155 Abstract Firstly, we investigate the structure of social capital based on several indicators selected on the basis of the European Social Survey’s 2006 wave for Romania. Secondly, the role of informal social interaction and some individual-level socio- demographical variables in determining formal, informal and community level volunteering is analyzed. The third part of the analysis investigates the role of these types of social capital in producing various forms of trust. The results of the analyses resulted in two specific variables, for informal and informal social capital: the former consists in variables accounting for informal networking with closer ties, while the latter consists in variables accounting for different kinds of voluntary activities. Regarding the cognitive dimension of social trust, we succeeded to separate specific types of trust. Once controlling for socio-demographics and religiosity, schmoozing exercises significant, positive influence only on formal volunteering. Our results suggest that the profile of those involved in the specific forms of volunteering is quite similar in at least two aspects: it seems that compared to the youngest age category, middle aged individuals are less involved in volunteering; compared to the lower educated respondents those who are medium and upper level educated people are the most involved in each types of volunteering. Keywords: formal social capital, informal social capital, volunteering, trust, European Social Survey. LINKAGES BETWEEN INFORMAL AND FORMAL SOCIAL CAPITAL AND THEIR RELATIONS WITH FORMS OF TRUST. A FOCUS ON ROMANIA* Laura NISTOR Cristina TÎRHAŞ Petru ILUŢ Laura NISTOR Lecturer, Department of Social Sciences, Faculty of Sciences Miercurea Ciuc, Sapientia University, Cluj-Napoca, Romania Cristina TÎRHAŞ Lecturer, Department of Social Sciences, Faculty of Social and Political Sciences, Avram Iancu University, Cluj-Napoca, Romania Petru ILUŢ (corresponding author) Professor, Department of Sociology, Faculty of Sociology and Social Assistance, Babes-Bolyai University, Cluj-Napoca, Romania Tel.: 0040-264-424.674 E-mail: pilut@socasis.ubbcluj.ro Transylvanian Review of Administrative Sciences, No. 34 E/2011, pp. 155-174 * This work was supported by the research program CNCSIS IDEI, contract 2572/2008. 156 1. Introduction Social capital constitutes a widely used concept in the social, economic and political sciences of nowadays and usually it is considered to be a notion describing social interaction and involvement, social trust and civil norms etc. which can be both reason, consequence and solution for a wide range of other social actions and phenomena, e.g. solidarity, democracy, economic development, civility etc. The broad concept of social capital is in fact full of distinctions, in the sense that there are many definitions within the literature which oppose different perspectives of definition and various types and forms of social capital. Such distinctions have already been present in the theoretical roots of the social capital literature. Bourdieu (1985, 1993), Coleman (1988) and Putnam (1993, 2000) who are considered the fathers of the concept had defined social capital on different levels and put forward different aspects and roles of it. Bourdieu defined social capital on the level of the individuals, considering it as individuals’ potential to mobilize their social connections in order to achieve various types of other capitals (e.g., economic, cultural capital). Coleman’s approach is directed, similarly to Bourdieu’s, towards the micro-level; however, Coleman defines social capital not so in terms of network-based resources, but by its functions. In the approach of Coleman, social capital is embedded in networks and facilitates collective actions, mainly through obligations, expectations, trustworthiness, information, norms and sanctions associated with such networks. In contrast to these micro-level assessments, Putnam theorizes the concept of social capital on the macro-level and defines it as the attribute of societies, a cultural phenomenon referring to civic participation, trust and trustworthiness. In his approach social capital can have a positive influence not only on the life of the individuals, but also on the whole society, since participation in networks and especially in formal associations, cooperation and trust resulting from participative behaviors can enhance the flourishing of democracy and civic life. The two approaches, the individual and macro-level assessment of social capital, were integrated by many scholars, especially in the case of those empirical studies which aimed at studying the determinants and consequences of social capital on both individual and country/regional level. In spite of such integrative endeavors, there is still a number of other distinctions between types and measures of social capital. Consequently, some authors define the concept as an umbrella term having various forms and dimensions and to which are associated equally various empirical indicators (e.g. van Oorschot, Arts and Gelissen, 2006). One of the major distinctions between types of social capital opposes bonding and bridging social capital, that is, strong intra-community oriented networks among similar people (e.g. family members, religious or ethnic groups) and weaker extra- community networks developed among individuals with different social status, for instance, work colleagues (Granovetter, 1973; Putnam, 2000). There exists, however, another important distinction between formal and informal types of social capital. While the former refers to participation in strictly organized networks, like civic associations, trade unions, charitable and voluntary associations, political parties 157 etc., and constitutes the major type of social capital in the work of Putnam (2000), the latter implies various forms of getting together with family members, friends, work colleagues, neighbors etc., and refers to more or less regular social interactions in the absence of a formal associational frameworks (Putnam, 2000). The importance of the informal networks was theorized especially in the works of Bourdieu (1985) and Coleman (1988), which identified social capital especially with informal connections. An additional separation occurs between the structural and cognitive dimensions of social capital (see Uphoff, 1999; Kaasa and Parts, 2008). The structural dimension refers to behaviors associated with social capital in forms of participation in both formal and informal networks, while the cognitive aspect refers to norms, beliefs and attitudes related to social capital, especially to forms of trust (general trust, institutional trust) and trustworthiness. Generally speaking, the relationship between these two dimensions of social capital is that they are mutually reinforcing and – even more important – they are so interrelated that it is very difficult to establish which one is the cause and the effect. As Putnam (2000, p. 137) assumed the causal arrows among forms of participation and social trusts “are as tangled as well-tossed spaghetti”. It is thus equally reasonable to presuppose that people become more trustworthy as a result of their sustained social interactions and that people who are participating in social interactions are more trustworthy than others. When we consider specifically the structural dimension of the social capital, that is forms of involvement, it is observable that formal and informal, bonding and bridging structures are usually treated as concurrent components and in spite of many attempts aiming at revealing the macro-level as well as micro-level linkages between them there are still relatively few articles which focus on the interrelations between various types and facets of social capital (e.g. Halman and Luijkx, 2006; van Oorschot, Arts and Gelissen, 2006). Consequently, little is known about the correlations between informal and formal participation (Palisi, 1985). In the present study we aim to investigate this kind of linkage in the case of Romania, a country which was considered in several studies an example for a weakly developed social capital, whether formal or informal (e.g. Voicu and Voicu, 2003; Pichler and Wallace, 2007). We frame our approach into the complementarity interaction theory (Fischer, 1982; Palisi, 1985; Iluţ and Tîrhaş, 2010) according to which formal and informal involvement can enhance one another and assume that these two specific forms of social capital (i.e. informal and formal) may interact even in the case of countries which show a social capital deficit, like Romania. Our analysis is built on the data set of the European Social Survey 2006. We separate between informal social interaction, informal help directed towards others than family members, community level involvement in helping activities and formal volunteering as structures of social capital. On the other hand, the cognitive dimension of social capital is measured through general trust and institutional trust, the latter separated in turn, in trust towards national and international institutions. The article develops around three main goals. Firstly, we are going to investigate the structure of the social capital based on the above mentioned indicators: that 158 is the investigation of how these different indicators associate together in forming different dimensions of social capital. Secondly, we are going to investigate the role of informal social interaction and some individual-level socio-demographical variables in determining formal, informal and community level help. In the third part of the analysis we will investigate the role of these types of social capital structures in producing various forms of trust. Before presenting the hypotheses, the methodology and the results of the analyses we briefly sketch some of the literature assumptions regarding the opposition/complementarity between formal/informal social capital, respectively their trust producing potential. 2. Formal and informal social capital. From substitution to complementary? Regarding the opposition and/or complementary of these two sub-types of social capital, i.e. formal and informal social capital, Haerpfer, Wallace and Raiser (2005) re-accentuated the role of the early views on social capital in the framing of the debate in one direction or another. Thus, their definition of informal social capital follows the approach of Bourdieu (1985) and Coleman (1988) and is based on the relational and reciprocal aspect of transactions conducted within social relationships. In contrast, formal social capital is most evident in Putnam’s (2000) approach, where social capital exists especially in forms of civic participation and thereafter social capital is complementary to formal institutions in supporting democracy and civility. Once the ground defined, a number of empirical studies tried to demonstrate the opposition or co-existence of these two types of social capital. The analyses usually were directed towards the investigation of whether formal or informal social capital is more important in producing trust, civility and trustworthiness. Putnam (2000) for instance argues that the engagement in civic organizations can enhance social capital, especially in terms of trust and social cooperation and thus, accentuates the role of the so called machers1 in social capital production. On the other hand, authors like Lin (2001) assess that participation in informal networks is more important in social capital production. There are however integrative approaches as well. Putnam (2000) himself when separates between machers (i.e. those involved in formal networks) and schmoozers2 assesses that these two types of social involvements may overlap to some extent, in spite of the fact that empirically they are rather distinct, since machers tend to be better educated, economically better off, compared to schmoozers who are more diffusely present throughout every social strata. Moreover, the dynamics of these two involvements is different: while participation in formal networks peaks in late middle age, informal involvement peaks among young adults. The relationship between formal/informal networking is specifically discussed and empirically analyzed later on by Pichler and Wallace (2007) in their European aggregate 1 Putnam (2000) uses the Yiddish term machers in order to denominate those persons who are active in formal organizations. 2 The Yiddish term describes those people who spend a lot of time for informal socialization. 159 level study. The authors followed Fukuyama’s (1995) approach and considered that formal social capital can lead to better informal networking and support, which then reinforce social norms of co-operation and trust. However, the authors noticed that this should not been considered business as usual. It happens frequently that these two forms of social capital appear rather exclusive than inclusive to each other, in the sense that strong informal networks can take the place of formal participation. Or, on the contrary, the proliferation of formal social capital in the absence of informal social capital might occur, especially in situations when in the presence of rich associative ties there is no need for informal networks (Fischer, 1975)3. Based on these considerations, the authors assumed that on the macro-level there might exist on the one hand, complementary regimes of social capital, within which both formal and informal networks are well developed. On the other hand, substitution regimes might be also present in which case one type of social capital, the most frequently formal, can be replaced by the other type of social capital. Based on a European level aggregate analysis Pichler and Wallace (2007) succeeded to bring empirical foundation to these hypotheses and confirmed that there are several social capital regimes throughout Europe, depending on the ways in which formal and informal social capital are both present and substituted. Thus, the authors showed that while especially Nordic countries and, to a lesser degree Western European countries, score high on both dimensions of social capital and thus are examples for complementary regimes, Southern countries are poorer in both kinds of social capital, while in Eastern Europe appears a substitution regime, where lower levels of formal social capital are replaced by high levels of informal social capital. Regarding the individual level, van Oorschot, Arts and Gelissen, 2006 revealed significant, positive correlations between three different facets (networks, trust and civism) of social capital. They concluded that on the level of the individuals it is legitimate to talk about an accumulation in terms of social capital, since there is a positive correlation between involvement in formal and informal networks (except the family), so that people who show these types of involvements are also those who trust others and institutions the most. The authors investigated these linkages on the level of aggregated, country-level data and concluded that there are positive, significant correlations between the formal and informal networks, in the sense that countries with higher level of participation in formal associations are also the countries in which people participate more in informal networks, particularly friends- based networks, so that it seems that people are becoming formal members via their 3 The so-called “crowding out” hypothesis constitutes a more radical approach which can be however quoted here for the purpose of analogy building: this approach assumes that social expenditures of the welfare state made informal involvement, caring relations and civic involvement useless and, consequently assumes that in the most developed welfare states there must be the lowest levels of involvement in both formal and informal networks (van Oorschot and Arts, 2005). Consequently, informal and formal social capitals may substitute or complement themselves, but they can be both eroded under certain contexts. 160 friends. Moreover, there is positive, significant correlation between participation in networks, both formal and informal, and generalized as well as institutional trust. This finding is in accordance not only with Putnam’s (1993) approach according to which participation enhances the flourishing of trust, but correlates also with the findings of Pichler and Wallace (2007). Regarding the linkages between informal and formal social capital, Kenny (1992), Huckfeldt (1979), Mutz (2002), etc. succeeded to confirm that people usually decide about formal political participation via their friends. The approach directed towards the investigation of the roles of informal social ties on volunteering (e.g. Smith, 1994) is also extremely rich. These studies put forward the so-called “contact frequency hypothesis” (see Wilson and Musick, 1997) which presupposes that the more frequent involvement in informal social networks increases the likelihood of formal volunteering. The already classical study of Wilson and Musick (1997) confirmed that volunteering is connected to individuals’ informal social networks and is often facilitated by these networks in such a great manner that in America social interaction can be considered as one of the most important predictors of volunteering. The explanation is that in most cases people end up as volunteers because they are asked to volunteer for an organization and a regular involvement in informal social ties enhances the chance of being asked for volunteering. There were revealed some nuances as well which must be considered when reflecting on the relationship between informal involvement and formal volunteering. Voicu and Voicu (2003) suggested that when considering the linkages between formal volunteering and informal social interaction, one should consider the bonding vs. bridging natures of the social interactions under discussion. Thus, while bridging social ties (e.g. colleagues) enhance volunteering, bonding social capital consisting in strong informal ties developed among peers are not expected to enhance formal volunteering, since these networks direct volunteering towards informal help. Wilson and Musick (1997) while signaled positive association between the frequency of involvement in informal social interactions and volunteering, demonstrated also that (at least in the context of the USA) the relationship between formal volunteering and informal volunteering (i.e. helping) was not mutually beneficial, in the sense that an increase in volunteering increased helping, but the reverse could not been proved. They also demonstrated that social ties are not linked to informal volunteering considered in the form of helping a neighbor. The authors’ explanation was that helping is probably perceived as an obligatory activity compared to formal volunteering and thus can be much more linked to the norm of obligation. In contrast, volunteering is much more rooted in altruism. Regarding the roots of trust, that is whether formal or informal social involvements contribute more to trust, Delhey and Newton (2003) put in opposition the theory of voluntary organizations and that of network theory. In accordance with the voluntary organizations theory the most important form of participation is direct, face-to-face involvement in voluntary groups in the local community, since in these associations 161 people learn to collaborate, develop empathy and reciprocity. In contrast, the network theory assumes that trust is produced mainly by involvement in informal networks. After summarizing a number of research findings, the authors considered that this latter theory should be particularly relevant in the case of post-communist societies in which during the period of communism people developed important circles of involvement in private, unofficial networks through which citizens succeeded to help each other in overcoming the shortage of resources and also created isles of trust and reciprocity “within a wider society that was pervaded by general suspicion and mistrust created by the state” (Delhey and Newton, 2003, p. 98). Li et al. (2005) concluded however that the relationship between formal associational membership and social trust is rather contradictory and that “it seems unlikely that social capital defined as associational involvement will be particularly powerful in generating outcomes such as social trust” (Li et al., 2005, p. 111). Similarly, Frane (2008) by analyzing European level longitudinal data showed that contrary to Putnam’s assumption according to which there is a connection between trust and formal participation, in Europe could not be signaled such a trend and occurs the fact that “while the reduction of generalized trust is quite visible – thought not in all countries – associational involvement shows a more complex but an average positive trend” (Frane, 2008, p. 167). In the post-communist societies, at least in Romania, quantitative, survey-based as well as qualitative-ethnographic data show that there is an extremely low level of trust, both in terms of general social trust and institutional trust. Moreover, it can be outlined that compared to the first years of the transition which were saturated by the enthusiasm of freedom, individuals’ trust has been following a down warding tendency. If we add to these data also those which show that according to a great part of the respondents life was better under socialism, in our opinion the low levels of institutional trust existing today can have two important roots. One of them consists in the disillusion with neoliberal capitalism, while the other might be located in the existing corruption on the level of the institutions. It is thus not surprising that the most trusted institutions by the Romanians are those which do not have a direct tangency with either capitalism, or they were not involved in corruption-scandals (e.g. the army and the church). The fact that market economy and neoliberalism have brought with themselves competition and struggle for resources in nearly every domains of life resulted also in the diminishing of interpersonal trust. On the other hand, in order to face competition and generalized corruption, people have associated in informal networks4. 4 Our theoretical point (which is based also on on-going qualitative research and which is not the subject of the present article) is that such informal networks are trustworthy networks in the sense that those involved in the network try to protect the interests of the network. Moreover, trust is not necessarily the result of the involvement, but mainly the root of the involvement. In the majority of cases, these networks are family-based, a fact which is especially visible in the case of the small urban and rural settings. It is also worth noting that such kinds of networks are very closed and difficult to access by researchers. 162 3. Data, hypotheses and methodology For the purpose of the present analysis we are relaying on the data of the European Social Survey’s (ESS) third (2006) round5. On the basis of the items of this ESS round questionnaire we selected several indicators of social capital which are presented in Table 1 together with their original types of measurement6. The first column of the table indicates the types of social capital we hypothesized and intended to measure. We presupposed that informal social capital is tapped through four items which account for: the frequency of socialization with friends, colleagues and relatives (Q1); the perceived existence of moral support for discussing intimate problems (Q2); the perceived frequency of social involvement in networks compared to others (Q3) and through the item which accounts for involvement in helping behaviors towards others than close network members (Q4). We hypothesized also that this latter type of involvement can be considered a form of informal bridging social capital, since it is explicitly focused on the informal help provided towards others than family members. It is however explicitly differentiated from formal volunteering as well (Q6). We presumed that involvement in community level help and activities (Q5) can be situated somewhere “in between” informal helping and formal volunteering. We appreciated that this question does not refer specifically, by definition, to either formal or informal social capital, and respondents may indicate their participation in both informal community activities (e.g. the cleaning of the green space by a group of neighbors) and formal activities (e.g. participation in activities organized by community organizations, e.g. by the association of flat inhabitants, i.e. asociaţia de locatari in Romanian)7. On the basis of these variables we succeeded to separate not only between informal social interaction and volunteering, but also between types of volunteering, i.e. formal and informal volunteering. According to Wilson and Musick (1997) helping behavior performed outside an organizational framework can be considered informal volunteering (or caring, if it is performed within the family) because it constitutes, similarly to volunteering, un-paid work. It is worth noting that while authors like Amato (1990) consider that the main difference between informal and formal volunteering consists in their different degree of formalization, Wilson and Musick (1997), Burr et al. (2005) etc. appreciate that the main difference between the two resides in their different motivational bases: while obligations have a powerful influence on informal help, volunteering is much more impacted by altruism. These rationales allowed us to put forward the question whether informal helping is more related to informal social 5 The data set was downloaded from the website of the Norwegian Social Science Data Services via de website of the European Social Survey (www.europeansocialsurvey.org) 6 In some cases it was necessary to recode the response variants, as far as we followed the logic that higher codes must indicate the strengthening or growing of the social capital facet under discussion. 7 In the case of this item, the ESS questionnaire explicitly asks respondents to associate this type of volunteering with any kind of activities they consider relevant in this respect. 163 interaction (based on its informal character) or to formal volunteering (based on its un- paid nature). Additionally, through the inclusion of the indicator for community-level help, becomes possible to analyze a different facet of volunteering, which according to Li et al. (2005) constitutes an involvement in a contextual, situational network. Regarding the cognitive dimension of social capital, similarly to the analysis of Halman and Luijkx (2006) which was based on an earlier version of the ESS, we relied on three specific items (Q7 – Q9) measuring the degree of trust in other people. Institutional trust was measured through seven items which asked respondents to indicate their confidence in specific national and international institutions. Table 1: Hypothesized types of social capital and their indicators (ESS, 2006) Type of social capital Questions Type of measurement Descriptives. Standard deviations in parentheses Informal How often do you meet socially with friends, relatives or work colleagues? (Q1) 1–7 ordinal scale (1 = never to 7 = every day) Mean = 4.09 (1.79) Do you have anyone with whom to discuss intimate and personal matters? (Q2) Dummy (0 = no, 1 = yes) Yes=70%; No=30% Compared to other people of your age, how often would you say you take part in social activities? (Q3) 1–5 ordinal scale (1 = much less than most to 5 = much more than the most) Mean = 2.50 (0.90) Not counting anything you do for your family, in your work, or within voluntary organizations, how often, in the past 12 months did you actively provide help for other people? (Q4) 1–6 ordinal scale (1 = never to 6 = at least once a week) Mean = 2.28 (1.48) Informal/ formal And in the past 12 months, how often did you help with or attend activities organized in your local area? (Q5) 1–6 ordinal scale (1 = never to 6 = at least once a week) Mean = 2.74 (1.59) Formal In the past 12 months, how often did you get involved in work for voluntary or charitable organizations? (Q6) 1–6 ordinal scale (1 = never to 6 = at least once a week) Mean = 1.64 (1.48) General trust Would you say that most people can be trusted, or that you can’t be too careful in dealing with people? (Q7) 0–10 ordinal scale (0 = you can’t be too careful to 10 = most people can be trusted) Mean = 4.07 (2.61) Do you think that most people try to take advantage of you if they got the chance, or would they try to be fair? (Q8) 0–10 ordinal scale (0 = most try to take advantage to 10 = most try to be fair) Mean = 3.81(2.48) Would you say that most of time people try to be helpful or that they are mostly looking for themselves? (Q9) 0–10 ordinal scale (0 = mostly looking for themselves to 10 = most of time try to be helpful) Mean = 3.36(2.44) Institutional trust How much do you personally trust each of the institutions? (Q10 – Q16) 0 – 10 ordinal scale (0 = no trust at all to 10 = complete trust) Country’s parliament Mean = 3.42 (2.57) Legal system Mean = 3.88 (2.69) Police Mean = 4.56 (2.83) Politicians Mean = 2.72 (2.35) Parties Mean = 2.80 (2.36) European Parliament Mean = 6.00 (2.77) United Nations Mean = 6.03 (2.85) 164 We hypothesized a positive linkage between these types of social capital, and specifically expected for a strong association among those informal and formal types of social capital which are very specific to each other. Thus, we assumed that besides the positive association between informal social interaction and different forms of volunteering, formal volunteering will be the most strongly linked to informal helping behavior, as much as this latter type of help in our case does not consider helping under its family-oriented aspect and which, according to Wilson and Musick (1997) should have different motivations than formal helping. As far as both informal help and formal volunteering in our case are oriented towards others than close network members, we presupposed strong association between the two variables. Given its quasi-formal character, we hypothesized that community-level help will be positively linked to both kinds of volunteering, as well as to informal social interaction. Regarding the roots of trust, we hypothesized the existence of some kind of specificity in this sense, that is, we expected that general trust – which can be regarded as a more diffuse form of trust – will have a stronger association with informal types of social capital than with formal volunteering, while in the case of institutional trust we expected the reversed situation. Besides investigating the structuring of these specific dimensions of social capital and the association between them, we were also interested in their most important socio-demographical determinants as well. We presupposed that if on the level of the socio-demographical variables we can reveal differential effects, then it is legitimate to talk about separate roots of these facets and, consequently, about the inexistence of linkages on the individual level. On the contrary, if similar socio-demographical roots are evident, than it is possible to reveal the linkages between these forms of social capital on these levels as well. In order to determine which are the socio-demographical roots of the types of social capital we relied on the most extensively used socio-demographical variables of the social capital literature: age (categorical variable: lower than 35, 36-55, 56+)8, gender (dummy: 0 = male, female)9, education (categorical: lower, medium, upper)10; income (ordinal: subjective rating of the household income on a four level scale)11; family size (continuous); type of residence (category: village, town, city)12. Additionally we introduced the variable of religiosity13, since it is frequently quoted as an important predictor of social capital. 8 The corresponding relative frequencies are the following: lower than 35 years = 33%; 36- 55 years = 31%; 56 years and older = 36%. 9 Female = 52% 10 We collapsed the original 8 level ordinal variable in the following categories with corresponding relative frequencies: lower educated = 34%; medium level educated = 46%; upper level educated = 20%. 11 The corresponding categories and relative frequencies are the following: very difficult to make ends meet = 24%; difficult = 38%; coping = 30%; living comfortably = 8%. 12 The corresponding relative frequencies are the following: village = 46%; town = 23%; city = 31%. 13 Determined as factorial score based on the judgment of personal degree of religiosity, private and public space religious practice. The resulted factor explains 70% of the original variance. 165 The influence of these variables on various facets of social capital has been well documented empirically, both in international and national studies (e.g. Wilson and Musick, 1997; Delhey and Newton, 2003; Voicu and Voicu, 2003; Halman and Luijkx, 2006; van Oorschot, Arts and Gelissen, 2006; Kaasa and Parts, 2008) and however the results are sometimes contradictory there are obviously some evident trends of influence. In spite of such evidence, we adopt here a rather inductive perspective, aiming at revealing those variables which have similar influence on both kinds of social capital and trust in our data. Our analysis proceeds as follows. Firstly, through exploratory factor analysis we are going to investigate the structure of social capital, based on the indicators presented in Table 1. Then, in order to better reveal the relations between various indicators of social capital we present the results of the bi-variate correlations. Thirdly, we run several regression analyses in which the dependent variables are represented by the three different specific indicators of informal, formal and semi-formal volunteering. The purpose of these analyses is to reveal which are those socio-demographical variables, if any, which contribute to each of these indicators and thus link the various types of help together, respectively, what is the contribution of informal socialization to the three forms of volunteering. In the followings, the forms of trust are going to be dependent variables of regression analyses and the aim will be to reveal what is the contribution of the several types of social capital structures to trust producing. Finally, results of a SEM analysis are reported in order to facilitate the visualization of the relationship between various forms of social capital and their socio-demographic determinants. 4. Analyses 4.1. The structuring of social capital indicators in forms of social capital In order to reveal how the selected indicators are structuring in different specific forms of social capital we used exploratory factor analysis; the results are presented in Table 2. The results of the factor analysis suggest that the considered indicators explain about 70% of the original variance and can be placed along five different, specific components, which do not replicate, however, our presupposition regarding the informal/formal types of social capital as presented in Table 1. Informal helping instead of saturating the component of informal socialization comes together with the other two indicators of volunteering. This aspect is revealed in Table 3 as well which, among others, presents the bi-variate correlations between the factor score of informal social networking and forms of volunteering and whose results show that while the correlation between informal social networking and informal volunteering is extremely low, the associations between informal volunteering and the other two types of volunteering are very consistent. Consequently, we have two very specific variables, for informal and formal social capital: the former consists entirely in variables accounting for informal networking 166 Table 2: Factor analysis of indicators of social capital Rotated component matrix Components 1 2 3 4 5 Most people can be trusted 0.844 Most people try to take advantage of you 0.890 Most of the time people helpful 0.861 Trust in country’s parliament 0.814 Trust in the legal system 0.791 Trust in the police 0.704 Trust in politicians 0.889 Trust in political parties 0.870 Trust in the European Parliament 0.871 Trust in the United Nations 0.881 How often socially meet with friends… 0.830 Take part in social activities compared to others 0.786 Anyone to discuss intimate problems 0.604 Involved in formal volunteering 0.736 Involved in helping others 0.854 Involved in community help 0.786 % of variance explained 23% 14% 12% 12% 10% KMO=0.781; Chi-Square=10768;p<0.001 Extraction method: Principal component analysis Rotation method: Varimax rotation with Kaiser nominalization Source: ESS, 2006 Table 3: Bi-variate correlations between social capital types and dimensions Types of involvement Types of trust Informal socialization Informal helping Formal volunteering Community- level helping General trust Institutional trust (national) Institutional trust (international) Informal socialization _ Informal helping 0.059** _ Formal volunteering 0.077** 0.476*** _ Community- level helping ns 0.506*** 0.313*** _ General trust 0.063* ns 0.058** 0.055* – Institutional trust (national) ns -0.072** ns ns 0.284*** – Institutional trust (international) 0.061* 0.073* ns ns ns ns _ ***p<0.001; **p<0.01; *p<0.05; ns=not significant correlation Source: ESS, 2006 167 with closer ties, while the latter consists in variables accounting for different kinds of voluntary activities oriented towards situational networks (i.e. community) or more departed others. We appreciate that these results are a proof not necessary for the split between informal and formal types of social capital, but for the role of specificity in structuring together informal and formal types of social capital. Thus, in the case of informal help, the activity towards which the indicator is oriented (un-paid helping) seems to have a more important role in the structuring of the variable together with the other variables accounting for volunteering, than the role of the type of help (informal) which is provided. Moreover, as already mentioned in the previous section, we have to give attention towards the bonding/bridging divide as well: informal socialization in our case measures a kind of bonding type of socialization, since it accounts for networking with close others, while the question about the involvement in informal helping is explicitly oriented towards others than close network members (Voicu and Voicu, 2003). Regarding the cognitive dimension, i.e. social trust, it is also obvious the separation of specific types of trust: general social trust, respectively two types of institutional trust, measuring the confidence in national, respectively international institutions. Table 3 further illustrates the associations between forms of trust and the structural dimensions of social capital and it shows that while general social trust is significantly and positively associated with the confidence in national institutions, general trust – interestingly – has no significant relationship with trust towards international institutions. Thus, it seems that the trust towards international institutions is not very much rooted in the general social trust, but, presumably in other cognitive dimensions like, for instance, aspirations, optimism, political efficacy etc., as much as the survey was realized in 2006 when the expectations from the European Integration were extremely high in Romania (Tufiş, 2007). General social trust is weakly, albeit significantly associated with schmoozing, formal volunteering and community level helping, but – interestingly – it has no significant correlation with informal helping. We can speculate here that this is quite logical as far as general social trust is more abstract and is directed towards trans- individual, less formal contexts, while informal helping has a more specific meaning and it is more linked to the personal characteristics of those involved in the informal networks than to their abstract confidence (Iluţ, 2009). Consequently, the hypothesis according to which social involvement and trust are interrelated can be confirmed only partially as much as institutional trust (whether we refer to national or international institutions) is even more ambivalent regarding its associations with the structural forms of social capital. In the followings we are turning to investigate further the interdependence between schmoozing and forms of volunteering. For this purpose we run three binary regression analyses in which the dependent variables are going to be represented by the three forms of volunteering (we created binary categories by collapsing the different frequencies of involvement to the value of 1 = involvement in various degrees vs. 0 = no involvement 168 at all). Besides schmoozing, we introduced among the independent variables also the socio-demographical variables and that of the religiosity. In this way we succeeded to measure the impact of schmoozing while controlling for socio-demographics and also to determine if socio-demographical variables exercise differential influence on the three dependent variables (Table 4). Table 4: Results of the binary logistic regression analyses Informal help Formal volunteering Community-level help B Exp (B) in parentheses Wald B Exp (B) in parentheses Wald B Exp (B) in parentheses Wald Age (reference: lower than 35) 36-55 0.260* (1.297) 4.039 0.392** (1.480) 8.081 0.566*** (1.761) 13.943 56+ 0.196 (1.216) 2.003 0.250 (1.284) 2.703 0.765*** (2.149) 21.157 Female 0.054 (1.056) 0.262 0.113 (1.120) 1.054 -0.086 (0.918) .498 Education (reference: lower level) Medium 0.459*** (1.583) 13.963 0.337** (1.401) 6.408 0.559*** (1.749) 15.621 Upper 0.860*** (2.364) 26.940 0.718*** (2.050) 18.501 0.698*** (2.010) 13.923 Income (reference: very diffi cult) Diffi cult 0.375** (1.455) 8.177 0.408** (1.504) 7.731 0.027 (1.027) 0.031 Coping 0.404** (1.497) 8.006 0.503*** (1.654) 10.384 0.076 (1.079) 0.207 Living comfortable 0.331* (1.393) 2.292 0.639*** (1.894) 7.996 0.162 (1.176) 0.401 Members of household 0.069* (1.072) 3.265 0.014 (1.104) 0.116 0.119** (1.127) 6.970 Residence (reference: village) Town -0.100 (0.905) .599 -0.053 (0.949) 0.149 -0.370* (0.691) 6.112 City 0.020 (1.020) .024 0.045 (1.046) 0.115 -0.384** (0.681) 6.874 Religiosity 0.118* (1.125) 4.831 0.207* (1.230) 12.967 0.133* (1.142) 4.775 Frequency of socialization 0.070 (1.054) 1.435 0.167** (1.181) 8.005 0.080 (1.084) 1.565 Constant -0.624** (0.536) 7.774 -1.595** (0.203) 42.380 0.124 (1.132) 0.241 Nagelkerke R Square: 0.050 Chi Square: 64.48/df 13; p<0.001 Valid N=1742 Nagelkerke R Square: 0.060 Chi Square: 73.78/df 13; p<0.001 Valid N=1736 Nagelkerke R Square: 0.056 Chi Square: 62.83/df 13; p<0.001 Valid N=1735 The results show that while controlling for socio-demographics and religiosity, schmoozing exercises significant, positive influence only on formal volunteering and thus confirms the already classical finding of the literature that informal social 169 networks are important sources for involvement in volunteering (Smith, 1994; Wilson and Musick, 1997; Putnam, 2000). Of course, the investigation of how exactly informal social networks become channels towards formal volunteering would need a deeper, preferably qualitative analysis which could provide further nuance. Interestingly, schmoozing has no significant influence on either informal help or community level help. We think that this result can be explained by the already articulated issue, according to which informal help towards others than close network members illustrates the bridging side of social capital which thus becomes less compatible with the socialization with close ties. We presuppose that a different kind of informal help (i.e. caring for family members) would be more strongly associated with schmoozing. Regarding socio-demographical variables, our results signal that the profile of those involved in these three specific forms of volunteering is quite similar in at least two aspects. It seems that compared to the youngest age category, middle aged individuals are less involved in volunteering; compared to the lower educated respondents those who are medium and upper level educated people are the most involved in each types of volunteering. While the influence from the part of the education does not constitutes a surprise, since it corresponds to the results of previous international and national level studies, the effect of age is much more curious. It seems, that contrary to previous findings reported in the case of Romania (Voicu and Voicu, 2003; Bădescu and Sum, 2005) and which signaled negative relationship between age and volunteering, we can report the reverse, which is in accordance with the findings of Putnam (2000) regarding the USA or that of Kaasa and Parts (2008) regarding Europe. Religious practice seems to be another linking factor: albeit its influence is not extremely strong, religiosity impacts significantly the involvement in forms of volunteering and thus our results confirm previous findings mentioned above, particularly that of Voicu and Voicu (2003) regarding the case of Romania. If we look at the influence of the different independent variables it becomes obvious that informal help and formal volunteering are more similar to each other than to community level help. Those involved in formal and informal volunteering seem to be middle aged individuals, while in the case of community level help the oldest age category has the highest chance for involvement, while perceived economic status has no impact in this latter regard. Moreover, while residence has no significant influence on informal and formal volunteering, residence in towns and cities, significantly reduces the chance of involvement in community level help (Kaasa and Parts, 2008). This aspect seems to confirm those previous results of the literature which accentuate the negative influence of the urban way of life on community-level solidarity. All in all, we can conclude that both general, informal help and formal volunteering have the highest chance to occur among better educated, economically better off and relatively religious middle aged individuals and community level help has the highest chance to occur among old and religious people from rural settings. 170 4.2. The roots of trust In the followings we conducted OLS regression analysis with the three forms of trust (factor scores) as dependent variables, while independent variables were represented, besides socio-demographical variables and religiosity by the factor scores of informal social networking and forms of volunteering (Table 5). Table 5: Results of OLS regression analyzes Generalized trust Institutional trust(national institutions) Institutional trust (international institutions) Age: 36-55 0.125* -0.027 -0.106 55+ 0.132* 0.087 -0.089 Female -0.013 0.001 -0.124+ Education: Medium level -0.087 -0.063 0.203+ Upper level -0.016 -0.188* 0.141+ Income: Diffi cult -0.047* -0.255** -0.077 Coping 0.118* 0.067 -0.044 Comfortable 0.067 0.051 -0.074 Members of household 0.047 0.006 0.001 Residence: Town -0.127* -0.224** 0.010 City -0.202** -0.465*** -0.146* Religiosity 0.068* 0.056* 0.039 Informal social interaction 0.047 0.008 0.018 Forms of volunteering 0.003 0.003 0.088* Constant 0.301** 0.311** -0.020 R Square=0.05 F(16,1575)=4.74; p<0.001 Valid N=1278 R Square=0.082 F(16,1275)=7.162; p<0.001 Valid N=1278 R Square=0.036 F(16,1275)=2.122; p<0.01 Valid N=1278 The low values of the R squares indicate that the set of the variables selected are not very suitable for configuring adequately the roots of trust and, compared to the structural dimension of social capital, this cognitive dimension can be less clearly modeled (Bădescu, 2001; Kaasa and Parts, 2008). Generally speaking, it can be concluded that older people and rural residents have higher levels of general trust and the urban residence goes together with lower levels of institutional trust as well (cf. the results regarding the determinants of community level help), in accordance with previous findings. Confidence in institutions is very difficult to be modeled based on socio-demographical variables, in any case, it seems that compared to lower educated respondents, medium and upper educated people are significantly less confident in national institutions and the situation is (at least marginally) the reverse in the case of the confidence in international institutions, so that our results replicate quite well the findings of Tufiş (2007) regarding the roots of institutional trust in Romania on the basis of the European Values Survey data. 171 In the light of the results based on the bi-variate correlations (Table 2), the fact that in the case of the regression analyses informal social interaction and the factor score based on the three forms of volunteering do not have a significant influence on the forms of trust (except the marginally significant influence of volunteering on the confidence in international institutions) does not constitute a surprise. It seems that in the case of our data neither general social trust, nor confidence in national institutions can be explained in terms of social interaction, whether formal or informal. 5. Instead of discussion: results of the path analysis In the followings, the above results of the regression analyses are presented in an integrated form in Figure 1, which includes all the significant influences from the part of socio-demographics on forms of social capital, respectively that of the volunteering and informal social interaction on forms of trust. For the sake of visibility, we skipped from the graphic the variables of religiosity and that of number of household members, which did not have significant influence (neither direct nor indirect) on any type and dimension of social capital when all the variables were integrated into the model. Moreover, the figure does not present those arrows of influence which were not significant. Source: Authors’ calculations Figure 1: Path analysis: determinants of the different structural and cognitive dimensions of social capital (CFI=0.922; RMSEA=0.050) 172 The path analysis suggests, once again, that the direct, significant effects from the part of the socio-demographical variables on the different latent dimensions of social capital are that of the age on both informal social interaction, forms of volunteering and general trust; that of the socioeconomic status (corresponding to the mean value of education and subjective income) on forms of volunteering and institutional trust; that of the residence on social interaction, general trust and institutional trust. Gender, while in the regression analyses turned to have no significant direct impact on any form of volunteering, now seems to have an indirect influence (through socioeconomic status) on forms of volunteering and confidence in national institutions. The latent variable of social interaction has a significant positive influence on the forms of volunteering and on general trust and, consequently, we are able to conclude that after all, informal social networking can be linked to the construct of volunteering which comprises the three, specific forms of volunteering. On the other hand, the confidence in national institutions seems to moderate between general trust and confidence in international institutions. As a general conclusion it can be deduced, that once the structural and cognitive forms of social capital are not considered as separate indicators, but in the form of groups of variables based on the factor loadings of the different indicators, the results seem to confirm the role of schmoozing on forms of involvement which is, thus, an important prove for the informal roots of the more formalized types of social capital, i.e. volunteering. In Romania, as well as in other post-communist countries, survey data show the diminishing social trust, both under its general and institutional form. For the first case the reason lies in corruption, in institutional un-accountability etc. In the same time, people develop trust in informal support networks. The association within family networks, quasi-legal or illegal networks of support etc. serve thus as a complementary regime for the eroded trust in institutions and low levels of participation in formal networks and frequently illustrate the dark side of social capital, a phenomenon which is otherwise very hard to be studied through quantitative data (Iluţ, 2009). Our present data confirm that informal networks constitute a form of compensation for the less flourishing formal networks, but informal networks are, in the same time and under certain conditions, corridors for accessing formal networks. References: 1. Amato, P., ‘Personality and Social Network Involvement as Predictors of Helping Behavior in Everyday Life’, 1990, Social Psychology Quarterly, vol. 53, no. 1, pp. 31-43. 2. Bădescu, G., Participare politică şi capital social în România, Cluj-Napoca: Editura Accent, 2001. 3. Bădescu, G. and Sum, P.E., ‘Historical Legacies, Social Capital and Civil Society: Comparing Romania on a Regional Level’, 2005, Europe – Asia Studies, vol. 57, no. 1, pp. 117-133. 4. Bourdieu, P., ‘The Forms of Capital’, in Richardson, J.G. (ed.), Handbook of Theory and Research for the Sociology of Education, New York: Greenwood, 1985, pp. 214-258. 173 5. Bourdieu, P., Sociology in Question, London: Sage Publications, 1993. 6. Burr, J.A., Choi, N.G., Mutchler, J.E. and Caro, F.G., ‘Caregiving and Volunteering: Are Private and Public Helping Behaviors Linked?’, 2005, The Journals of Gerontology, vol. 60, no. 5, pp. S247-S256. 7. Coleman, J., ‘Social Capital in the Creation of Human Capital’, 1988, American Journal of Sociology, vol. 94 (Supplement), pp. S95-S120. 8. Delhey, J. and Newton, K., ‘Who Trusts? The Origins of Social Trust in Seven Societies’, 2003, European Societies, vol. 5, no. 2, pp. 93-137. 9. European Social Survey, ESS Round 3: Round 3 Data. Data file edition 3.3, 2006, [Online] available at http://ess.nsd.uib.no/ess/round3/, accessed on July 26, 2011. 10. Norwegian Social Science Data Services, Norway, 2006, [Online] available at http:// essedunet.nsd.uib.no/cms/data/statdata.html, accessed on July 26, 2011. 11. Fischer, C., The Urban Experience, New York: Harcourt Brace Jovanovich, 1975. 12. Frane, A., ‘Mapping Social Capital across Europe: Findings, Trends and Methodological Shortcomings of Cross-national Surveys’, 2008, Social Science Information, vol. 47, no. 2, pp. 159-186. 13. Fukuyama, F., Trust: The Social Virtues and the Creation of Prosperity, New York: Free Press, 1995. 14. Granovetter, M.S., ‘The Strength of Weak Ties’, 1973, American Journal of Sociology, vol. 78, no. 6, pp. 1360-1380. 15. Haerpfer, C.M., Wallace, C. and Raiser, M., ‘Social Capital and Economic Performance in Post-communist Societies’, in Koniordos, S.M. (ed.), Networks, Trust and Social Capital: Theoretical and Empirical Investigations from Europe, Aldershot: Ashgate, 2005, pp. 243-278. 16. Halman, L. and Luijkx, R., ‘Social Capital in Contemporary Europe: Evidence from the European Social Survey’, 2006, Portuguese Journal of Social Science, vol. 5, no. 1, pp. 65-90. 17. Huckfeldt, R., ‘Political Participation and the Neighborhood Social Context’, 1979, American Journal of Political Science, vol. 23, no. 3, pp. 579-592. 18. Iluţ, P., Psihologie socială şi sociopsihologie. Teme recurente şi noi viziuni, Iaşi: Polirom, 2009. 19. Iluţ, P. and Tîrhaş, C., ‘Capitalul social informal şi formal. Modalităţi şi procese de întrepătrundere’, 2010, Sociologie Românească, vol. VIII, no. 2, pp. 145-160. 20. Kaasa, A. and Parts, E., ‘Individual-level Determinants of Social Capital in Europe’, 2008, Acta Sociologica, vol. 51, no. 2, pp. 145-168. 21. Kenny, C., ‘Political Participation and Effects from the Social Environment’, 1992, American Journal of Political Science, vol. 36, no. 1, pp. 259-267. 22. Li, Y., Pickles, A. and Savage, M., ‘Social Capital and Social Trust in Britain’, 2005, European Sociological Review, vol. 21, no. 2, pp. 109-123. 23. Lin, N., Social Capital, Cambridge: Cambridge University Press, 2001. 24. Mutz, D.C., ‘The Consequences of Cross-cutting Networks for Political Participation’, 2002, American Journal of Political Science, vol. 46, no. 4, pp. 838-855. 25. Palisi, B.J., ‘Formal and Informal Participation in Urban Areas’, 1985, The Journal of Social Psychology, vol. 125, no. 4, pp. 429-447. 174 26. Pichler, F. and Wallace, C., ‘Patterns of Formal and Informal Social Capital in Europe’, 2007, European Sociological Review, vol. 23, no. 4, pp. 423-435. 27. Putnam, R.D., Leonardi, R. and Nanetti, R.Y., Making Democracy Work. Civic Traditions in Modern Italy, Princeton: Princeton University Press, 1993. 28. Putnam, R.D., Bowling Alone: The Collapse and Revival of American Community, New York: Simon and Schuster, 2000. 29. Smith, D.H., ‘Determinants of Voluntary Association Participation and Volunteering: A Literature Review’, 1994, Nonprofit and Voluntary Sector Quarterly, vol. 23, no. 3, pp. 243-263. 30. Tufiş, C., ‘Încrederea instituţională – victimă a tranziţiei postcomuniste’, in Voicu and Voicu (eds.), Valori ale românilor 1993 – 2006, Iaşi: Institutul European, 2007, pp. 117-149. 31. Uphoff, N., ‘Understanding Social Capital: Learning from the Analysis and Experiences of Participation’, in Seregeldin, I. and Dasgupta, P. (eds), Social Capital: A Multifaceted Perspective, Washington DC: World Bank, 1999, pp. 215-253. 32. van Oorschot, W. and Arts, W., ‘The Social Capital of European Welfare States: The Crowding Out Hypothesis Revisited’, 2005, Journal of European Social Policy, vol. 15, no. 1, pp. 5-26. 33. van Oorschot, W., Arts, W. and Gelissen, J., ‘Social Capital in Europe: Measurement and Social and Regional Distribution of a Multifaceted Phenomenon’, 2006, Acta Sociologica, vol. 49, no. 2, pp. 149-167. 34. Voicu, M. and Voicu, B., ‘Volunteering in Romania: A rara avis’, in Dekker, P. and Halman, L. (eds.), The Values of Volunteering. Cross-Cultural Perspectives, New York: Kluwer Academic, 2003, pp. 143-160. 35. Wilson, J. and Musick, M., ‘Who Cares? Toward and Integrated Theory of Volunteer Work’, 1997, American Sociological Review, vol. 62, no. 5, pp. 694-713.