, again, makes Putnam right, and suggests that social capital strengthens a sense of belonging to a Nation, leading to inclusiveness, not a populist nationalism based on race. Future research could compare levels of social capital across different elections to test this ideological affinity. Tim Reeskens and Matthew Wright’s work on “cross-group trust” and the “civic” or “ethnic” definition of the Nation could be useful (Reeskens & Wright, 2013).
What these results show, is also a somewhat more contrasted picture of the Trump electorate. While previous research has described the Trump electorate as very virulent on social media and very strongly vocal about its White identity, this paper shows a more tamed picture. Maybe, the Trump electorate was less of a virulent mob of White supremacists, strongly active on social media and more of a collection of socially disconnected and “alienated” (Carney, 2019) individuals, brought together not by a strong group consciousness, but rather by an unconscious shared despair, capitalized on by the candidacy of Donald Trump.