The ways in which society and politics interacts is a central theme in political sociology. One of the topics that can be derived directly from this line of reasoning is concerned with links between social relations and political regime. In this respect, the research on social capital is noteworthy. Introduced to wider social science research by Robert Putnam (1993), it was a highly investigated and discussed phenomena that received considerable amount of attention, albeit only for a decade. Although his inquiry into this issue has received substantial amount of criticism (e.g. Levi 1996; Tarrow 1996), he posed an important question: do relations between individuals have an effect on the functioning of democratic institutions? Or, more precisely in the context of current paper: does social capital promote democracy?

The latter question has been addressed by various authors, but probably one of the most cited examples here would be a study by Pamela Paxton (2002). Comparing different countries using a rather novel method, she found that that social capital indeed promotes democracy, despite that the relationship was found to be reciprocal (Paxton 2002:272). However, if one is to draw any further conclusions from the Paxton’s article it is of critical importance to be mindful of avoiding ecological fallacy by claiming that individuals who possess more social capital are somehow more beneficial to democratic development. As Kenneth Newton (2001:211) suggests, there is a relationship between social capital and its democratic consequences, but this cannot be observed on the individual level and even on the system level this interaction is indirect and complicated as Paxton’s findings seem to confirm. When using countries as unit of analysis, the only conclusion we can make is that a country that’s population possess more social capital is in general more likely to be democratic. I argue that this says little about the actual association between the two phenomena and in order to better grasp the relationship it is necessary to approach this issue at the lowest, i.e. individual level.

This is the task that the current paper will undertake. It seeks to ascertain whether individuals with a higher level of social capital are also more likely to hold democratic values and tend to behave in a way that promotes democracy. The paper begins with describing the results of some previous research on this issue, then continues to conceptualize the two phenomena under investigation. This is then followed by explaining the theoretical relationship between social capital and democracy and stating of hypotheses to be tested. In the last part of the paper the methodology of a quantitative analysis is described and the analysis itself is carried out.

Previous research

In social capital research the previously mentioned Putnam’s work (1993, 2000) has been in a way seminal to this field of study and has hence gained much attention. His aim in the 1993 book was to examine how social capital affects democratic governance in Italy (Levi 1996:48), but for the most part he was focused simply on the performance of democratic institutions (Paxton 2002:255; Tarrow 1996:395), which is different from the questions posed in the current paper and will not be thus described here in further detail.

Although the results of Paxton’s (2002) work were already briefly outlined, her theoretical as well as methodological approach would still deserve some more attention. For the measurement of social capital she employed data on social trust and belonging to different associations mainly collected from World Values Survey and democracy was assessed by utilizing Bollen’s measure of liberal democracy (Paxton 2002:262). She concluded that on a system level there is a correlation between social capital and democracy, but she had crucial difficulties in establishing exact causal relationship between the two phenomena.

This finding regarding the correlation on a national level is also supported by Newton (2001) who assessed the relationship between social and political trust and its implications on civil society and democracy. Also making use of World Values Survey data he was, however, unable to confirm a relationship between social and political trust, or even between trust and membership in voluntary associations on individual level.

Perhaps most relevant findings in the context of current paper have been made by Robert M. Marsh (2005). Focusing explicitly on individuals, he found evidence supporting the assumption that belonging to different organizations adds to democratic political behavior, while it has no effect on political attitudes. Contrary to associational participation, trust seemed to have no impact on such political behavior and even had a negative effect on democratic attitudes (Marsh 2005:607-8). These findings are yet somewhat limited, since Marsh’s work was only concerned with Taiwan.

This is the first respect in which the current paper intends to differ from previous studies. The relationship will be assessed among a substantial number of different societies by employing a considerably large dataset. Second, the investigation will not be limited to only national level, but individuals will be used as unit of analysis.

Social capital and individual democracy in theory

In this section the two central phenomena in this paper, namely social capital and individual democracy will be conceptualized.

Social capital

Even though the term social capital was first mentioned almost a century ago (Putnam 2000:19), as noted above, it was Putnam who brought this idea into macrosociological theory only a few decades ago (Paxton 2002:256). In this Putnam’s first work on social capital he described the term as referring to “features of social organization, such as trust, norms, and networks that can improve the efficiency of society by facilitating coordinated actions” (Putnam 1993:167). Or, as summarized by Farr (2004:49), Putnam thought of social capital as “the network of associations, activities, or relations that bind people together as a community via certain norms and psychological capacities”. From various works that are involved with social capital it is possible to identify at least 20 different definitions of the concept (Adler, Kwon 2002:20) that typically seem to emphasize the importance of different networks that promote cooperation. This suggests that social capital is most dependent on associations that establish links between people, as well as on the strength of these associations and the ways in which individuals gain from them (Sobel 2002:139). As such, social capital can be perceived as an important resource (Adler, Kwon 2002:27) that could be used for social bargaining by individuals. Social capital can then be said to be not only cultural, but rational in essence. Although maintaining of social capital also requires resources, these “expenses” are not commonly material, while the gains from social capital can be material (Sobel 2002:143-4). Hence, social capital is not simply a romanticized vision of how societies work.

As follows, social capital can be said to encompass three phenomena: trust, social relations and norms. These are frequently thought of as and measured by trust that individuals have towards others and the number of associations they belong to (e.g. Marsh 2005; Newton 2001; Paxton 2002; Temple, Johnson 1998:985). Trust is perhaps the most significant of these and it’s important to note that in this context it refers to social and not political trust. Also, since the exact meaning of this term is highly ambiguous (Newton 2001:203), it is usually based on plainly subjective judgment of respondents. The main function of associations in social capital is to establish links (Newton 2001:206) and therefore also trust (Marsh 2005:594) between people. However, it has often been argued that instead of “bridging” individuals, some organizations could only “bound” people together (Sobel 2002:151-52), which has a negative, rather than positive effect on the functioning of social capital (Marsh 2005:612; Paxton 2002:272). Still, at least on a theoretical level, all kinds of associations create some kind of social capital (Levi 1996:52) and excluding such associations from analysis would be overly normative and might thus result in biased outcome of an analysis. For instance, Putnam’s (2000) approach has been criticized precisely because of his limited understanding of social capital that included only traditional, hierarchic and elitist organizations which allowed him to observe the decline of social capital (Inglheart, Welzel 2005:117-8).

Democracy

The literature on democracy is vast and for this reason the different meanings of this term will not be covered as extensively as in case of social capital here. The approach taken here is derived from Vanhanen’s broadly acknowledged and often adopted definition of democracy. According to his conceptualization, democracy refers to a political system that (1) includes the largest possible part of population to electing political power holders who are responsible to the people and (2) allows for a government change which presumes the possibility for different groups to compete for political power (Vanhanen 2000:251-2). Two fundamental requirements for a democracy can be identified from this perspective. First, a procedural condition is that democracy presupposes an inclusive mechanism which allows for political participation, resulting in genuine responsiveness of the power holders to electorate. Second, a more normative requirement assumes that the competition for political power is inclusive and pluralist and would under no circumstances exclude any group for ideological affiliations.

In the current paper, however, it is necessary to conceptualize democracy not on the state, but instead on individual level. Thus, the term “individual democracy” is developed here. The meaning of this concept is derived from the two outlined aspects of democracy and involves both of these. Individual democracy refers to a person’s active participation in democratic practices and maintaining of attitudes that uphold democracy and democratic values. The latter element is substantially different from what Vanhanen described as competition and pluralism, but it more genuinely embraces the normative aspect of democracy on individual level. It is important to keep in mind that this measure of individual democracy is in no way equivalent to the conceptualization of democracy on a state level (as adopted e.g. by Paxton 2002), in which case it indicates the democratic functioning of a system rather than attitudes of individuals.

Theoretical relationship between social capital and individual democracy

Social capital

Before examining the relation between social capital and democracy, the first logical step would be to investigate if it is reasonable to combine these two terms under a single concept. For instance, regarding democracy, trust functions largely only as a supporter for voluntary organization membership and in order to explain this relation, we should be able to observe a relationship between these two phenomena that underlie social capital as well. This relationship may be seen as mutually reinforcing. People who are more trusting towards others are therefore also more prone to join those organizations, whereas this membership itself generates trust between the members of organizations (Marsh 2005:594). Hence, it should be possible to observe a relation between the two components of social capital:

H1. The more trusting people are, in more organizations they belong to and conversely, in more organizations they belong to, the more trusting they are.

Social capital and participation

The main function of trust in social capital is to bolster mutual understanding and solidarity. This supports a sense of community which according to Putnam (2000:35) is firmly linked to voting as it increases the responsibility to community. In brief, people who are more trusting, also tend to care more for the community and are thus more interested in public issues, although this causal relationship might somewhat vague. The ways in which associations encourage political participation are, however, less ambiguous. Voluntary organizations, much like trust, “instill in their members habits of cooperation, solidarity and public-spiritedness”, but more importantly, they articulate and aggregate interests (Wollebaek, Selle 2002:35), which eases decision-making. Some research has already found that individuals who tend to join different organizations more often become engaged in political activities, including voting (La Due Lake, Huckfeldt 2002:579; Marsh 2005:594). Thus, the following hypothesis can be posed:

H2: The more trusting people are and in more organizations they belong to, the more likely are they to participate in democratic political activities.

Social capital and democratic attitudes

In terms of democratic attitudes, more trusting people should uphold democratic attitudes since trust essentially means being more reliant on others and willing to depend on them, which is also inherent to democracy. In addition to strengthening democratic values simply by bringing different people together (Paxton 2002:258), most voluntary organizations also “serve as schools of democracy” (Marsh 2005:594) through various democratic practices, most notably electing leadership for the organization. This should contribute to higher endorsement of democracy and thus the following hypothesis is proposed:

H3: The more trusting people are and in more organizations they belong to, the more likely are they to have positive views on democracy.

Social capital and democracy on individual and system level

As previous research has found, the relationship between social capital and democracy seems to be very much evident on the country level (Paxton 2002), while social capital seems to have little effect on individuals’ democratic behavior and attitudes (Marsh 2005). This has also been argued by Newton (2001:211). As follows, the findings in previous research imply the following hypothesis:

H4: The variation in democracy is better explained by social capital on the system, rather than individual level.

An inverse causal relationship

As the exact causal relationship between social capital and democracy is highly ambiguous, it is crucial to also recognize the limitations of the approach taken here. In contrast to the causality described above, there are number of ways in which both, individual democracy one the one side and trust and belonging to voluntary associations on the other are influenced by the same independent variables. It is plausible to assume that countries where there are better conditions for the functioning of a genuine democracy and thus more positive perception towards it, the social environment is also more beneficial to the development of social capital (Newton 2001:211). Democratic institutions strengthen not only participation in democratic processes, but also civic culture (Seligson, Booth 1993:779), which in a way is equivalent to social capital. This highlights the difference between developed and developing countries. People can be expected to be more interested in politics and democracy in countries that are more modernized (Inglehart, Welzel 2005:56), while it is in these countries that people are also likely to be more trusting as well as trustworthy and have better organized and accessible voluntary organizations.

Thus, it is of critical importance to include the differences among countries as well as different social groups to the analysis. In particular, political participation above all, but also other phenomena assessed in the current paper are very much dependent on different socio-economic indicators (Marsh 2005:595). Some previous research (Paxton 2005; Marsh 2005) on the relationship between social capital and democracy has included education, social class, income, gender, ethnicity and religiosity to the analysis as control variables that should have an effect on both, the independent and dependent variable in the theoretical relationship.

Methodology

Data

All the variables in the analysis are measured by utilizing World Values Survey data. In particular, an unofficial release of longitudinal multiple-wave dataset of the survey (WVS 2014) is used here as it allows to analyze a larger sample than any of the datasets for separate waves. The variables are defined in this section using the respective variable dictionary (WVS 2015). However, as the exact questions and their coding have not been consistent over the waves, only the results of the last two waves are used for analysis here, which limits it to years 2005-2014. Additionally, all respondents with missing values in any of the variables under investigation here were excluded from the analysis, which results in a total of 87 880 observations from 67 countries.

Operationalization of main variables

Trust is measured using four variables that indicate whether or not the respondent trusts people she knows personally (G007_33_B), people she meets the first time (G007_34_B), people of another religion (G007_35_B) and people of another nationality (G007_36_B). Measuring trust in this way might be problematic as it should evaluate how trusty a person is of others, but it can be said to indicate, rather, how trustworthy are others deemed to be (Newton 2001:203-4). Nevertheless, trust has often been measured in this subjective way (Marsh 2005:603; Newton 2001:208; Paxton 2002:260), presumably since there is no better indicator for evaluating trustiness and even the described method should express it quite accurately.

The measurement of membership in voluntary associations is rather straightforward. This is evaluated by respondents’ answers to questions if they are active or inactive members or not members at all in following voluntary organizations: church or religious organizations (A098); sport or recreational organizations (A099); art, music or educational organizations (A100); labor unions (A101); political parties (A102), environmental organizations (A103); professional organizations (A103) or charitable/humanitarian organizations (A105). Just like in case of trust, this measurement may also lack validity, since voluntary associations might not be a good indicator of social capital in developing countries where creating and maintaining strong social associations is presumably more difficult. For instance, it has been suggested that associations do not increase social capital as such in Africa (Widner, Mundt 1998:14) where social ties are likely to be organized along very different lines. I argue, however, that these lines are not capable of substituting formal networks which are in the focus of the current paper.

Participation, the first component of individual democracy, is measured by two distinct expressions of this activity: voting in elections and other political action. Voting is indicated simply by the answer to questions if the respondent voted in recent parliament elections (E257) and if the respondent always, usually or never votes in national elections (E264). Other political activity includes signing a petition (E025), joining boycotts (E026) and attending demonstrations (E027). The inclusion of these two different aspects of participation is necessitated on the one hand by the fact that voting depends highly on the democratic status of a country and is likely to be higher in more liberal societies while those activities expressing dissent are more likely to be observed in countries where different liberties are restrained. Including two different aspects of participation should then balance these deficiencies to some extent. The problem still remains that in some countries voting is compulsory, but since the number of countries in the analysis is quite high, this should not cause too much bias in the results.

The second element of democracy, democratic attitudes, are measured by respondents’ opinion on democracy. This involves questions on how well respondents’ think of having a democratic political system (E117) as well as how they would evaluate the importance of democracy (E235). This approach does not involve any democratic values, but as such, it provides a simple and unambiguous measure.

The measures for all of the four variables are created by a simple method. The original variables included to each of the four variables in the causal relationship are first scaled to an identical range and then simply the mean of these is calculated on the scale of 0 to 1. This results in measures that equally express all the variables that were used to calculate them, although the values of these variables do necessarily follow a standard normal distribution.

Control variables

As previously noted, including control variables to the analysis is essential when attempting to explain phenomena such as social capital and democracy. Thus, in addition to the four variables, another four variables expressing different socio-economic characteristics of the respondents are introduced to the analysis. These are age (X003), income (X047), social class (X045), religiosity (A006) and education (X025), all measured on an interval scale. It’s important to note that income is measured on a relative scale, reflecting differences only inside each country.The ways in which society and politics interacts is a central theme in political sociology. One of the topics that can be derived directly from this line of reasoning is concerned with links between social relations and political regime. In this respect, the research on social capital is noteworthy. Introduced to wider social science research by Robert Putnam (1993), it was a highly investigated and discussed phenomena that received considerable amount of attention, albeit only for a decade. Although his inquiry into this issue has received substantial amount of criticism (e.g. Levi 1996; Tarrow 1996), he posed an important question: do relations between individuals have an effect on the functioning of democratic institutions? Or, more precisely in the context of current paper: does social capital promote democracy?

Analysis

In this section the proposed hypotheses will be tested using previously described variables which are outlined in table 1. In case of regression analyses, standardized coefficients are reported instead of regular coefficients since this allows for better evaluation of the total magnitude of variables and all the main variables are nevertheless measured on an obscure scale and thus arbitrary to simple interpretation.

Table 1. Summary statistics for the variables included in the analysis

  Count Mean Median Standard Deviation Minimum Maximum
Trust 87880 0.456 0.417 0.207 0 1
Membership 87880 0.14 0.063 0.183 0 1
Participation 87880 0.529 0.5 0.342 0 1
Attitudes 87880 0.819 0.833 0.187 0 1
Social capital 87880 0.069 0.026 0.104 0 1
Individual democracy 87880 0.441 0.389 0.315 0 1
Education 87880 4.863 5 2.209 1 8
Income 87880 4.914 5 2.127 1 10
Social class 87880 3.24 3 0.97 1 5
Religiosity 87880 3.107 3 1.05 1 4
Age 87880 41.259 39 16.32 15 99


H1: The more trusting people are, in more organizations they belong to and conversely, in more organizations they belong to, the more trusting they are.

As the relationship between trust and voluntary organization membership (hereafter referred to as “membership”) is in theory reciprocal, it will be evaluated by simple correlation analysis. The results of this are presented in table 2 along with other variables which allows to comprehend how these relate to each other. Although the relation between the two variables is statistically significant and positive, the correlation can be considered as weak with Pearson’s correlation coefficient of just 0.123. This indicates that we have no reason to believe that individuals with higher level of trust also belong to substantially more voluntary organizations or the other way around as suggested by theory. Thus, the first hypothesis is rejected.

Table 2. Pearson’s correlation coefficient between main and selected variables.
* p < 0.05, ** p < 0.01, *** p < 0.001

  Trust Membership Participation Attitudes
Trust 1      
Membership 0.123*** 1    
Participation 0.186*** 0.245*** 1  
Attitudes 0.121*** 0.0354*** 0.109*** 1
Education 0.128*** 0.118*** 0.176*** 0.0536***
Income 0.0947*** 0.127*** 0.0747*** 0.0513***
Social class -0.106*** -0.107*** -0.0924*** -0.0425***
Religiosity -0.105*** 0.0685*** -0.177*** 0.0436***
Age 0.119*** -0.0339*** 0.0579*** 0.0436***


H2: The more trusting people are and in more organizations they belong to, the more likely are they to participate in democratic political activities.

The effect of trust and membership on participation is measured by four different models in order to assess the influence of two main independent variables separately, together and with control variables. As demonstrated in table 3, the effect of membership tends to be stronger than that of participation in all models. Not only can the influence of membership considered to be about twice as high in model 4, it also explains almost twice as much variation among participation when compared to the effect of trust. Still, the R2 of just 0.085 in case of model 3 implies that even both main independent variables combined have a very weak linear relationship with participation. Hence, while participation seems to be somewhat more dependent on membership rather than trust, both trust and membership have very little if any effect on participation, which means that the second hypothesis is rejected.

Table 3. Beta coefficients from OLS regression of participation on selected variables.
Standardized beta coefficients
|* p < 0.05, ** p < 0.01, *** p < 0.001

  1 2 3 4
  Participation Participation Participation Participation
Trust 0.186***   0.158*** 0.118***
Membership   0.245*** 0.225*** 0.228***
Education       0.119***
Income       0.001
Social class       -0.007
Religiosity       -0.160***
Age       0.057***
R2 0.035 0.06 0.085 0.131


H3: The more trusting people are and in more organizations they belong to, the more likely are they to have positive views on democracy.

To test the third hypothesis, the effect of independent variables is evaluated in the same manner as in previous case. The impact of trust and membership on positive attitudes towards democracy seem to be even lower than in case of participation as table 4 illustrates. However, here the independent variables function contrarily in respect to democratic attitudes. While the influence of membership is entirely negligible, the attitude has significantly higher beta coefficient. Yet, even in model 3 both independent variables explain only 1.5% of variation among democratic attitudes and even in model 4 none of the variables included to the analysis seem to affect democratic attitudes substantially. This means that hypothesis 4 is rejected.

Table 4. Beta coefficients from OLS regression of democratic attitudes on selected variables.
Standardized beta coefficients
* p < 0.05, ** p < 0.01, *** p < 0.001

  1 2 3 4
  Attitudes Attitudes Attitudes Attitudes
Trust 0.121***   0.119*** 0.112***
Membership   0.035*** 0.021*** 0.009**
Education       0.044***
Income       0.032***
Social class       -0.005
Religiosity       0.065***
Age       0.046***
R2 0.015 0.001 0.015 0.023


H4: The variation in individual democracy is better explained by social capital on the system, rather than individual level.

The assessment of the examined relationship on aggregate level is simplified by coding the four main variables into just two. Social capital variable is the product of trust and membership variables. Individual democracy variable is likewise obtained by multiplying participation and attitudes variables. The results of a regression analysis reported in table 5 imply again that on individual level, social capital is a relatively weak predictor of individual democracy at R2 value of just 0,074. This figure, however, alters drastically when the regression analysis is repeated after data has been aggregated by country. When observing means among countries, social capital explains as much as 36,9% of variation among individual democracy without control variables. It’s also noteworthy that the impact of none of the control variables is statistically significant anymore at such a low number of observations.

Table 5. Beta coefficients from OLS regression of individual democracy on selected variables on individual level (models 1 and 2) and on country level (models 3 and 4).
Standardized beta coefficients
* p < 0.05, ** p < 0.01, *** p < 0.001

  1 2 3 4
  Individual democracy Individual democracy Individual democracy Individual democracy
Social capital 0.271*** 0.251*** 0.608*** 0.602***
Education   0.130***   -0.125
Income   0.017***   -0.051
Social class   -0.012***   -0.119
Religiosity   -0.142***   -0.175
Age   0.080***   0.205
Observations 87880 87880 67 67
R2 0.074 0.123 0.369 0.494


The linear relationship observable in table 5 is better illustrated on graph 1. While most countries seem to have a low mean value of individual democracy as well as social capital, those with higher levels seem to follow a linear relationship, albeit with relatively high scatter. With some exceptions, it is even possible to identify two distinct regional or cultural patterns as there seem to be countries with (1) high level of individual democracy and medium to high level of social capital (established western democracies) and (2) high level of social capital, but medium level of individual democracy (sub-Saharan Africa and Southeastern Asia). In sum, there is enough evidence to confirm the fourth hypothesis.

Graph 1. Individual democracy plotted against social capital. India is excluded from the graph.

Conclusions

Even though a number of authors have examined the relationship between social capital and democracy, they have in most cases if not always remained somewhat doubtful of the relation and the exact causality behind it. It has been only on the aggregate level where some, although ambiguous, links between the two phenomena have been established. In this paper the stance was taken that the relationship should be examined on individual level as this is where it would be most relevant. Interpreting social capital as a combination of trust and voluntary organization membership and democracy on individual level as democratic participation and attitudes, the relationship between these two phenomena was then examined using quantitative analysis. The main finding of the analysis was that among people, there is almost no relation between the variables of social capital and individual democracy. While extremely moderate share of variation in participation can be explained by trust and membership, in case of democratic attitudes neither of the two social capital variables indicated almost any correlation. This suggests that in a general sense, social capital has no effect on democracy on individual level. Interestingly, however, the behavior of trust and membership is not consistent in relation to democratic participation and attitudes. Whereas participation seems to be more dependent on membership, attitudes in contrast depend more on trust. The weakness of these relationships does not permit one to draw any firm conclusions, but it nevertheless calls the relevance of social capital as a concept into question. Despite that trust and membership are positively correlated, this correlation is still weak at best. The fact that these components of social capital behave in different ways when examining their effects on different phenomena, implies that trust and membership should perhaps not be united under a single concept, but instead studied separately. Another claim that can be made from the results is that while on individual level the correlation between social capital and individual democracy is negligible, it becomes surprisingly substantial on country level. What this implies is that there seems to be some kind of relationship between individual democracy and social capital, but it is mediated or caused by a phenomenon that was not captured by any of the control variables employed in the current analysis. As the relation is only observable on country level and certain regional patterns seem to occur, there is reason to believe that the possible explanation for the relation is very likely cultural, although the exact causality still remains unclear. What the results do however confirm is that there is no direct relationship between social capital and individual democracy.

References

  1. Adler, Paul S.; Kwon, Seok-Woo. “Social capital: Prospects for a new concept.” Academy of management review 27.1 (2002): 17-40.
  2. Farr, James. “Social capital a conceptual history.” Political theory 32.1 (2004): 6-33.
  3. Inglehart, Ronald; Welzel, Christian. Modernization, cultural change, and democracy: The human development sequence (2005). Cambridge University Press.
  4. La Due Lake, Ronald, and Robert Huckfeldt. “Social capital, social networks, and political participation.” Political Psychology 19.3 (1998): 567-584.
  5. Levi, Margaret. “Social and unsocial capital: A review essay of Robert Putnam’s Making Democracy Work.” Politics & society 24.1 (1996): 45-55.
  6. Marsh, Robert M. “Social capital and democracy in a new democracy.” The sociological quarterly 46.4 (2005): 593-615.
  7. Newton, Kenneth. “Trust, social capital, civil society, and democracy.” International Political Science Review 22.2 (2001): 201-214.
  8. Paxton, Pamela. “Social capital and democracy: An interdependent relationship.” American sociological review (2002): 254-277.
  9. Putnam, Robert. Bowling alone: The collapse and revival of American community. Simon and Schuster, 2000.
  10. Putnam, Robert. Making democracy work: civic traditions in modern Italy (1993). Princeton University Press.
  11. Seligson, Mitchell A.; Booth, John A. “Political culture and regime type: Evidence from Nicaragua and Costa Rica.” The journal of politics 55.03 (1993): 777-792.
  12. Sobel, Joel. “Can we trust social capital?” Journal of economic literature 40.1 (2002): 139-154.
  13. Tarrow, Sidney. “Making social science work across space and time: A critical reflection on Robert Putnam’s Making Democracy Work.” American political science review (1996): 389-397.
  14. Temple, Jonathan; Johnson, Paul A. “Social capability and economic growth.” The Quarterly Journal of Economics 113.3 (1998): 965-990.
  15. Vanhanen, Tatu. “A new dataset for measuring democracy, 1810-1998.” Journal of Peace Research 37.2 (2000): 251-265.
  16. Widner, Jennifer; Mundt, Alexander. “Researching social capital in Africa.” Africa (1998): 1-24.
  17. Wollebaek, Dag, and Per Selle. “Does participation in voluntary associations contribute to social capital? The impact of intensity, scope, and type.” Nonprofit and Voluntary Sector Quarterly 31.1 (2002): 32-61.
  18. World Values Survey (WVS) (2014). “WVS_Longitudinal_1981-2014_stata_dta_v_2014_06_17_Beta”. http://www.worldvaluessurvey.org/WVSDocumentationWVL.jsp (25.11.2014)
  19. World Values Survey (WVS) (2015). „WVS_Values Surveys Integrated Dictionary_TimeSeries_v_2014-04-25“ http://www.worldvaluessurvey.org/WVSDocumentationWVL.jsp (12.01.2015)