University Students' Social Capital, Social Media Usage and Democratic Citizenship Behaviours

The aim of the study is to investigate the relationship between university students’ level of democratic citizenship behaviors and their social media usage through social capital theory. University students’ involvements in democratic issues are both affected by and happens though social media due to the rapid developments in mobile technologies. Thus, it is worthwhile to explore its effect. The study utilized a quantitative approach for both data collection and analysis. The participants of the study were students of Turkish state universities. The total number of the students involved in the study was 2253. The data was gathered by 7 scales developed for the study. One of the scales was yes/no type and the rest were Likert type scales. The data was subjected to stepwise regression analysis. Among all the variables that were in regression models, it appears that ‘trust’ is the most important single variable in explaining students’ different citizenship behaviors.


INTRODUCTION
Social capital theory has been around for some time as a useful tool for researchers. In Turkey, mostly researchers majoring in politics, economics and development appear to be having an interest in social capital. There is only limited attention paid to the relationship between social capital and democracy in educational settings. Recent developments in social media and its use in democratic movements make it important to pay special attention to the relationships among social capital, democratic citizenship, and social media in Turkey.
There are many opinions about the definitions and dimensions of social capital which, however, still has several agreed upon basic features. In other words, it is accumulated by people through interaction with the others, and they can utilize its benefits freely (Coleman, 1998;Putnam 2000;Lin, 2005). Social capital is generated by engaging in voluntary activities for the good of the small group or public in general through the organized social networks (Pattie, Seyd & Whiteley, 2003;Ranieri, Manca & Fini, 2012).
According to social capital theory, involvement in local networking, and carrying out voluntary work within the network generate trust among individuals (Putnam, 1993cited in Pattie, Seyd & Whiteley, 2003. Societies with high trust levels appear to engage more in social, political, and environmental activities as opposed to societies in which the level of trust among individuals is low (Hall, 1999;Whiteley, 2000). Consequently, it is theoretically expected that there should be a link between trust, voluntary work, and active democratic citizenship. Literature in this field is very limited in Turkey (see Ozdemir, 2007;Ekinci, 2008;Gungor, 2011;Gunkor, 2011;Ekinci & Karakus, 2011;Yucel et al., 2013).

Social trust and democracy
Social capital theory emphasizes the importance of trust in healthy democratic societies. It is an indicator of social capital in socially integrated, economically effective, and democratically strong societies (Newton, 2001). In the theory, the concept is used in the sense of both trust at an individual level and trust in democratic institutions (Newton & Zmerli, 2011).
Trust plays a central role in creating both bonding and bridging social capital. Bonding social capital occurs in small, close, and comprehensive groups of people, while bridging social capital happens in relatively larger and distant groups, and allows intergroup relationships. As with trusting each other, trusting in institutions in vivid democracies is very important for people to act upon their beliefs.
As established above, trust is the basis for people to act upon. The intricateness and depth of the relationships depend on people's trust in each other as well as in lawfully governed governmental and civilly managed non-governmental democratic institutions such as courts of justice, NGOs charities, parliament and so on (van Deth, 2010).

Bonding and bridging social capital
Bonding social capital is built among people who know and support each other personally such as family and close friends (Putnam, 2000). This provides members of any close group with access to jobs, emotional and financial support, and so on. Individuals find this kind of support at their disposal very comforting in difficult times. The kind of closed grouping that generates bonding social capital may also function as an isolating barrier to the individual, denying access to any other group activities (Paxton, 2002;Zmerli, 2010). This might especially be true for strictly religious and/or very exclusive marginal political groups. In theory social capital is desirable in order for a democratic society to function in cohesion. Social capital is not a one-off happening, rather it is created through the collective actions of individuals with some sort of regularity in groups (Maritz & Prinsloo, 2015).
If intergroup contacts are limited or prohibited because of the grouping that creates bonding social capital, then the desired effect of social capital in any democratic society is far from being achieved. It is for this reason that theorists are inclined to assert that, for the democratic societies, bridging social capital is desired (Lin, 2005).
Bridging social capital, on the other hand, happens when an individual or a group of people interact with other individuals in other groups, thus creating networks that are driven to work for the good of the public in general. Networking is the core of social capital (Maritz & Prinsloo, 2015), the power generated by the weak links in the society (Granovetter, 1973). Networks and relationships are important as they affect the performance of the individual or organizations (Angervall, Gustafsson & Silfver, 2018). Trust being the central term in explaining how weak links can culminate such power for the bridging social capital, one should also pay attention to it when trying to explain the relationship between social capital and self-esteem, and satisfaction with life.
There are conflicting research findings about the relationship between the social media and self-esteem and satisfaction with life. In other words, some researchers assert that individuals with low levels of self-esteem and life satisfaction use social media to compensate for the lack of interaction and trust in real life, while others claim that individuals' self-esteem and life satisfaction levels have no bearing on their social media usage -on the contrary, the very nature of the internet, which allows for connection and interaction without time and space restrictions, creates another kind of social capital (Hofer & Aubert, 2013).

Social capital and social media
A person's perception of trust constructed through the interpretation of his/her views about the events and social phenomena has become ever more important for the political system and its institutions. That is, a person's trust in his/her close proximity comes from first hand life experiences. However, his/her perception of trust in the political system and its institutions is generated through the reality presented by the media. In this case there are many new factors at play in the creation of political trust. Some of these factors are media ownership and the relationship between the government and media patronage, media literacy of the public and sources of information for individuals.
Social media provides a special and in a way stable social address for individuals without time and space restrictions. Moreover, they can create connections with individuals and groups again and again. Advanced algorithms make it possible for individuals to give updates, automated replies, reminders, and so on. Thus, they can make the most of social media for creating and improving close relationships. In this way, they create and use online social capital through social media when and where needed (Quinn, 2016).
One of the most important factors for a democratic system to work is the existence of independent media. The media's role was so crucial even before the internet and social media that theorists studying democracy theory called the media the fourth estate alongside the legislative, judicial, and executive powers (Erdogan, 1999). As in every part of human life, the internet and social media have much more impact on the democratic system and its institutions. Its power on people's opinions and perceptions about politics is immense (Afsar, 2004). Thus, the power relationship between media patronage and the government is always important for a democratic system to function as it should. The social media have accumulated a similar power in recent years by affecting people's ideas and creating public opinion (Vural & Bat, 2010;Babacan, Haslak & Hira, 2011;Babacan, 2012;Shin & Choi, 2017).
The political uprising and protests in Arab states, the so called Arab Spring, demonstrated very effective uses of social media in creating public opinion and organizing protests (Babacan, Haslak & Hira, 2011).

Democracy, active citizenship, and social media
'In today's society, anti-pluralist and anti-democratic views are on the rise, with subordination of minorities and questionable truths being launched, resulting educational equality being threatened' (Bergmark & Westman, 2018, p. 1,354). What is needed is the promotion of democratic values as a counter measure.
The most important indicator of democratic societies is their citizens' democratic citizenship behaviours. The 'active citizenship' term is sometimes used in relation to voluntary charity work in social capital theory (Pattie, Seyd & Whiteley, 2003;Dalton, 2008;Zmerli, 2010).
However, active citizenship includes individuals' acts of mundane citizenship behaviours arising from being citizens of any state. What is important here is that individuals should be taking initiatives to make a difference in societal or environmental issues that they are not primarily responsible for. In that case we would be talking about the relationship between social capital and active citizenship behaviours (Pattie, Seyd & Whiteley, 2003;Dalton, 2008;Zmerli, 2010).
Social media acts as a platform for people to be active and responsible citizens in a strong democratic state (Smith, 2013;Warren, Sulaiman & Jaafar, 2015).
Similarly, Shin and Choi (2016) have also found that social media not only help people perform democratic engagement at the macro level but also mediate people's relationships for civic engagement. There is a strong correlation between online and offline civic engagement and social media use (Hwang & Kim 2015).

METHOD Research design and subjects of the study
The study utilizes quantitative research approaches for both the collection and analysis of data. A demographic data collection sheet, six Likert type scales, and a yes/no question scale were developed for the study.
The data was collected from 2,253 students who were involved in the study with varying response rates from several universities. It should be noted that 'the sample of the study' does not reflect the intended representation of the population under investigation. Therefore, the results cannot be generalized.

Data gathering tools
The demographic data collection tool included many items to create a socio-economic status index. It also included the students' grand point average of, gender, membership of any NGOs on or off campus, political views, type of social media use, and frequency of social media use. Seven other data gathering tools developed for the study were subjected to the same procedure. A pool of items was developed for each scale drawn from the literature. Then, the validity check procedures were completed by utilizing expert opinions for each scale. A pilot study was carried out for all scales for both validity and reliability purposes. In order to test whether the scales were suited for factor analysis, the Kaiser-Meyer-Olkin test was run. Cronbach's Alpha was used to test the reliability of the scales.
Purposes of social media use scale (KMO= .77, p<.001; .80 Cronbach's Alpha; total variance explained = 62%): The factor analysis shows that the instrument contains 12 items and four sub factors. Those factors are named 'entertainment and killing time', 'keeping up to date with politics and political groups', 'research' and 'networking'.
Self-esteem scale (KMO= .84, p<.001; .80 Cronbach's Alpha; total variance explained = 56%): The factor analysis shows that the instrument contains five items and one factor. Trust (KMO= .75 p<.001; .88 Cronbach's Alpha; total variance explained 74%): The factor analysis shows that the instrument contains four items and one factor.
Life Satisfaction (KMO= .87 p<.001; .87 Cronbach's Alpha; total variance explained 66%): The factor analysis shows that the instrument contains five items and one factor.
Online and face to face bridging social capital (KMO= .91; and .88 p<.001; .91 and .87 Cronbach's Alpha; total variance explained 69.97% and 64.31% respectively): The factor analysis shows that the instrument contains eleven items and three sub factors.
Online and face to face bonding social capital (KMO= .93; and .91 p<.001; .93 and .89 Cronbach's Alpha; total variance explained 69.50% and 60.72%): The factor analysis shows that the instrument contains eleven items and two sub factors.
Sources of information (KMO= .85, p<.001; .86 Cronbach's Alpha; total variance explained 71.64 %): The factor analysis shows that the instrument contains 18 items and six sub factors.
Trust in democratic institutions (KMO= .94, p<.001; .95 Cronbach's Alpha; total variance explained 74.25%): The factor analysis shows that the instrument contains 27 items and five sub factors.

Analysis
Stepwise regression techniques were used to explain variance in students' democratic citizenship behaviors.

RESULTS
The analysis of the democratic citizenship behaviours scale shows that the lowest mean of university students' democratic citizenship behaviours is observed in the 'extreme political activism' dimension. Almost all of the students regularly vote, however their voting behaviours do not correlate with their participation in the activities of political parties.
In order to understand the factors affecting students' democratic citizenship behaviours, a stepwise regression analysis procedure was carried out. The findings are presented in Table 1 below. A total of 19 independent variables were included in the regression model to explain students' organized citizenship behaviours. A linear combination of 19 independent variables, altogether, explained 23 percent of variance in 'organized citizenship behaviors' of the students (R= .48, R2= .23, Ad. R2 = .22, Std. Err. = .22, F = 27.10, p<.01). All the other variables did not contribute or increment into the variance in organized citizenship behaviors above and beyond the contribution of these variables.
The highest unique contribution in explaining the variance in students' organized citizenship behaviours is made by students' level of trust in non-governmental organizations (3.5%).
The students who use the social media to create networks and follow the people and groups that are politically in line with their own political orientation have more tendency to exhibit organized citizenship behaviours. Those students' inclination to exhibit organized citizenship behaviours is higher than the students who use social media to do research and homework, gather information, kill time and entertainment.
As for trust in the institutions of the democratic system, trust in NGOs is crucial for organized citizenship behaviours. As opposed to this, those students whose trust in the judiciary and law enforcement agencies is high have less inclination towards exhibiting organized citizenship behaviours.
Those students who define themselves as apolitical do not exhibit organized citizenship behaviours. Contrary to this, the students who define themselves as socialist/Marxist or social democrats are more inclined to exhibit organized citizenship behaviours. There is no statistically significant relationship between organized citizenship behaviours and defining oneself as conservative, liberal, nationalist, anarchist, ultranationalist, or Islamist. These students' level of organized citizenship behaviours' are very similar to and almost indistinguishable from one another.
Students who have membership of an NGO or a students' club in the university, as opposed to those who do not, tend to exhibit democratic citizenship behaviours such as taking part in democratic protests, organizing petitions and boycotts to certain companies and products. However, not every kind of club or NGO membership is a determinant in explaining their democratic citizenship behaviours. Being a member of a club which is about sports, entertainment, education, art, women's rights, or the environment does not make any difference in democratic citizenship behaviours compared to being a member of a religious club. A total of 19 independent variables were included in a stepwise regression model to explain students' membership of political parties and their voting behaviours. A linear combination of 19 independent variables, altogether, explained 23.8% of the variance in membership of political parties and voting behaviors of the students (R= .49, R 2 = .24, Ad. R 2 = .23, Std. Err. = .28, F = 38.62, p<.01). All the other variables did not contribute or increment into the variance in membership of political parties and voting behaviours of the students above and beyond the contribution of these variables.
Trust in politics explains 10.9% of the variance in 'membership of political parties and voting behaviors' of the students when all the other variables remain fixed ( Table 2). The students who trust in political parties, members of parliament, the prime minister, and the president tend to exhibit membership of political parties and voting behaviours while the students who trust in the judiciary system do not show such tendencies. A total of 18 independent variables were included in a stepwise regression model to explain students' extreme political activist behaviours. A linear combination of 18 independent variables, altogether, explained 38.6% of the variance in extreme political activist behaviours of the students (R= .62, R 2 = .39, Ad. R 2 = .38, Std. Err. = .28, F = 60.33, p<.01).
Trusting law enforcement officers makes up the largest part for not showing extremist behaviours, explaining about 6.2% of the variation alone when the other variables remain fixed ( Table  3).
The students who trust the political parties, members of parliament, the prime minister's office, and the president's office have low levels of 'political extremism' behaviours. On the other hand, the students who trust the media do tend to show political extremism behaviours. Like those who trust law enforcement officers, the students who trust the good nature of mankind tend not to be political extremists.
Students who define themselves as socialist / Marxist, anarchist, or social democrats, exhibit extreme citizenship behaviours higher than other students. Conservative, liberal, nationalist, idealist, Islamist, and apolitical affiliations are not determinants of being extreme activists for university students.
Similar to students who are members of a group within or outside the university, university students in political communities, social assistance and solidarity communities, recreational communities and environmental communities tend to exhibit higher politically extreme democratic citizenship behaviours than those students taking part in clubs such as sports, arts, education, religion, and women rights.
University students who are political activists are less inclined to use religious resources and religious figures as a source of information. Gender is an explanatory variable in the behaviour of political extremism and activism. Other variables such as self-esteem, satisfaction with life are not explanatory variables. As a result of the regression analysis conducted to explain the behaviour of the university students in joining civil society, 14 variables that could explain the dependent variables were identified, and this accounted for 23.4% of the total variation. Those variables were treated as dependent variables (R= .48, R 2 = .23, Ad. R 2 = .23, Std. Err. = .27, F = 37.81, p<.01).
It is seen that confidence in non-governmental organizations, which explains about 8% of the variance in political activism behaviour in non-governmental organizations alone when the other variables are fixed, has the highest explanation power among the variables included in the model (Table 4). University students who are members of any group within or outside the university tend to be more active in non-governmental organizations than students who are not members of such a community.
Students who define themselves as apolitical tend to be passive in non-governmental organizations. Nevertheless, just being conservative, liberal, nationalist, socialist/Marxist, anarchist, social democrat, nationalist or Islamist does not seem to be the sole determiner in students' being active or passive in civil society organizations. Students' confidence in civil society institutions alone is the strongest determinant of being active in NGOs. If students' confidence in the judiciary and law enforcement agencies is low, their tendency to be active in non-governmental organizations increases. As a result of the regression analysis conducted to explain the behaviours of the university students interacting with the politicians they met, 11 variables that could explain dependent variables were identified. These 11 variables accounted for 12.3% of the total variation in interaction behaviours with politicians (R= .35, R 2 = .12, Ad. R 2 = .12, Std. Err. = .34, F = 21.99, p<.01).
Students' participation in political communities alone explains 2.3 % of the variance in interacting with politicians when the other variables are fixed (Table 5). The ability to connect with the close family and friends through the internet increases the tendency of university students to interact with politicians. Students with higher socioeconomic status were found to be more likely to interact with politicians.
When trust in the institutions of the democratic system was examined, it was found that students who have a positive tendency towards trusting in human nature and trusting in politics have a higher tendency to interact with politicians. However, trusting the judiciary, the media, civil society organizations or law enforcement agencies have no effect on exhibiting trusting behaviour to politicians'. While students who use social media to be aware of political issues and groups tend to be more inclined to interact with politicians, those who use the social media to kill time, do research, and stay connected with the groups they are interested in are not prone to this behaviour. As a result of the regression analysis conducted to explain the voting behaviour of university students, 11 variables that could explain dependent variables were identified. Those 11 variables, altogether, explain 11.5% of the variance in the voting behaviour of university students (R= .34, R 2 = .12, Ad. R 2 = .11, Std. Err. = .29, F = 17.26, p<.01).
University students' confidence in politics is the strongest variable that explains the behaviour shown at this dimension by explaining about 2.8% of the voting behaviour alone when the other variables are fixed (Table 6). In addition, students with higher socioeconomic status were found to be more inclined to the voting behaviour.
The university students who tend to devote themselves to others' good in their face-to-face relationships are inclined to the voting behaviour. Students who define themselves as social democrats are inclined to vote. Students with an Islamist, apolitical, or anarchist worldview, on the other hand, tend not to vote. Having a conservative, liberal, nationalist, socialist/Marxist, or idealist worldview has not been found to be a determinant in the voting behaviour.
Students' participation in sports communities increases their tendency to vote while nonparticipation in political, religious, entertainment, arts, education, women's rights, environment or social assistance communities does not cause a change in this behaviour. It was also found that students with a higher GPA were more likely to vote than those with a lower GPA. While the university students who use social media for research purposes tend to vote, no relationship was found between the voting behaviour and social media usage to know about political issues and other students' clubs. In addition, students who are members of any group, whether inside or outside the university, are less inclined to vote.

DISCUSSION
Among the six different dimensions of the democratic citizenship behaviours scale, the highest score was observed in the voting dimension. The lowest behaviours are seen in the dimensions of interaction with politics, participation in political party activities, and political extremist activism. Everyone votes, with or without regular affiliation, because it is seen as a civic duty. On the other hand, other political activities take a lot of time and effort, and sometimes carry risks for the future. Moreover, no one trusts the institution of politics, as proved by the evidence obtained from the measurement of confidence in the institutions that operate within the democratic system, which is not reported here. As such, it is normal for political attitudes of democratic citizenship to remain low. Creating a culture of democracy in universities is essential not only for political participation but also for motivating students to achieve better standards (Bergmark & Westman, 2018).
One of the two notable findings here is participation in civil society activities, albeit relatively low. One of the pioneering indicators of social capital of both the bridging and the bonding dimension in terms of university students is their membership of any non-governmental organization. According to social capital theory, groups of people with significant functions in the social structure voluntarily agree to take initiative as indicators of the healthy functioning of the community (Pattie, Seyd & Whiteley, 2003). About 65% of university students are not involved in any student clubs or NGOs. However, it is understood that students are engaged in a variety of activities through various nongovernmental organizations, albeit not so much to the point of changing the society and the world they live in, as explained in social capital theory.
Research on university students and social capital reveals that the internet and social media are important in two respects. The first of these is the situation that the accumulation of social capital can be realized due to the formed associations on the internet. Second, the use of social media tools has an effect on social capital accumulation both on the face-to-face relationships and through the internet (Williams, 2006;Greenhow & Burton, 2011;Hofer & Aubert, 2013). It can even affect international students' mobility decisions (Herman & Kombe, 2019).
The other noteworthy finding is that the score is above the average 6 in the dimension of organized citizenship reaction. When the items of this dimension are examined, it can be seen that students' behaviours such as signing petitions and protests in various social, economic, educational, and political matters are more likely to happen if they join in any activities with NGOs. This is the most important aspect of the active citizenship dimension of democracy theory (Pattie, Seyd & Whiteley, 2003).
A number of regression analyses were conducted to examine the democratic citizenship behaviour in greater depth. It is understood that the most important variable in explaining students' organized citizenship behaviour is trust in NGOs. At the same time, those who are very confident in the functioning of the judiciary and the law enforcement agencies within the democratic system are less likely to exhibit organized citizenship behaviour. This is a sign that those who are satisfied with the possibilities that the system offers to them believe that they are not responsible for any organized citizenship behaviour. However, those who are traditionally at the edge of the political spectrum (Marxist / communist / socialist) are the ones who exhibit organized citizenship behaviour the most. Again, as might be expected, those who define themselves as apolitical, on the contrary, do not exhibit any organized citizenship behaviour.
The same situation is observed in the dimension of extreme citizenship behaviour. The trust variable for the law enforcement officers who are most vulnerable to not showing these behaviours can account for around 6.2% of variation when the other variables are fixed. The students who express confidence in political leaders, parties, members of parliament, prime minister and presidency, and politicians are less inclined to show political extreme behaviour in the political mechanism.
In the context of social media and democratic citizenship behaviour, some of the findings need to be addressed. While it is evident that the students who use social media to stay informed about political issues and groups tend to be more active in non-governmental organizations, the students who use social media to do research, prepare assignments, and gather information about topics that they are curious about are found to be more passive in non-governmental organizations.
Students' level of trust in institutions and citizenship behaviours that will contribute to the accumulation of bridging and bonding social capital are too low to witness in established democracies with a democracy culture.
In light of the above results, it is possible to offer various suggestions for both research and education.

CONCLUSION AND A RECOMMENDATION
The first conclusion to be drawn from the research is its limitation in terms of a good representation of the population. In such a quantitative study, researchers should start with the stratified sampling or cluster sampling technique from the very beginning in order to best represent the population.
Another conclusion is that, contrary to popular belief in society, the majority of students do not use social media very often. However, considering the usage purposes of those who do, entertainment, killing time, and socialization stand out. For purposes of better use, educational activities could be organized earlier in the year to raise awareness of more meaningful purposes beyond killing time.
Because of its importance for strong democracies, social capital accumulation among individuals and groups can also develop with educational measures in the early years. For example, awareness raising in the conservation of the environment, encouraging participation in voluntary activities, and developing cooperation and collaboration skills can be addressed in early years' education. When these requirements are met, it can be expected that social capital, especially bridging social capital, will increase if it is combined with the belief that doing so is necessary.
For the effective use of information resources, it is necessary for the students to acquire the skills to find, manipulate, and present scientific, up-to-date information. It is also evident that the acquisition of scientific process skills to acquire scientific knowledge as the most enduring source of knowledge must happen at the level of very early education and not at university level.
The development of trust in democratic institutions and democratic citizenship behaviour is not possible with any single suggestion that any single person or institution can put forward. However, the research findings clearly show that the existing system has resulted in mistrust in democracy and democratic institutions among individuals in Turkey.
What then needs to be done is to take legal, administrative, economic, and educational measures to solve the structural problems of democratization. The responsibility lies with the state. The state has the power to create culture.