ABSTRACT
The process generally referred to as ‘the transition’ implied a complex set of social changes in Central and Eastern European countries (CEE). On the one hand, there were institutional reforms which were launched to abolish the socialist economic and political systems. On the other hand, the provision of welfare was also affected by ‘transition’. The change could be observed at the institutional level (i.e., through the welfare state) as well as at the social and individual level – in the ways individuals associated, how they perceived new risks and with whom they engaged to confront them. In this article we focus on social support provision from informal sources, i.e., ego-centered social support networks, which are an important foundation for the quality of everyday life. We would like to find out whether changes in the socio-economic system are reflected in and accompanied by changes in social support providers. How have people responded to transition? Did the patterns of support and social support providers change in this period? Consequently, the research question addressed in this article is the following: do the current types of social networks differ from those in the 1980s? We try to answer it by analyzing and comparing the data on social support networks in 1987 and 2002. The data are interpreted in the context of the transition that was happening in Slovenia at the time.
1. Introduction
The processes of ‘the transition’ in all Central and Eastern European countries (CEE), including Slovenia, implied a complex set of social changes. On the one hand, there were institutional reforms to the political and economic systems, while on the other hand, the provision of welfare was also affected – i.e., which systems were retained or newly established to ensure quality of life, especially in times of crisis. The changes implemented in all the CEE countries were profound, and Slovenia was no exception. Many of these changes have already been widely discussed in literature by various authors, but have more often been observed at the institutional level (i.e., through the welfare state), while at the interpersonal level – how individuals associated, how they perceived new risks and with whom they engaged to confront them – they were less often addressed.
The article addresses the question of how people responded to transition – did patterns of support and social support providers change during this period? We analyze and compare the data on social support networks in 1987 and 2002. Firstly, we present some transitional changes in Slovenia that have an important link to patterns of support – mainly those in the systems of social protection and health, along with general changes in the labour force market and housing. Then we present the theoretical background for studying social networks and their importance in times of crisis and change. This is followed by a description of the methodology and presentation of the results of the study.
2. Transition in Slovenia – new risks
Owing to transition, changes have been introduced in important spheres of society – the political sphere (see, for example, Adam 1994; Adam and Roncevic 2004) and the economic system, along with changes to the welfare system. All these changes have brought many positive developments; however, they have also brought new social risks for individuals.
While this may have been less evident in the political sphere, the changes in the economic sphere brought, for example, high unemployment as well as intensification and flexibilization of work (Ignjatovic 2002; Kanjuo-Mrcela cited in Ignjatovic 2004). The transition to a market economy has also meant a transition for the labor force from agriculture and industry to services. It has led to an increase in unemployment and to a decline in the active population (early retirement schemes) (Ignjatovic 2002: 177–82). There has also been a significant shift from a passive to an active employment policy, which has transferred many new obligations onto individuals (Kopac 2005). The situation of full employment, which was typical before the transition, has prevailed; however, flexible work forms have increased (part-time and limited contracts), but mainly among the younger population (Ignjatovic 2002).
There were significant changes to the social welfare system. The current system is based on the state-socialist welfare system developed in Yugoslavia (Kolaric 1992). The specific structure that formed in the 1990s (Kolaric 1992; Kolaric et al. 2009) was a tripartite system with a well developed and regionally dispersed network of public/state organizations and institutions (production and distribution of services and financial compensation by the state and by the framework of enterprises for employed people), and voluntary and unpaid services within the informal sector.
There has been a gradual introduction of reforms to individual social policies, and the new welfare system can be described as a ‘welfare mix’, which more closely resembles western European welfare systems (Kolaric et al. 2009). The state still has the dominant position as a financial force, whereas the production of social services has been redistributed to the private sector. In the transition from socialist to post-socialist society, the Slovenian welfare system has been constituted as a dual model with elements of the conservative-corporate welfare system (a compulsory social insurance system, based on social partnership) and of the social-democratic welfare system (a strong public/state sector is still the dominant service provider of all types of services, to which all citizens are equally entitled). A nonprofit-voluntary sector has been evolving. On the other hand, the state with its social policy measures did not pay attention to the informal sector (comprising self help and mutual aid among members of family, kin, neighbors and friends) in the past, while at present its support for this sector is increasing.
Social support networks have been vital for ensuring quality of life, and we seek to discover how informal providers of social support have responded to transition. It is possible that they have become even more significant because some of the more generous parts of the social protection system were gradually limited during the reforms. For example, the eligibility criteria for social security measures have in some cases become stricter and the extent of support smaller (see Crnak Meglic 2005). An example can be found in social security for the unemployed. Socialist social policy was primarily based on a policy of full employment and social services provided by the state. After the independence of Slovenia in 1991, the government still tried to ensure a high level of social security for unemployed persons. However, at present the responsibility to ensure employment lies mainly in the hands of individuals and not with the state. Changes that have led to stricter conditions of attainment and maintenance of rights to cash benefits were introduced (Crnak Meglic 2005; Kopac 2005).
Changes in the health system have also led to increasing risk for individuals. Basic (compulsory) health insurance (CHI) still covers almost the entire Slovenian population, and therefore, at least in principle, the health security system is accessible to all. However, some of the health services, medicines, etc. require additional, voluntary health insurance (VHI), which limits its accessibility for the more vulnerable groups (e.g., the unemployed, the retired, the poor). VHI covers payments for health services above the share covered by CHI and for services that represent low value for money and are not covered by CHI. It has been criticized for enlarging social inequality in Slovenia (Stropnik et al. 2003).1
Changes have also been significant in the field of housing and have brought additional risks. Before the transition, housing was guaranteed by the state, but there was also a significant share of ‘self-help building’ (individual building of houses). Among the policy areas, housing has seen the most noticeable decline in the role of government since transition. The most important policies have been denationalization and privatization of housing. The latter led to a significant decline in non-profit rental housing (see Mandic 1994). This has increased the risk of housing vulnerability for low income households and single households, especially when taking into account risk in other sectors (see, e.g., Mandic and Filipovic Hrast 2008). The declining availability and accessibility of housing after transition can also be seen in the rising proportion of young adults still living with parents (Mandic 1996: 36).
The changes described above indicate that several new risks for individuals to cope with have come to the fore in the period of transition. The state has withdrawn support in some cases and extended it in others. Social support networks are thus an important resource for coping with the above mentioned risks, which are translated into a variety of problems that have to be solved at the individual level (e.g., financial problems, illness, and so on).
3. Social networks
The quality (size and composition) of social networks is theoretically and empirically related to social integration (e.g., Berkman and Syme 1979; Hampton and Wellman 1999, 2000; Lin et al.1999; Wright 1999). Network size allows us to identify isolated individuals; however, one has to be aware of the quality of a social network, its structure (e.g., density, heterogeneity and geographical distance) and composition (e.g., proportions of relatives, friends, neighbors and formal ties), as well as the quality of ties and exchange of social support (e.g., duration, intimacy, importance of ties and frequency of contacts).
Social support networks are important because they reduce stress (the main effect model) and also owing to their buffering function, i.e., in protecting persons from the pathogenic consequences of stressful events (the buffering effect model). Empirical evidence can be found for both models (e.g., Franks and Stephens 1996). Cassel (1976), Caplan (1974) and Cobb (1976), who are among the founding fathers of systematic research in the field of social support, stressed the emotional dimension of support and significantly influenced later development in the field. Recent definitions (e.g., House 1981; Vaux 1988, 1992; Burleson et al. 1994) stress social support as a complex interaction and communication process among people. One of the most comprehensive definitions of social support is Vaux's (1985, 1988; see also Thoits 1985) definition of social support as a complex concept of a higher order. Vaux's (1988) concept contains three basic elements: sources of social support, types of social support and social support appraisal. The three elements are connected by a complex dynamic of exchange and communication processes between an individual and his/her social environment.
In this classification, the sources represent the part of the social network to which an individual turns for help and support (or which gives support spontaneously without the individual's explicit request). Usually it is assumed that support networks are relatively stable in size and composition, except in times of life transitions and greater changes (e.g., moving to a higher level of schooling, marriage, divorce, severe accident, job loss, retirement or death of a spouse). Network characteristics (structure, composition and quality of ties) have an effect on the value of the network – its sensitivity, accessibility and capability as a source of support. Types of social support, as the second element of the concept of social support, are specific actions or behaviors that are generally seen as actions with an intent of helping, whether spontaneously or at the individual's request (Vaux 1988, 1992). The third element of social support is the individual's subjective perception of social support. It is an indicator of how well the support network functions and fulfills its aims (Vaux 1988).
In the literature there are a number of definitions of social support dimensions, types and functions (for an overview, see Vaux 1988; Veiel and Baumann 1992). It seems that a relative consensus has been established that there are four major types of social support (e.g., Vaux 1988; Cauce et al. 1990; Walker et al. 1994; Wan et al. 1996): instrumental (material; e.g., help in small or large household tasks), informational support (e.g., advice in times of important life changes such as looking for a new job), emotional support (e.g., providing support when feeling a little depressed or in times of distress) and social companionship (e.g., visiting, going to dinner).
Some network characteristics are very important for studying social support (Walker et al. 1977; Vaux 1988). These characteristics are network size, tie strength, density (the number of actual ties over all possible ties in a network), homogeneity (e.g., by age, gender, etc.) and geographical distance (e.g., Granovetter 1973, 1982; Hirsch 1980; Fischer 1982; Kadushin 1982; Marsden 1987; Sarason et al.1990; Wellman and Wortley 1990; Acock and Hurlbert 1993). Numerous studies show that there is some level of specialization among network members by the type of social support. Emotional support is usually provided by the closest network members (partner/spouse, the closest relatives and friends). On the other hand, the closest network members provide different types of social support. By contrast, a large part of instrumental and informational support is given by emotionally more distant ties (acquaintances, co-workers or neighbors) (e.g., Granovetter 1973). The most frequent social companions are friends (e.g., Wellman and Wortley 1990).
The affective network generator that elicits the people with whom respondents discuss important matters is used in the American General Social Survey (GSS), which focuses on emotionally close and important ties (Burt 1984). Discussion partners are most likely to be friends, relatives, and co-workers who are especially close to respondents. Discussion partners show a high degree of homogeneity with regard to sex, age, religion, and ethnicity (Burt 1984; Marsden 1988). Women tend to report more kin than men do (Marsden 1987). The composition of these discussion networks tends to change over a life-span (Burt 1991).
The structure of networks is influenced by the structure of the social environment, offspring, the history of exchanges and so on. Empirical studies also show changes in network composition in different life cycles (e.g., Fischer 1982; Vaux 1985, 1988; Ishii-Kuntz and Seccombe 1989; van der Poel 1993; Wellman et al. 1997 [1988]; Iglic 1998; Kogovšek 2001). Even though networks are usually stable and their composition varies mainly in response to life cycle changes, significant historical and social changes as have happened in the CEE countries might have an effect on the structure of these networks. What these changes might be in Slovenia in the period of transition is analyzed in the following section.
4. Methodology
Evaluating changes in social support provision over time requires careful consideration of possible explanations for differences and similarities found in survey data. Special attention to survey design is necessary whenever a researcher uses secondary data collected with other than comparative intentions. It has been shown (Hlebec and Kogovšek 2005a, 2005b) that the indicators used in this paper to evaluate social support provision in Slovenia in the 15-year period are equivalent and enable substantive analyses of changes in social support providers.
The aim of the study is to evaluate changes in the structure and composition of social support networks of residents of Slovenia over time and to establish whether and to what extent the change in the social, political and economic system played a part in those changes. Data sets that allow such comparison date from 1987 and 2002. However, those data were collected for different purposes and with considerable differences in methodology (Table 1). In the 1987 study the original Burt name generator was used (a 6-month time frame, actual discussion partners, information about alters was collected only for the first five reported alters), but in the 2002 survey the Burt name generator was somewhat changed (no time frame, the usual discussion partners, data about alters was collected for all named alters). Based on the Bailey and Marsden (1999) study, it can be assumed that response patterns to the Burt name generator can be affected by the context of the questionnaire (preceding questions). It has been established (Hlebec and Kogovšek 2005a, b) that context may have had some effect on the interpretation of ‘important matters’, but not on the network composition.
| Indicator/Study | The Stratification and Level of Living Survey in Yugoslavia, 1987 | Social Support Networks of Residents of Slovenia, 2002 |
| Discussion partners | Name generator, Actual interactions, Time limitation, Reduced to the first five support providers | Name generator, Actual provision, Usual providers, No reduction in number of support providers |
| Indicator/Study | The Stratification and Level of Living Survey in Yugoslavia, 1987 | Social Support Networks of Residents of Slovenia, 2002 |
| Discussion partners | Name generator, Actual interactions, Time limitation, Reduced to the first five support providers | Name generator, Actual provision, Usual providers, No reduction in number of support providers |
Source: Hlebec and Kogovšek 2005b.
A thorough methodological investigation using experimental design (Hlebec and Kogovšek 2005a, b; Kogovšek and Hlebec 2005) revealed that, despite differences in wording, the two indicators were equivalent and could be used for comparative purposes. In the 2002 data set, the data used in comparison was limited to the first five named alters.
4.1. Description of the studies
| Survey | The Stratification and Level of Living Survey in Yugoslavia, 1987 | Social Support Networks of Residents of Slovenia, 2002 |
| Research institute | Institute of Sociology – Ljubljana, Slovenia | CMI – Centre for Methodology and Informatics, Faculty of Social Sciences, University of Ljubljana SPIRS– Social Protection Institute of the Republic of Slovenia |
| Data | Social Science Data Archive – Ljubljana, Slovenia, July 1999 | CMI, SPIRS |
| Data collection | Market Research Centre, Zagreb ZIT/CEMA | CATI center, Ljubljana |
| Sample | Random substitute units replace non-responses within clusters. The substitute units are predefined on a sampling list. The interviewers were allowed to employ the substitute unit only after five attempts to obtain an interview | Random sample of telephone users in Slovenia |
| n | 289 | 5013 |
| Age | 15–75 | 18+ |
| Data collection mode | Personal interview, face-to-face (Burt's name generator), self-administered questionnaire (informal sources of social support, ISSP86 module) | Computer assisted telephone interview |
| Data collection | May 1987–July 1987 | February 2002–April 2002 |
| Survey | The Stratification and Level of Living Survey in Yugoslavia, 1987 | Social Support Networks of Residents of Slovenia, 2002 |
| Research institute | Institute of Sociology – Ljubljana, Slovenia | CMI – Centre for Methodology and Informatics, Faculty of Social Sciences, University of Ljubljana SPIRS– Social Protection Institute of the Republic of Slovenia |
| Data | Social Science Data Archive – Ljubljana, Slovenia, July 1999 | CMI, SPIRS |
| Data collection | Market Research Centre, Zagreb ZIT/CEMA | CATI center, Ljubljana |
| Sample | Random substitute units replace non-responses within clusters. The substitute units are predefined on a sampling list. The interviewers were allowed to employ the substitute unit only after five attempts to obtain an interview | Random sample of telephone users in Slovenia |
| n | 289 | 5013 |
| Age | 15–75 | 18+ |
| Data collection mode | Personal interview, face-to-face (Burt's name generator), self-administered questionnaire (informal sources of social support, ISSP86 module) | Computer assisted telephone interview |
| Data collection | May 1987–July 1987 | February 2002–April 2002 |
Source: Hlebec and Kogovšek 2005b.
Regardless of these differences, both studies give information about support provisions. Both surveys provide representative samples of the Slovenian adult population after weighting. Demographic characteristics of the samples are presented in Table 3.
| Age categories | |||||
| Study | 18–24 | 25–34 | 35–49 | 50–64 | 65–75 |
| SLLSY 1987 | 17 | 21 | 26 | 21 | 15 |
| SSNRS 2002 | 13 | 19 | 30 | 24 | 14 |
| Gender | |||||
| Study | Male | Female | |||
| SLLSY 1987 | 46 | 54 | |||
| SSNRS 2002 | 48 | 52 | |||
| Education1 | |||||
| Study | Elementary s. or below | Vocational school | High school | College and above | |
| SLLSY 1987 | 58 | 13 | 23 | 7 | |
| SSNRS 2002 | 30 | 28 | 29 | 12 | |
| Age categories | |||||
| Study | 18–24 | 25–34 | 35–49 | 50–64 | 65–75 |
| SLLSY 1987 | 17 | 21 | 26 | 21 | 15 |
| SSNRS 2002 | 13 | 19 | 30 | 24 | 14 |
| Gender | |||||
| Study | Male | Female | |||
| SLLSY 1987 | 46 | 54 | |||
| SSNRS 2002 | 48 | 52 | |||
| Education1 | |||||
| Study | Elementary s. or below | Vocational school | High school | College and above | |
| SLLSY 1987 | 58 | 13 | 23 | 7 | |
| SSNRS 2002 | 30 | 28 | 29 | 12 | |
1Differences are statistically significant (χ2=99.27, P<0.001).
Source: Hlebec and Kogovšek 2005b.
5. Analysis
The analysis was done by Multiple Classification Analysis (MCA). MCA (Andrews et al. 1973) is a multivariate method by which relationships between multiple independent variables (or predictors) and a dependent variable are analyzed. It is similar to multiple regression, with the advantage that nominal measurement level variables need not be dichotomized. Multiple classification analysis gives us the following information:
- -
the overall (grand) mean and group means of the dependent variable for each combination of categories of predictors;
- -
tests of significance of the effects of single predictors as well as of interactions between them;
- -
the effect of each predictor is shown by parameter β, which tells us the effect of the predictor if other predictors are held constant; the rank order of βs shows us the relative importance of a single predictor in explaining the dependent variable;
- -
deviations from the grand mean of the dependent variable for each category of a predictor (therefore, how much would the grand mean of the dependent variable increase or decrease as a result of the effect of a certain predictor); and
- -
the percentage of explained variance for all predictors included in the analysis (R2).
In Table 4 results of the MCA analyses are presented. For each dependent variable (network composition variables), two separate MCA models were estimated. In this paper we are primarily interested in whether there is any effect of time on network composition, and other variables are used mainly as control variables.
| . | . | Close kin2 . | Friends . | Co-workers . | Neighbors . | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| . | . | Grand mean=55 . | Grand mean=30 . | Grand mean=6 . | Grand mean=3 . | ||||||||
| . | N . | . | Multivariate . | . | Multivariate . | . | Multivariate . | . | Multivariate . | ||||
| . | . | . | Beta . | deviation . | . | Beta . | deviation . | . | Beta . | deviation . | . | Beta . | deviation . |
| YEAR | *** | * | *** | *** | |||||||||
| QL 1987 | 210 | −18 | −6 | 11 | 8 | ||||||||
| SSN 2002 | 4464 | 0.086 | 1 | 0.032 | 0 | 0.118 | −1 | 0.111 | 0 | ||||
| AGE | *** | ** | *** | *** | |||||||||
| 18–29 | 1268 | 4 | 0 | −2 | −2 | ||||||||
| 30–49 | 1900 | −4 | 2 | 3 | −1 | ||||||||
| 50–75 | 1506 | 0.075 | 1 | 0.056 | −3 | 0.112 | −2 | 0.134 | 3 | ||||
| GENDER | *** | *** | * | ||||||||||
| male | 2257 | 6 | −4 | −1 | 0 | ||||||||
| female | 2417 | 0.126 | −5 | 0.098 | 4 | 0.030 | 1 | 0.021 | 0 | ||||
| MARITAL STATUS | *** | *** | * | ** | |||||||||
| single | 1309 | −16 | 17 | −2 | 0 | ||||||||
| divorced/widowed | 450 | −12 | 7 | 1 | 2 | ||||||||
| married | 2915 | 0.259 | 9 | 0.274 | −8 | 0.054 | 1 | 0.050 | 0 | ||||
| MODEL 1: multiple R2 | 0.083 | 0.089 | 0.032 | 0.034 | |||||||||
| EDUCATION | ** | *** | * | *** | |||||||||
| Elementary or below | 792 | 3 | −7 | −1 | 3 | ||||||||
| secondary | 2927 | 0 | 0 | 0 | 0 | ||||||||
| College or above | 956 | 0.052 | −4 | 0.095 | 6 | 0.041 | 2 | 0.097 | −1 | ||||
| AGE | *** | *** | *** | *** | |||||||||
| 18–29 | 1268 | −7 | 12 | −3 | −2 | ||||||||
| 30–49 | 1899 | 2 | −3 | 3 | −1 | ||||||||
| 50–75 | 1508 | 0.092 | 4 | 0.181 | −6 | 0.122 | −1 | 0.124 | 3 | ||||
| MODEL 2: multiple R2 | 0.040 | 0.051 | 0.033 | 0.041 | |||||||||
| Partner | Parents | Children | Siblings | ||||||||||
| Grand mean=34 | Grand mean=7 | Grand mean=6 | Grand mean=7 | ||||||||||
| N | Multivariate | Multivariate | Multivariate | Multivariate | |||||||||
| Beta | deviation | Beta | deviation | Beta | deviation | Beta | deviation | ||||||
| YEAR | *** | ||||||||||||
| QL 1987 | 210 | −12 | −2 | −2 | −2 | ||||||||
| SSN 2002 | 4464 | 0.060 | 1 | 0.018 | 0 | 0.021 | 0 | 0.021 | 0 | ||||
| AGE | *** | *** | *** | * | |||||||||
| 18–29 | 1268 | 6 | 6 | −5 | −2 | ||||||||
| 30–49 | 1900 | −1 | 0 | −3 | 1 | ||||||||
| 50–75 | 1506 | 0.095 | −4 | 0.177 | −4 | 0.286 | 9 | 0.057 | 1 | ||||
| GENDER | *** | * | *** | * | |||||||||
| male | 2257 | 9 | −1 | −2 | −1 | ||||||||
| female | 2417 | 0.210 | −9 | 0.035 | 1 | 0.086 | 2 | 0.036 | 1 | ||||
| MARITAL STATUS | *** | *** | *** | *** | |||||||||
| single | 1309 | −23 | 5 | −1 | 3 | ||||||||
| divorced/widowed | 450 | −21 | −2 | 8 | 3 | ||||||||
| married | 2915 | 0.406 | 13 | 0.144 | −2 | 0.124 | −1 | 0.113 | −2 | ||||
| MODEL 1: multiple R2 | 0.203 | 0.085 | 0.137 | 0.012 | |||||||||
| EDUCATION | * | * | *** | ||||||||||
| Elementary or below | 792 | −3 | 2 | 4 | 1 | ||||||||
| secondary | 2927 | 1 | −1 | 0 | 0 | ||||||||
| College or above | 956 | 0.039 | −1 | 0.037 | 0 | 0.095 | −3 | 0.031 | –1 | ||||
| AGE | *** | *** | *** | ||||||||||
| 18–29 | 1268 | −9 | 9 | −7 | 0 | ||||||||
| 30–49 | 1899 | 6 | −1 | −3 | 0 | ||||||||
| 50–75 | 1508 | 0.148 | 0 | 0.276 | −6 | 0.321 | 9 | 0.007 | 0 | ||||
| MODEL 2: multiple R2 | 0.089 | 0.076 | 0.133 | 0.004 | |||||||||
| . | . | Close kin2 . | Friends . | Co-workers . | Neighbors . | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| . | . | Grand mean=55 . | Grand mean=30 . | Grand mean=6 . | Grand mean=3 . | ||||||||
| . | N . | . | Multivariate . | . | Multivariate . | . | Multivariate . | . | Multivariate . | ||||
| . | . | . | Beta . | deviation . | . | Beta . | deviation . | . | Beta . | deviation . | . | Beta . | deviation . |
| YEAR | *** | * | *** | *** | |||||||||
| QL 1987 | 210 | −18 | −6 | 11 | 8 | ||||||||
| SSN 2002 | 4464 | 0.086 | 1 | 0.032 | 0 | 0.118 | −1 | 0.111 | 0 | ||||
| AGE | *** | ** | *** | *** | |||||||||
| 18–29 | 1268 | 4 | 0 | −2 | −2 | ||||||||
| 30–49 | 1900 | −4 | 2 | 3 | −1 | ||||||||
| 50–75 | 1506 | 0.075 | 1 | 0.056 | −3 | 0.112 | −2 | 0.134 | 3 | ||||
| GENDER | *** | *** | * | ||||||||||
| male | 2257 | 6 | −4 | −1 | 0 | ||||||||
| female | 2417 | 0.126 | −5 | 0.098 | 4 | 0.030 | 1 | 0.021 | 0 | ||||
| MARITAL STATUS | *** | *** | * | ** | |||||||||
| single | 1309 | −16 | 17 | −2 | 0 | ||||||||
| divorced/widowed | 450 | −12 | 7 | 1 | 2 | ||||||||
| married | 2915 | 0.259 | 9 | 0.274 | −8 | 0.054 | 1 | 0.050 | 0 | ||||
| MODEL 1: multiple R2 | 0.083 | 0.089 | 0.032 | 0.034 | |||||||||
| EDUCATION | ** | *** | * | *** | |||||||||
| Elementary or below | 792 | 3 | −7 | −1 | 3 | ||||||||
| secondary | 2927 | 0 | 0 | 0 | 0 | ||||||||
| College or above | 956 | 0.052 | −4 | 0.095 | 6 | 0.041 | 2 | 0.097 | −1 | ||||
| AGE | *** | *** | *** | *** | |||||||||
| 18–29 | 1268 | −7 | 12 | −3 | −2 | ||||||||
| 30–49 | 1899 | 2 | −3 | 3 | −1 | ||||||||
| 50–75 | 1508 | 0.092 | 4 | 0.181 | −6 | 0.122 | −1 | 0.124 | 3 | ||||
| MODEL 2: multiple R2 | 0.040 | 0.051 | 0.033 | 0.041 | |||||||||
| Partner | Parents | Children | Siblings | ||||||||||
| Grand mean=34 | Grand mean=7 | Grand mean=6 | Grand mean=7 | ||||||||||
| N | Multivariate | Multivariate | Multivariate | Multivariate | |||||||||
| Beta | deviation | Beta | deviation | Beta | deviation | Beta | deviation | ||||||
| YEAR | *** | ||||||||||||
| QL 1987 | 210 | −12 | −2 | −2 | −2 | ||||||||
| SSN 2002 | 4464 | 0.060 | 1 | 0.018 | 0 | 0.021 | 0 | 0.021 | 0 | ||||
| AGE | *** | *** | *** | * | |||||||||
| 18–29 | 1268 | 6 | 6 | −5 | −2 | ||||||||
| 30–49 | 1900 | −1 | 0 | −3 | 1 | ||||||||
| 50–75 | 1506 | 0.095 | −4 | 0.177 | −4 | 0.286 | 9 | 0.057 | 1 | ||||
| GENDER | *** | * | *** | * | |||||||||
| male | 2257 | 9 | −1 | −2 | −1 | ||||||||
| female | 2417 | 0.210 | −9 | 0.035 | 1 | 0.086 | 2 | 0.036 | 1 | ||||
| MARITAL STATUS | *** | *** | *** | *** | |||||||||
| single | 1309 | −23 | 5 | −1 | 3 | ||||||||
| divorced/widowed | 450 | −21 | −2 | 8 | 3 | ||||||||
| married | 2915 | 0.406 | 13 | 0.144 | −2 | 0.124 | −1 | 0.113 | −2 | ||||
| MODEL 1: multiple R2 | 0.203 | 0.085 | 0.137 | 0.012 | |||||||||
| EDUCATION | * | * | *** | ||||||||||
| Elementary or below | 792 | −3 | 2 | 4 | 1 | ||||||||
| secondary | 2927 | 1 | −1 | 0 | 0 | ||||||||
| College or above | 956 | 0.039 | −1 | 0.037 | 0 | 0.095 | −3 | 0.031 | –1 | ||||
| AGE | *** | *** | *** | ||||||||||
| 18–29 | 1268 | −9 | 9 | −7 | 0 | ||||||||
| 30–49 | 1899 | 6 | −1 | −3 | 0 | ||||||||
| 50–75 | 1508 | 0.148 | 0 | 0.276 | −6 | 0.321 | 9 | 0.007 | 0 | ||||
| MODEL 2: multiple R2 | 0.089 | 0.076 | 0.133 | 0.004 | |||||||||
1Significance levels are marked as follows: * (0.01<P<0.05), ** (0.005<P<0.001), *** (P<0.001).
2Partner, parents, siblings and children together.
In general, it can be seen that the effect of time on network composition is relatively weak. In most cases the effects of control variables are stronger, especially age and marital status, in some cases also the effect of gender. This is an expected result, as we have already emphasized that social networks tend to be relatively stable over time.
The effects of control variables are mainly consistent with previous studies on social support networks (e.g., Wellman 1979; Burt 1984; Marsden 1987; van der Poel 1993; Schweizer et al. 1998). The percentage of (close) kin tends to be higher in the youngest and the older age groups, with persons of lower rather than higher education and with married persons. Parents tend to be rather important for younger persons and children for the older age group. Siblings also tend to become more important with ageing and also for single, divorced or widowed persons. Friends play a more prominent role with younger, single, divorced or widowed and highly educated persons. Co-workers are usually among the weak ties, but may be more prominent in the active life cycle (middle age group). Neighbors, as a geographically close potential source of social support, tend to be important for older persons.
However, the effect of time is relatively strong and highly statistically significant (the first or second strongest) in the case of the percentages of co-workers and neighbors, which are considerably higher in 1987 than in 2002. The difference between the years is also quite prominent in the case of the percentages of partner and close kin together (both lower in 1987 than in 2002). The differences are moderate for friends and low for some categories of close kin (parents, children and siblings).We conclude that there were changes in the composition of social networks during the transition. Discussion partners have become less numerous and more intimate by 2002, lowering the number of co-workers and neighbors and focusing on close kin. In the next section we interpret this change in the context of the transitional changes happening in Slovenia.
6. Discussion
The changes that occurred during transition show a narrowing of discussion partners and more focus on close kin. In this section we will offer some explanations for these changes. It could be speculated that an explanation for such dramatic changes could be found in the political situation in Slovenia and Yugoslavia in the late 1980s. There were many actual events around 1987 that could be outlined to illustrate the general situation in Slovenia. Let us mention just two of these: in 1986 the Yugoslav army started threatening to take over if the civil leadership refused to lead Yugoslavia along Tito's way, and in February 1987 a special edition of the journal ‘Nova Revija’ proposed a program for establishing a Slovenian nation in opposition to the Serbian nationalist program proposed by the Serbian Academy of Sciences and Arts. In the late 1980s politics was vitally important as a discussion topic, since the turmoil, which later led to the independence of Slovenia and the terrible war in the Balkans, had already started at the time. This situation led to heated discussions in daily newspapers and magazines and among people – therefore potentially involving a wider variety of discussion partners, such as co-workers and neighbors. After the transition and with the gradual ‘normalisation’ of the new political and economic system, these discussions might have changed, with political topics perhaps giving way to more intimate personal matters, which might also have meant limiting this wider network of discussing partners to closer family members and friends.
Furthermore, less discussion with co-workers might also be linked to changes in the labour market. Intensification at work and less stability and security in the work place might have led to looser relations among co-workers, lowering their number among the support providers in social networks.
Another interesting change is the decline in the role of neighbors as discussion partners. This might again be linked to the already described broader political context and the discussions that were stimulated among the wider public (also neighbors). It might also be linked to the changes that the transition brought at the community level. Reorganization of and changes to local communities at the beginning of the 1990s (establishment of smaller municipalities, abandonment of local communities within cities and their combination into larger city quarters) in some cases led to loss of community places for socializing and discussion (see Dragoš and Leskošek 2003). Similarly, other authors (Filipovic et al. 2005; Hlebec and Mandic 2005; Mandic and Hlebec 2005; Filipovic 2007) claim that the development of community networks was not supported during the period of transition. Participation in and integration into local communities or city neighborhoods was not, for instance, supported by targeted local policy (see Ploštajner et al. 2004). Consequently, the transition brought some degree of alienation at the local level, possibly leading to fewer neighbors among closer network members.
The general changes in welfare might also have had an effect on the composition of social networks, but perhaps that would be more evident in the financial support networks and networks in the case of illness than in the discussion networks – which, however, could not be analyzed in this paper.
Furthermore, housing issues might also have had some link to these changes. Before the transition, the state was the dominant supplier of housing. However, the share of self-building (especially outside cities) was also significant. This kind of self-help building was usually done by families themselves and their close kin, with the help of neighbors, thus establishing stronger ties among neighbors. After the transition this kind of building declined, potentially leading to looser relations with neighbors.
Finally, we would like to add some comments as to the strengths and limitations of our study and provide some possibilities for further work. The strength of our study lies especially in the attempt to study and explain possible effects of a profound transition in a country that recently underwent a change from one political and socioeconomic system to another on the lives of ordinary people at the micro level; i.e., how their informal social support networks (an important factor for any person's well-being) responded to such large changes in the society. Most studies of which we are aware deal with these changes at the macro or meso level. We believe that the observed differences in support providers are substantive ones.
However, our study has weaknesses as well. The main weakness lies in the data used, which, as already mentioned in the methodological section of the paper, were secondary data, collected for purposes other than ours. Additionally, there were several important methodological differences between the two data sets, which, despite our careful examination and adaptation, could not be made completely comparable.
In addition to this, it is somewhat difficult to confirm whether the discussed changes in the network composition are the result of the described changes brought by the transition, or they are perhaps also a significant byproduct of the general societal changes such as globalization, rise of the new media (e.g., Internet) and general changes in personal relationships, found all over Europe. A general trend toward individualization, which has been present in all European societies for some time, combined with emotional closing and a turn to family and friends, could be linked to the decline in discussions with and support from co-workers and neighbors. This would be best analysed if we could compare our data with a similar data in another transition country and/or a Western European country that did not experience the transition. To our knowledge, unfortunately, no such directly comparable data exists. However, for example, a longitudinal study of importance of neighboring in USA showed no such strong decline in time (see Guest and Wierzbicki 1999), and similarly no such change in neighborhood networks was found in Sweden when comparing 1983 and 1993 (see Henning and Lieberg 1996).
Therefore, careful examination and delineation of transition factors (changes in Slovenia owing to the transformation of the political and socioeconomic system) from developmental factors (changes in the broader society – e.g., Europe) are needed in the future. One possibility would be to complement the survey data used in this study with other sources of data, such as archival and statistical data from the period. The interpretation could be further enriched by contextualizing the findings with, for instance, an examination of documents, historical sources and studies and other survey data.4 Therefore, our conclusions and discussion should be seen as a starting point for further, deeper and more precise sociological explanation of the observed changes in social support provision.
Footnotes
These percentages are estimated in two stages. Firstly, the precentages are estimated on the level of egocentered networks, i.e., for each respondent his/her individual network composition is obtained. These percentages are the dependent variables in the MCA model. In the MCA model the precentages are then estimated across all respondents: the total (grand) mean and the means by subgroups, defined by included independent variables (represented in the model by deviations from the grand mean). This ‘double aggregation’ of data is the reason why the network composition percentages do not add up to 100 percent in the MCA model (also the category ‘other’ is excluded from the model).
Despite a rather large sample in year 2002 it is impossible to include all predictors in one model. Therefore, two models were estimated with a slightly different combination of predictors. The first model includes the following variables: year of study, age, gender and marital status. In the second model education is used instead of marital status (other three predictors are the same). Effects estimates between the models differ only for age, which is in interaction with education. Therefore, the model estimates change slightly with variable age and are therefore included in Table 4 for model 2 as well. Model 2 estimates are the same for year of study and gender.
For instance, in Slovenia a large longitudinal survey, Slovene Public Opinion, has been carried out since 1968. In the study, rich data on opinions and attitudes about various important topics is collected at least yearly on representative samples of the general population. Since some batteries of questions are repeated from time to time, the study also enables examination of trends. These data could therefore provide another rich source of contextual information for our research problem.
References
Valentina Hlebec is associate professor for sociology at University of Ljubljana, Faculty of Social Sciences. Her research interests cover topics in sociology of ageing, social networks and survey methodology, especially designing and testing survey questionnaires. She has been involved in several national and international project groups in these fields.
Maša Filipovič Hrast is a researcher and assistant professor at the Faculty of Social Sciences, University of Ljubljana. Her research topics are social cohesion and social inclusion issues, social policy and housing policy. The emphasis is on social networks and cohesion at the community and neighbourhood level, as well as situations of vulnerable groups (e.g. elderly and homeless) and their quality of life and inclusion in society.
Tina Kogovšek is associate professor for social science methodology at University of Ljubljana, Faculty of Arts and Faculty of Social Sciences. Her research interests cover social science methodology, especially social network analysis and measurement and quality of measurement. She has been involved in several national and international project groups in these fields.