Social Networks and Subjective Wellbeing in Australia: New Evidence from a National Survey

Abstract

The article draws on data from a national survey in Australia in 2014 to examine how social networks affect life satisfaction and happiness. Findings show that social network composition, social attachment, perceived social support and the volume of social resources are significantly positively associated with life satisfaction and happiness. Stress about social commitments, feeling restricted by social demands and being excluded by a social group are negatively associated with life satisfaction and happiness. These results indicate that social networks have both ‘bright side’ and ‘dark side’ effects on subjective wellbeing.

Keywords

Australia happiness life satisfaction social networks subjective wellbeing

Introduction

Across the social sciences, social networks have been found to play an important role in educational attainment (Furstenberg and Hughes, 1995), parenting (Hagan et al., 1996), crime reduction (Hagan and McCarthy, 1995) and employment (Bian and Huang, 2009; Bian et al., 2015; Huang and Western, 2011). In recent years, social scientists and policy-makers have asserted that social networks also matter for health and wellbeing (Christakis and Fowler, 2009; Huxhold et al., 2013). However, despite the ubiquity of the social network concept scholars have highlighted a number of problems with theoretical and empirical work (Durlauf and Fafchamps, 2005; Lin, 2001; Lyons, 2011). In this article we introduce a new model for the relationship of social networks to subjective wellbeing (SWB) and present data from a national Australian survey to explore the model.

Subjective wellbeing refers to people’s evaluations of their own lives, including life satisfaction (a cognitive evaluation) and happiness (an affective evaluation) (Bian et al., 2017; Diener, 2006). In psychology, research on SWB was triggered by Brickman and Campbell (1971) who argued that most individuals had an equilibrium (set-point) level of SWB which was only temporarily disturbed by significant life events. The idea of a set-point dominated SWB research for nearly 30 years (Headey, 2007) but began to be undermined by evidence that certain events such as marriage (Gottman, 1996) or repeated unemployment (Clark et al., 2004) were associated with long-term changes in individuals’ SWB. This led some (Headey, 2007) to argue that the study of SWB should concentrate on the factors leading to enduring changes in SWB such as having life goals linked to social networks (e.g. family, friends, social participation). However, network research on subjective outcomes has remained comparatively uncommon (Hauberer, 2011; Western and Tomaszewski, 2016) and early influential social network research on SWB (Christakis and Fowler, 2009; Helliwell, 2006; Wilkinson, 2002) has also been convincingly criticised (Lyons, 2011). We attempt to address both the gap in the literature and some of these criticisms in this study.

The concept of social networks has been defined in various ways. It may refer to digital and online networks, personal relationships, political associations, economic transactions and geopolitical relations among nations and international systems (Kadushin, 2012; Scott, 2017). In empirical studies, social networks have been defined as formal and informal social relationships (Bian et al., 2017; Coleman, 1988), properties of network structure (e.g. size and frequency of contact) (Huxhold et al., 2013), the norms (e.g. trust, reciprocity) that characterise networked relationships (Putnam, 2000) and positive outcomes from such relationships (e.g. economic output, social cohesion, health) (Helliwell and Putnam, 2004; Huxhold et al., 2013). As Durlauf and Fafchamps (2005) note, much research fails to clearly distinguish elements within bundles of related concepts (e.g. norms, trust, networks, positive outcomes), specify the mechanisms through which networks have effects and develop rigorous measures and appropriate empirical analyses where network variables and theoretical assertions are tested alongside alternative explanations. In addition, prior research tended to assume that networks were only related to intended beneficial outcomes and did not consider adverse unintended consequences such as social exclusion or economic inefficiency (Coleman, 1988). Moreover, there was a lack of scholarly attention to the dark side of social networks, such as relational and psychological stress, reduced SWB (Fox and Moreland, 2015; Kross et al., 2013) and negative relationships that impacted on workplace outcomes in organisations (Labianca and Brass, 2006). In the following sections, we offer a clear definition of social networks and propose a comprehensive and explicit model to examine both ‘bright side’ and ‘dark side’ effects of social networks on SWB.

Theoretical Model and Hypotheses

In this study, a social network is broadly defined as a structure made up of a set of social actors or entities (such as individuals or organisations) and the relationships among them (Robins, 2015; Scott, 2017). The central research questions we want to address are: Are social networks associated with SWB? If so, what network characteristics or mechanisms are involved and to what effect? As mentioned above, the existing research of social networks on SWB has not been common but a few welcome exceptions have produced initial findings contributing to this field. Huang and Western (2015) found SWB was significantly associated with social support people obtained from family members, friends and acquaintances in Australia. In related studies of networks, trust and SWB, Li and his co-authors (2005) showed that good neighbourly relations and wider social support networks significantly affect social trust. Again it was revealed that the social status of contacts, rather than the social distance between them, influenced social trust (Li et al., 2008). More recently Li (2016) showed that having good neighbourly relations and being actively engaged in civic activities was conducive to people’s health, happiness and life satisfaction but network diversity moderated these relationships. Bian and his collaborators (2017) compared the relative importance of formal networks (e.g. group membership) and informal networks (e.g. contact with friends and family) in affecting people’s SWB in Australia, Britain and China. The different research findings reflect the fact that research in this area uses different data sources and different analytical frameworks, rendering it difficult to establish a consensus about how and why social networks (e.g. types, facets and levels) influence SWB. We attempt to address this gap by proposing and examining a new model of social network influences on SWB. To do so, we first review classical accounts of social networks and then utilise the insights of the literature to develop our model and hypotheses.

Networks provide actors with resources to pursue desired outcomes. However, researchers disagree about why social networks matter (Lin, 2001). One key dispute is whether personal resources associated with network locations or embedded (i.e. social) resources associated with network connections matter. Within the network locations account, Bourdieu (1986) and Coleman (1988) hold that network positions lead to social benefits because they allow network members to recognise and mutually reward each other and exclude outsiders. Network location, in other words, allows group members to use social closure and group solidarity to pursue their interests. Alternatively, Burt (1992) emphasises the importance of a strategic location or ‘structural hole’, such as a bridge linking networks, through which occupants gain advantages by accessing more diverse resources and influences or by controlling their access by other network members. Granovetter (1995) likewise stresses how an individual’s weak ties (e.g. acquaintanceship) serve as information bridges between networks in job searches. These network mechanisms are linked to the positions or locations actors occupy and thus are individual positional attributes.

In contrast, in the embedded or social resources account (Lin, 2001), network resources are a meso-level property accessible through direct and indirect ties. Network resources may include things like the number of ties or the resources generated through ties. These resources are network, not actor, properties (Lin, 2001). Network locations facilitate access to better embedded resources but such resources are a property of the whole network, not an actor’s position in it.

The foregoing discussion shows that the network location account calls attention to factors like network closure, density and strategic position while the social resources account stresses resources that can be accessed or mobilised. However, no research has incorporated these elements in one comprehensive model using empirical data to test related assumptions. To address this gap we propose a new theoretical model incorporating network characteristics and network mechanisms (Figure 1). We include network mechanisms that may adversely affect SWB as well as those expected to enhance it.

Figure 1.

Theoretical model of network characteristics and mechanisms on SWB.

Network characteristics include properties of network composition (e.g. size, density and other features) and strength of ties between network members. These factors work with or through network mechanisms to influence SWB. The positive mechanisms associated with beneficial outcomes include social attachments, social resources and social support. Negative network mechanisms (i.e. mechanisms which we hypothesise as being associated with worse SWB) are relational constraints, which can be regarded as ‘downward levelling norms’ (Portes, 1998: 17) or the ‘dark side’ of social networking experience (Fox and Moreland, 2015). Our theoretical model distils previous arguments (e.g. Granovetter, Lin and Coleman) and posits separately measured network mechanisms to address conceptual and methodological criticisms that prior research does not specify or test (Durlauf and Fafchamps, 2005).

Based on this model, we formulate the hypotheses below. The first two hypotheses focus on the association between network characteristics (i.e. network composition and strength of ties) and SWB reflecting arguments which emphasise network location (Bourdieu, 1986; Burt, 1992; Coleman, 1988; Granovetter, 1995). These hypotheses make explicit the positive associations between networks and SWB assumed in the literature (Durlauf and Fafchamps, 2005) and they supplement each other: Hypothesis 1 addresses density of networks (as a kind of network property), which indicates the closeness between and among all relevant network members. Hypothesis 2 stresses the strength of ties between the individual actor and his/her social contacts.

Hypothesis 1: Having dense networks or close social connections tends to enhance life satisfaction and happiness.

Hypothesis 2: Having strong ties based on frequent contact and reciprocity tends to enhance life satisfaction and happiness.

Our other hypotheses investigate the associations between relational mechanisms and SWB (see Figure 1). In particular, Hypotheses 3–5 examine assertions about social attachment, social support and social resources respectively. Some of these factors have been examined in previous research (e.g. Li, 2016; Lin, 2001; Portela et al., 2013; Van Der Gaag and Snijders, 2005) but few studies have ever incorporated them in one model for empirical verification.

Hypothesis 3: Having strong social attachments indicated by social participation, neighbourhood belonging and social trust tends to enhance life satisfaction and happiness.

Hypothesis 4: Having a higher level of perceived social support tends to enhance life satisfaction and happiness.

Hypothesis 5: Having more social resources tends to enhance life satisfaction and happiness.

Hypothesis 6 aims to investigate negative relational mechanisms that have been rarely studied in previous empirical research. We borrow ideas from Becker’s (1991) rational choice theory and Honneth’s (1996) account on social exclusion and social recognition and formulate the hypothesis below:

Hypothesis 6: Feeling stressed about social commitments, restricted by social demands or excluded by social groups tends to decrease life satisfaction and happiness.

Methods and Variables

Data

We drew on data from the National Survey on Social Networks and Subjective Wellbeing in Australia (AuSNet 2014) for our statistical analysis (Western et al., 2016a). AuSNet 2014 is a comprehensive personal network survey of the Australian adult population, based on a large national probability sample of adult respondents (N = 2310). Our questionnaire is too varied in content and format to be administered by telephone interview and too expensive to be administered face-to-face by interviewers. We therefore opted for self-completion by respondents, either online or by mail. This is a cost-effective approach that allows respondents to see the questionnaire format.

We invited respondents into the study by telephone with numbers selected randomly using landline and mobile telephone sampling frames. Respondents had to be at least 18 years of age to be included. The landline recruitment sample was geographically stratified by state while the mobile recruitment was via a simple random sample, because Australian mobile phone numbers do not identify state of residence. Respondents who agreed to participate then completed questionnaires online or in hardcopy depending on their preference. The dual-frame (i.e. mobile and landline sampling frames) telephone recruitment strategy addresses population non-coverage issues associated with relying solely on a landline-based sampling frame (Baffour et al., 2016). Respondents were first invited to complete the survey online, and if they declined were offered a hardcopy option through the mail. Offering our preferred survey mode first was more likely to lead to higher response rates than providing respondents with a choice of survey modes at the outset (Millar and Dillman, 2011). After follow-ups the final response rates for web and hardcopy completed surveys were 18.6 per cent for recruitment from the landline sample, 26.9 per cent for the mobile sample and 22.7 per cent for the overall sample (Western et al., 2016b). The response rates are within ranges for web and mail surveys noted in a comprehensive meta-analysis of existing studies (Manfreda et al., 2008). For instance, an online survey investigating social capital and the use of social networking sites by college students had a lower response rate of 8.2 per cent (Valenzuela et al., 2009).

Dependent Variables

SWB is defined in terms of two key dimensions: a cognitive evaluation of one’s life (life satisfaction) and an affective, emotional evaluation (happiness) (Davern et al., 2007; Diener, 2006). We measure these concepts as follows.

Life Satisfaction

Life satisfaction is measured with the Satisfaction with Life Scale (Diener et al., 1985) which is among the most widely used measures in psychology and has been employed in a series of social surveys around the world (see Dolan et al., 2008 for a review). Respondents were asked to indicate their level of agreement about their life at present by giving a score (1–7, from ‘strongly disagree’ to ‘strongly agree’) to each of the following statements: ‘In most ways my life is close to my ideal’; ‘The conditions of my life are excellent’; ‘I am satisfied with my life’; ‘So far I have gotten the important things I want in life’; and ‘If I could live my life over, I would change almost nothing’ (Cronbach’s α = .881). As shown in Table 1, in the analytical sample of this article, the average score of life satisfaction is 4.71 on a scale of 1–7.

Table 1.

Descriptive statistics of variables, AuSNet Survey, 2014 (N = 1713).

Variables	Mean (SD) or %
Subjective wellbeing (SWB)
Life satisfaction (on a scale of 1–7)	4.71 (1.35)
Happiness (on a scale of 1–11)	8.07 (1.79) (N = 1708)
Network composition
Size of whole discussion network	5.97 (12.55)
Number of kin in core discussion network	1.65 (1.27)
Number of non-kin in core discussion network	2.08 (1.36)
Density of core discussion network (0–1)	.67 (.20)
Strength of ties
% Reciprocity of providing advice/help (strong = 1)	83.65
% Frequency of contact (high = 1)	79.80
Social attachment
% Having one club membership	22.42
% Having more than one club membership	28.20
Extent of social participation (on a scale of 1–6)	2.19 (.81)
Neighbourhood belonging (on a scale of 1–7)	4.65 (1.35)
Social trust (on a scale of 1–7)	3.94 (.81)
Social support
Types of perceived support (0–10)	8.75 (1.73)
Social resources
Number of contacts in listed occupations (0–20)	9.58 (4.56)
Total status score of generated occupations	583.35 (278.54)
Highest status score of generated occupations	93.02 (11.39)
Range of status scores of generated occupations	68.58 (19.56)
Relational constraints and exclusion
% Stress about social commitments	72.50
% Restriction of social demands	70.11
% Exclusion of social groups	39.23
% Rules of club membership	18.80
% Reluctance of providing help	69.12
Demographic & socioeconomic characteristics
% Gender (male = 1)	41.21
Age	48.13 (16.01)
Age²	2572.92 (1585.05)
% Certificate or diploma	31.29
% Bachelor degree and above	44.60
% Other education qualification (ref.)	24.11
% Employee or self-employed (ref.)	59.02
% Unemployed or looking for work	3.21
% Retired	13.31
% Full-time student	5.49
% Household duties	7.76
% Other employment status	11.21
% Low household income ($799 per week and below)	17.40
% Middle income ($800–$1499 per week)	20.14
% High income ($1500 per week and above) and other (ref.)	62.46
% Single/never married (ref.)	19.61
% Married or de facto	65.38
% Divorced or separated	11.56
% Widowed	3.44
% Migrant	26.21
% Urban resident	64.80
% Having a religion	53.01
Weekly hours on social activities	10.64 (11.86)
$ Weekly expenditure on social activities	79.03 (81.93)

Note: The results in this table are based on variables that are included in the full models in Table 4.

Happiness

A scale of happiness was prepared for respondents for self-evaluation (Fordyce, 1977). Again this has been a classical measure in social science research: ‘People often talk about “happiness” in their life. Where would you place yourself on a scale from 0 to 10, where 0 means “extremely unhappy” and 10 means “extremely happy”?’ The mean score of Australians’ happiness is 8.07 on a scale of 1–11, which is close to the results from previous surveys in Australia and Britain (Bian et al., 2017). There is also a positive correlation between life satisfaction and happiness (r = .723, p < .01).

Independent Variables: Network Characteristics

Network Composition

We used name generators (Burt, 1984) to measure network composition and asked respondents about their ‘core discussion networks’ (McPherson et al., 2006; Marsden, 1987), a type of strong-ties network that has been well documented in the literature. The questions were as follows: ‘From time to time, most people discuss important matters with other people. Looking back over the last six months, have you discussed matters important to you with other people?’ If so, ‘with how many people have you discussed matters important to you?’ Respondents were required to report their relationship with the named discussant (i.e. family or relative, including spouse/partner, friend and other), how well they knew the named discussants and the mutual familiarity between the named discussants (i.e. on a five-point scale, from ‘Don’t know each other at all’ to ‘Extremely familiar with each other’). Based on the information we construct four variables to indicate respondents’ network composition, namely, the size of whole discussion network, the number of kin in the core discussion network, the number of non-kin in the core discussion network and the density of the core discussion network (see Table 1).¹

Strength of Ties

Granovetter (1973, 1995) suggested four dimensions measuring strength of ties: frequency of contact; emotional intensity; intimacy; and reciprocal services. Bian (1997) added measures of role relationships such as kin, friends and acquaintances. In this study, the available information in the AuSNet questionnaire allowed us to construct two variables about the strength of ties among the members of the core discussion network: reciprocity of providing advice or help (strong = 1) and frequency of contact (high = 1), which are based on the following two questions: ‘Did this person ask you for any advice or help in the last six months?’ and ‘Generally, how often do you contact each other?’ Strong reciprocity of helping means that at least 50 per cent of the named discussants asked the respondent for advice or help. High frequency of contact indicates that the ‘mean frequency of contact’ (McPherson et al., 2006: 360) among these discussion network members was at least once a week (i.e. 52 days per year).

Independent Variables: Relational Mechanisms

Social Attachment

Social attachment can be measured in terms of strength of connectedness across formal and informal relationships with others. We examine social attachment with variables measuring strength of formal organisational attachment, informal attachment, neighbourhood belonging and social trust:

(1) Club membership. Respondents were asked if they were currently an active member of a sporting, hobby or community-based club or association, and if it was the case, how many groups they were currently an active member of.

(2) Extent of social participation. Respondents were asked ‘In general, how often do you get together socially with “Family or relatives not living with you”, “Friends”, “Work colleagues”, “Neighbours”, “People at your place of worship”, “People at sporting activities or clubs” and “People at service organisations (e.g. Rotary or Lions Club)”?’ The frequency ranges on a six-point scale, from ‘Less than once every three months’ to ‘Daily’. The sample average score of social participation is 2.19 on the scale of 1–6.²

(3) Neighbourhood belonging. Respondents were asked how strongly they agreed with the following three statements about their neighbourhood on a scale of 1–7, from ‘Strongly disagree’ to ‘Strongly agree’: ‘You live in a close-knit neighbourhood’; ‘You have a sense of belonging to this neighbourhood’; and ‘People around here are willing to help their neighbours.’ The average score of neighbourhood belonging is 4.65 on the seven-point scale.

(4) Social trust. Respondents were asked to what extent they agreed or disagreed with the following statements: ‘Generally speaking, most people can be trusted’; ‘Generally speaking, you can’t be too careful in dealing with people’; and ‘Generally speaking, most people would try to take advantage of you if they got a chance.’ These questions are variants of the generalised trust scale well established in previous surveys in different countries. We reversed the scores of the last two statements when recoding and found that on a scale of 1–7 the average score was 3.94.

Social Support

Social support refers to the kinds of assistance people can receive from others. We employed Resource Generators (Van Der Gaag and Snijders, 2005) to identify different types of support people receive in their daily lives (Huang and Western, 2015) by asking respondents if they had anyone to talk to for each type of support or problem out of 10 concrete scenarios when in need. Example scenarios included having problems with a partner or visiting someone for a chat. The average score is 8.75 out of 10 (see more details in Figure 2 in the Appendices).

Social Resources

Social resources are resources generated through networks. We used Position Generators (Bian and Li, 2000; Li et al., 2008; Lin, 2001) to measure various aspects of network resources. Respondents were asked if they knew anyone who had any of the occupations from a list of 20 selected occupations (Table 2). We chose occupations which cover the spectrum of occupational status measured by the Australian Socioeconomic Index 2006 (AUSEI06) (McMillan et al., 2009). These occupational status measures proxy socioeconomic resources available through networks across the occupational status hierarchy.

Table 2.

Listed occupations and status scores.

Occupation	AUSEI06
General Practitioner	100.00
University Lecturer	92.30
Solicitor/Barrister	90.70
Engineer	86.10
Primary/Secondary School Teacher	86.10
Accountant	83.70
Finance Manager	81.50
Nurse	80.70
Journalist	77.40
Police Officer	64.00
Secretary	44.80
Clerk in Public Service	54.30
Electrician	39.60
Waiter/Waitress	36.50
Farmer	34.00
Shop/Sales Assistant	30.80
Bus/Coach Driver	29.90
Chef	26.60
Cleaner	20.40
Factory Worker	12.10

We constructed four key variables: the number of occupations accessed, total status scores of occupations accessed, occupational ceiling (the highest occupational score) and occupational range (i.e. the difference between the highest and the lowest scores). These are measures of potential socioeconomic resources accessible to respondents through their networks (Bian and Li, 2000; see Table 1). In particular, the first two variables measure the volume of network resources while the other two measure the diversity of network resources.

Relational Constraints and Exclusion

We created five dummy variables to measure relational constraints and exclusion (1 = Yes): (1) Stress about social commitments. Respondents were asked ‘Over the last three months, did you feel stressed about balancing your own needs against your social commitments (e.g. spending time with friends, taking care of family members and joining activities of clubs)?’ (2) Restriction of social demands. The question asked ‘Over the last three months, do you feel your own freedom or ability to do what you want has been restricted by the demands of family or relatives, friends, acquaintances or groups you are part of?’ (3) Exclusion by social groups. Respondents were asked ‘Have you ever felt excluded by a social group?’ (4) Rules of club membership. Respondents were asked ‘Does any social group you are associated with have rules for choosing members?’ (5) Reluctance to provide help. Respondents were asked ‘Have you ever been reluctant or unwilling to provide help when asked by family, relatives, friends or acquaintances?’

Control Variables and Method of Analysis

Our analyses include demographic and socioeconomic control variables for gender, age, educational level, employment status, household income, marital status, migration status, place of residence, religion, weekly hours spent on social activities and weekly expenditure on social activities (see Table 1). These variables have been documented in previous studies on SWB (Dolan et al., 2008). To examine associations between social network variables and SWB we regress life satisfaction and happiness, respectively, on the network and control variables, using ordinary least squares (OLS) linear regressions and then compute the average marginal effects (AME) to interpret model results.³

Table 1 shows that our sample, like others, tends to under-represent men and over-represent the more highly educated. We can adjust for potential response bias associated with these variables by controlling them in our analyses. However, we cannot adjust for other potential sources of bias that are unobserved, such as respondent sociability. If more sociable people are more likely to fill in questionnaires, and also more likely to be happier or more satisfied with their lives, not including sociability will bias our results. While this is a potential problem, we also expect sociability to be highly correlated with our network variables, and since we incorporate more diverse network variables in our analysis than other researchers, we potentially also better control any sociability response bias.

Findings

Table 3 shows the impact of network characteristics on SWB. All measures of network composition, namely, the size of whole discussion network, the number of kin and non-kin in core discussion networks and the density of core discussion networks are significantly associated with life satisfaction and happiness respectively. However, the two variables measuring strength of ties, namely, reciprocity of helping and frequency of contact among the members of the core discussion network are not significantly associated with SWB.

Table 3.

Average marginal effects of network characteristics on SWB, AuSNet Survey, 2014.

Independent variables	Life satisfaction	Happiness
Network composition
Size of whole discussion network	.001!	.001!
Number of kin in core discussion network	.026***	.009!
Number of non-kin in core discussion network	.024***	.016**
Density of core discussion network	.221***	.209***
Strength of ties
Reciprocity of helping (strong = 1)	.004	−.002
Frequency of contact (high = 1)	.017	.017
Control variables^a
R²	.144	.120
Number of cases	1813	1805

Note: Two tailed test: !p < .10; *p < .05; **p < .01; ***p < .001.

Control variables are included but not presented here due to space constraints. Models with full details are presented in Table 4. The results related to control variables are very similar in both tables.

Table 4 extends the analysis by including the variables measuring relational mechanisms. This strategy helps achieve a greater understanding of the robustness of relationships between network variables and SWB as ‘variables are systematically introduced into different models’ (Dolan et al., 2008: 111). Among network characteristics, the number of kin in the core discussion network is significantly associated with life satisfaction but not happiness, and density of the core discussion network is significantly associated with life satisfaction and happiness. These results lend strong support to Hypothesis 1 that having dense networks or close social connections among network members tends to enhance life satisfaction and happiness. By contrast, the two variables measuring strength of ties are not significantly associated with either life satisfaction or happiness, implying no real support for Hypothesis 2.

Table 4.

Average marginal effects of network characteristics and relational mechanisms on SWB, AuSNet Survey, 2014.

Independent variables	Life satisfaction	Happiness
Network composition
Size of whole discussion network	.001	.001
Number of kin in core discussion network	.018**	.005
Number of non-kin in core discussion network	.010	.006
Density of core discussion network	.130**	.130***
Strength of ties
Reciprocity of helping (strong = 1)	.001	−.004
Frequency of contact (high = 1)	.010	.005
Social attachment
Having one club membership	.020	.024!
Having more than one club membership	.035*	.025!
Extent of social participation	.010	.012!
Neighbourhood belonging	.039***	.027***
Social trust	−.004	−.010
Social support
Types of perceived support	.008!	.005!
Social resources
Number of contacts in listed occupations	−.012	−.010!
Total status score of generated occupations	.268*^a	.226*^a
Highest status score of generated occupations	−.001	−.010
Range of status scores of generated occupations	.198^a	.226^a
Relational constraints and exclusion
Stress about social commitments	−.070***	−.041**
Restriction of social demands	−.042**	−.047***
Exclusion of social groups	−.071***	−.060***
Rules of club membership	−.014	−.005
Reluctance of providing help	−.033*	−.015
Control variables
Gender (male = 1)	−.038**	−.008
Age	−.010***	−.005*
Age²	.100***^a	.049*^a
Education (ref. = other education qualification)
Certificate or diploma	−.016	−.014
Bachelor degree and above	.016	.012
Employment status (ref. = employee or self-employed)
Unemployed or looking for work	−.111**	−.106**
Retired	.035	.013
Full-time student	.020	.005
Household duties	−.013	−.010
Other employment status	−.046*	−.019
Household income (ref. = high income and other)
Low income	−.066**	−.039*
Middle income	−.023	−.003
Marital status (ref. = single/never married)
Married or de facto	.093***	.084***
Divorced or separated	−.024	.017
Widowed	.059	.052
Migrant	−.010	−.359^a
Urban resident	−.030*	−.021!
Having a religion	.016	.014
Weekly hours on social activities	.002**	.001**
Weekly expenditure on social activities	.055^a	.063^a
R²	.250	.221
Number of cases	1713	1708

Note: Two tailed test: !p < .10; *p < .05; **p < .01; ***p < .001.

Coefficient multiplied by 1000.

We anticipated that social attachments, social support and social resources would be positively associated with SWB and relational constraints or exclusionary relationships would be negatively associated. Our findings are as follows: (1) Social attachment. Having club memberships and a strong sense of neighbourhood belonging are significantly associated with life satisfaction and happiness. This supports Hypothesis 3 that having strong social attachments tends to enhance life satisfaction and happiness. (2) Social support. Respondents with more types of perceived support are more satisfied and happier than respondents with fewer types of support. This lends support to Hypothesis 4 that having accessible social support tends to enhance life satisfaction and happiness. (3) Social resources. Other things being equal, we found only the total occupational status score of generated occupations is significantly associated with life satisfaction. The total score is a measure of the volume of socioeconomic resources available through networks (Bian and Li, 2000). This dimension of these occupations refers to their hierarchical positioning on a socioeconomic status (SES) continuum. The volume measure indicates that the quantity of socioeconomic resources available through networks is associated with happiness and life satisfaction, but the diversity of resources, as indicated by the status range of networks, is not. This finding offers moderate support for Hypothesis 5. (4) Relational constraints. As predicted, stress about social commitments, restrictions from social demands and exclusion by social groups are negatively associated with life satisfaction and happiness. Having been reluctant to provide help is negatively associated with life satisfaction as well. There is therefore significant support for Hypothesis 6. We summarise these results in Table 5.

Table 5.

A summary on the results of hypotheses testing.

Independent variables under assumption	Hypothesis testing
Network composition	H1 supported
Strength of social ties	H2 not supported
Social attachment	H3 supported
Social support	H4 supported
Social resources	H5 moderately supported
Relational constraints	H6 supported

Some demographic and socioeconomic factors are also significantly associated with SWB (Table 4). Women, high income earners, married and partnered individuals, those in paid employment, those spending more hours on social activities weekly and rural residents are more satisfied with their lives and/or happier than men, those on lower incomes, single people, the unemployed and those not in the labour force, those spending less time on social activities and urban residents, respectively. Older people are also less satisfied with their lives than younger people are.

Summary and Discussion

Drawing on the most recent and comprehensive national survey of Australians’ social networks (AuSNet 2014) we contribute to understanding people’s wellbeing in Australia empirically. Our results suggest that both network characteristics (i.e. features of networks associated with an individual’s network location) and network mechanisms are associated with life satisfaction and happiness. These findings imply that SWB is not explained simply by sociostructural and economic factors (Diener, 2006). Using different approaches and measures of network properties (e.g. name generators, position generators and resource generators) we have shown preliminary findings about the composition of individuals’ core discussion networks, strength of social ties, social attachment, perceived social support and social resources but also identified relational constraints and social exclusion. Overall, the analysis supports much of our proposed theoretical model, showing that network characteristics and positive and negative relational mechanisms are associated with Australians’ life satisfaction and happiness. Moreover, the cognitive and affective evaluations of SWB, life satisfaction and happiness are intertwined as social networks associations with them show similar patterns. In addition, the negative associations of social networks and SWB shed light on the study of the ‘dark side’ of social networks (Fox and Moreland, 2015; Kross et al., 2013) and help explore the ‘social ledger’, namely, both the potential benefits and liabilities of social relationships (Labianca and Brass, 2006).

Not all of our expectations were supported. For instance, the strength of social ties among the core discussion network is not significantly associated with SWB and the association between social resources and SWB is only moderately supported. How can we make sense of these results? In other research, strong and weak ties (e.g. Bian, 1997; Bian et al., 2015; Granovetter, 1995) and access to social resources (Lin, 2001) have been found to matter for status attainment. However, these network factors do not associate with SWB in our analysis. Lin (2001) differentiated two types of actions (i.e. instrumental and expressive actions) in his social resources theory. Following this classification we argue that enhancing SWB is an expressive action, which usually motivates an individual to ‘seek out others with similar characteristics and lifestyles in order to share and confide’. Comparatively, an instrumental action ‘motivates one to seek out others with dissimilar characteristics and lifestyles in order to access information and influence to achieve the expected return of more and/or better resources’ (Lin, 2001: 59). Expressive actions tend to evoke homophilous interactions (i.e. interactions with those like ourselves) while instrumental actions are inclined to be associated with heterophilous interactions (interactions with those who differ from us). The varied socioeconomic network resources measured by position generators may be more closely related to heterophilous interactions and consequently to instrumental socioeconomic processes such as status attainment than to expressive processes such as maintaining and enhancing SWB. This may explain why the diversity of social resources is not significantly associated with SWB even though the volume of social resources is. Moreover, a recent study revealed that weak ties are better able to channel information and strong ties are better able to mobilise forms of favouritism in job searches (Bian et al., 2015), which might imply that strong ties are more inclined to be involved in instrumental actions. Hence, strength of ties is probably not a significant factor associated with expressive outcomes such as SWB.

The next step in future study is to use longitudinal data where possible. Some existing longitudinal datasets incorporate some of the network variables in our analyses. In future work longitudinal data may offer some advantages for assessing causality, in cases where appropriate network variables are available. In addition, our model could also be extended by considering how it is shaped by macro-contextual factors. We consider two such factors particularly important, relational cultures and social structures. First, in some societies such as China and other East Asian countries, network relationships are tightly structured by cultural and normative expectations which emphasise interdependence and social connectedness over individualism and independence (Gold et al., 2002). In contrast, dominant cultures in western countries (e.g. Australia and the UK) are more strongly individualistic. Strong, widespread relational understandings lead to shared expectations that linked actors will behave predictably in ways that reflect collective and relational expectations and obligations. In more individualistic societies, we might expect people to behave more on the basis of individual calculation and rationality. This argument suggests that our proposed relational mechanisms would exert greater impacts on people’s SWB in a society like China than in Australia or the UK. Second, the rigidity of social boundaries is another factor that may moderate network effects. Where social boundaries are rigid, cross-boundary interactions are limited and structured. Where social boundaries are fluid, cross-boundary interactions are multifarious and varied. Australia has relatively high levels of social mobility and social barriers to the formation of friendship and marriage ties are comparatively weak (Western, 1991), so social relationships are relatively heterogeneous, providing access to more diverse network resources. We might thus expect network effects on SWB to be enhanced in less rigid societies like Australia compared to other countries with more rigid social interaction and mobility. In future research we will pursue these two macro-level hypotheses about relational culture and the rigidity of social boundaries through focused comparative contrasts of our micro-level network theory in different countries such as Australia, China and the UK.

Finally, it should be recognised that social networking sites (SNSs) play a growing role in individuals’ social networks, especially for younger people. We have not considered online social networks in this article but they offer opportunities to test our model in online as well as face-to-face networks. Previous studies of the impact of online social networks on individuals’ SWB have produced mixed results, paralleling the situation with face-to-face network research. Most research focuses on the use of Facebook, the most popular SNS globally. Some scholars suggested that there were positive associations between intensity of Facebook use and college students’ life satisfaction, social trust and civic engagement (Valenzuela et al., 2009). Facebook usage interacted with psychological wellbeing and offered greater benefits for users experiencing low self-esteem and low life satisfaction (Ellison et al., 2007). The number of Facebook friends and self-presentation strategies were found to help enhance SWB by different paths (Kim and Lee, 2011). On the other hand, it was revealed that Facebook use predicted negative effect on young adults’ SWB (Kross et al., 2013). Even though Facebook users often experienced negative emotions and psychological stress, they felt pressured to access the site frequently to keep up with relationship maintenance demands (Fox and Moreland, 2015). Our future work will attempt to collect relevant data and investigate the mechanism through which online social networks (not only about Facebook) influence SWB, positively and negatively, and to compare the roles of online versus offline social capital. This can be a promising direction of inquiry in the field of social networks and SWB.

Supplemental Material

Huang_online_appendix – Supplemental material for Social Networks and Subjective Wellbeing in Australia: New Evidence from a National Survey

Supplemental material, Huang_online_appendix for Social Networks and Subjective Wellbeing in Australia: New Evidence from a National Survey by Xianbi Huang, Mark Western, Yanjie Bian, Yaojun Li, Rochelle Côté, Yangtao Huang in Sociology

Footnotes

Acknowledgements

We are grateful to Sociology editors and anonymous reviewers for their insightful comments on the earlier draft of this article.

Funding

This research was supported by the Australian Research Council under DP130100690 ‘Social networks and subjective wellbeing in Australia, China and the United Kingdom’.

Supplementary Material

Supplementary material is available for this article online.

Notes

Xianbi Huang is Senior Lecturer in Sociology at La Trobe University, Australia. Her main research interests include social networks, social stratification, employment processes and social wellbeing. She has published articles in Sociology, Work, Employment and Society, Social Networks, American Behavioral Scientist, Research in the Sociology of Work, The China Quarterly, SAGE Research Methods Cases and other outlets. She has been involved in large scale national survey projects in Australia and China.

Mark Western is Director of the Institute for Social Science Research at The University of Queensland and Professor of Sociology. He is also a Chief Investigator on the Australian Research Council Centre of Excellence for Children and Families Over the Life Course where he directs the Disadvantage Systems Research Programme. His main areas of research include social stratification and social inequality, sociology of education and social science research methods. He has published widely in these areas and has also worked extensively with government and non-government organisations on collaborative and commissioned research projects.

Yanjie Bian is Professor of Sociology at the University of Minnesota, the United States, and Director of the Institute for Empirical Social Science Research at Xi’an Jiaotong University, China. His main research interests include social networks and social capital, social stratification and mobility, economic sociology and contemporary Chinese society. His current research is focused on explaining the increasing significance of guanxi networks in China’s market transition, with a panel study on the causal effect of network social capital on labour market outcomes in Chinese cities.

Yaojun Li is Professor of Sociology at Manchester University, UK and a Fellow of the Royal Society of Arts (FRSA). His research interests are in social mobility, social capital and the socio-economic integration of minority ethnic groups. He has published over 90 papers in these areas and conducted around 20 projects funded by academic and government agencies in Britain, USA, Australia, China and Qatar. He edited a book on Social Capital in 2015 and is doing another on Social Inequality in China with Yanjie Bian. His co-authored book with Anthony Heath on Social Mobility will be published in 2018.

Rochelle Côté is Assistant Professor of Sociology at Memorial University of Newfoundland. Broadly, her programme of research considers the variable experiences of Indigenous Peoples, ethnic minorities and immigrants as they engage with mainstream society – through the migration and incorporation experience, treatment in host societies and participation in cultural, social and economic venues. Key to her work is understanding the role of network social capital in fostering tolerance and facilitating participation.

Yangtao Huang is Postdoctoral Research Fellow at Institute for Social Science Research, The University of Queensland. His research interests include social and economic inequality and mobility, education and employment, subjective wellbeing and quantitative research methods.

References

Baffour

Haynes

Western

et al . (2016) Weighting strategies for combining data from dual-frame telephone surveys: Emerging evidence from Australia. Journal of Official Statistics 32(3): 549–578.

Becker

(1991) A Treatise on the Family. Cambridge, MA: Harvard University Press.

Bian

(1997) Bringing strong ties back in: Indirect ties, network bridges, and job searches in China. American Sociological Review 62: 266–285.

Bian

Huang

(2009) Network resources and job mobility in China’s transitional economy. In: Keister

(ed.) Research in the Sociology of Work, Volume 19. Bingley: Emerald Books, 255–282.

Bian

(2000) Social network capital of the Chinese family. Tsinghua Sociological Review 2: 1–18.

Bian

Hao

(2017) Social networks and subjective well-being: A comparison of Australia, Britain, and China. Journal of Happiness Studies. Epub ahead of print 1 November 2017. DOI: https://doi.org/10.1007/s10902–017–9926–2.

Bian

Huang

Zhang

(2015) Information and favoritism: The network effect on wage income in China. Social Networks 40: 129–138.

Bourdieu

(1986) The forms of capital. In: Richardson

(ed.) Handbook of Theory and Research for the Sociology of Education. New York: Greenwood, 241–258.

Brickman

Campbell

(1971) Hedonic relativism and planning the good society. In: Apley

(ed.) Adaptation Level Theory. New York: Academic Press, 287–302.

10.

Burt

(1984) Network items and the General Social Survey. Social Networks 6: 293–339.

11.

Burt

(1992) Structural Holes: The Social Structure of Competition. Cambridge, MA: Harvard University Press.

12.

Christakis

Fowler

(2009) Connected: The Surprising Power of Our Social Networks and How They Shape Our Lives. New York: Little, Brown and Company.

13.

Clark

Georgellis

Lucas

et al . (2004) Unemployment alters the set point for life satisfaction. Psychological Science 15: 8–13.

14.

Coleman

(1988) Social capital in the creation of human capital. American Journal of Sociology 94: 95–120.

15.

Davern

Cummins

Stokes

(2007) Subjective wellbeing as an affective-cognitive construct. Journal of Happiness Studies 8: 429–449.

16.

Diener

(2006) Guidelines for national indicators of subjective well-being and ill-being. Journal of Happiness Studies 7: 397–404.

17.

Diener

Emmons

Larsen

et al . (1985) The Satisfaction with Life Scale. Journal of Personality Assessment 49: 71–75.

18.

Dolan

Peasgood

White

(2008) Do we really know what makes us happy? A review of the economic literature on the factors associated with subjective well-being. Journal of Economic Psychology 29: 94–122.

19.

Durlauf

Fafchamps

(2005) Social capital. In: Aghion

Durlauf

(eds) Handbook of Economic Growth, Vol. 1. Amsterdam: Elsevier, 1639–1699.

20.

Ellison

Steinfield

Lampe

(2007) The benefits of Facebook ‘friends’: Social capital and college students’ use of online social network sites. Journal of Computer-Mediated Communication 12: 1143–1168.

21.

Fordyce

(1977) Development of a program to increase personal happiness. Journal of Counselling Psychology 24: 511–520.

22.

Fox

Moreland

(2015) The dark side of social networking sites: An exploration of the relational and psychological stressors associated with Facebook use and affordances. Computers in Human Behavior 45: 168–176.

23.

Furstenberg

Hughes

(1995) Social capital and successful development among at-risk youth. Journal of Marriage and the Family 57: 580–592.

24.

Gold

Guthrie

Wank

(2002) Social Connections in China: Institutions, Culture, and the Changing Nature of Guanxi. Cambridge: Cambridge University Press.

25.

Gottman

(1996) What Predicts Divorce? The Relationship between Marital Processes and Marital Outcomes. Hillsdale, NJ: Lawrence Erlbaum.

26.

Granovetter

(1973) The strength of weak ties. American Journal of Sociology 78: 1360–1380.

27.

Granovetter

(1995) Afterword 1994: Reconsiderations and a new agenda. In: Getting a Job. Chicago, IL: University of Chicago Press, 139–182.

28.

Hagan

McCarthy

(1995) Getting into street crime: The structure and process of criminal embeddedness. Social Science Research 24: 63–95.

29.

Hagan

MacMillan

Wheaton

(1996) New kid in town: Social capital and the life course effects of family migration on children. American Sociological Review 61: 368–385.

30.

Hauberer

(2011) Social Capital Theory: Towards a Methodological Foundation. Wiesbaden: Springer Fachmedien Wiesbaden GmbH.

31.

Headey

(2007) The set-point theory of wellbeing needs replacing – on the brink of a scientific revolution? Discussion Paper 753, German Institute for Economic Research, Berlin.

32.

Helliwell

(2006) Wellbeing, social capital and public policy: What’s new? Economic Journal 116(510): C34–C45.

33.

Helliwell

Putnam

(2004) The social context of well-being. Philosophical Transactions of the Royal Society 359(1449): 1435–1446.

34.

Honneth

(1996) The Struggle for Recognition: The Moral Grammar of Social Conflicts. Cambridge, MA: MIT Press.

35.

Huang

Western

(2011) Social networks and occupational attainment in Australia. Sociology 45(2): 269–286.

36.

Huang

Western

(2015) Social capital and life satisfaction in Australia. In: Li

(ed.) The Handbook of Research Methods and Applications on Social Capital. Cheltenham: Edward Elgar, 225–241.

37.

Huxhold

Fiori

Windsor

(2013) The dynamic interplay of social network characteristics, subjective well-being, and health: The costs and benefits of socio-emotional selectivity. Psychology and Aging 28(1): 3–16.

38.

Kadushin

(2012) Understanding Social Networks: Theories, Concepts, and Findings. New York: Oxford University Press.

39.

Kim

Lee

(2011) The Facebook paths to happiness: Effects of the number of Facebook friends and self-presentation on subjective well-being. Cyberpsychology, Behavior, and Social Networking 14(6): 359–364.

40.

Kross

Verduyn

Demiralp

et al . (2013) Facebook use predicts declines in subjective well-being in young adults. PLoS ONE 8(8): e69841.

41.

Labianca

Brass

(2006) Exploring the social ledger: Negative relationships and negative asymmetry in social networks in organizations. Academy of Management Review 31(3): 596–614.

42.

(2016) Social mobility, social network and subjective wellbeing in the UK. Contemporary Social Research 11(2–3): 222–237.

43.

Pickles

Savage

(2005) Social capital and social trust in Britain. European Sociological Review 21(2): 109–123.

44.

Savage

Warde

(2008) Social mobility and social capital in contemporary Britain. British Journal of Sociology 59(3): 391–411.

45.

Lin

(2001) Social Capital: A Theory of Social Structure and Action. Cambridge: Cambridge University Press.

46.

Lyons

(2011) The spread of evidence-poor medicine via flawed social-network analysis. Statistics, Politics, and Policy 2(1): Article 2.

47.

McMillan

Beavis

Jones

(2009) The AUSEI06: A new socioeconomic index for Australia. Journal of Sociology 45: 123–149.

48.

McPherson

Smith-Lovin

Brashears

(2006) Social isolation in America: Changes in core discussion networks over two decades. American Sociological Review 71: 353–375.

49.

Manfreda

Bosnjak

Berzelak

et al . (2008) Web surveys versus other survey modes: A meta-analysis comparing response rates. International Journal of Market Research 50: 79–104.

50.

Marsden

(1987) Core discussion networks of Americans. American Sociological Review 52: 122–131.

51.

Millar

Dillman

(2011) Improving response to web and mixed-mode surveys. Public Opinion Quarterly 75: 249–269.

52.

Portela

Neira

Salinas-Jimenez

(2013) Social capital and subjective wellbeing in Europe: A new approach on social capital. Social Indicators Research 114: 493–511.

53.

Portes

(1998) Social capital: Its origins and applications in modern sociology. Annual Review of Sociology 24: 1–24.

54.

Putnam

(2000) Bowling Alone. New York: Simon & Schuster.

55.

Robins

(2015) Doing Social Network Research: Network-Based Research Design for Social Scientists. London: SAGE.

56.

Scott

(2017) Social Network Analysis. London: SAGE.

57.

Valenzuela

Park

Kee

(2009) Is there social capital in a social network site? Facebook use and college students’ life satisfaction, trust, and participation. Journal of Computer-Mediated Communication 14: 875–901.

58.

Van Der Gaag

Snijders

(2005) The resource generator: Social capital quantification with concrete items. Social Networks 27: 1–29.

59.

Western

(1991) Class structure and demographic class formation. In: Baxter

Emmison

Western

et al . (eds) Class Analysis and Contemporary Australia. Melbourne: Macmillan, 166–201.

60.

Western

Tomaszewski

(2016) Subjective wellbeing, objective wellbeing and inequality in Australia. PLoS ONE 11(10): e0163345.

61.

Western

Huang

Bian

et al . (2016a) National Survey on Social Networks and Subjective Wellbeing in Australia (AuSNet Survey), 2014. Computer file, Institute for Social Science Research, The University of Queensland.

62.

Western

Huang

et al . (2016b) Technical report for the National Survey on Social Networks and Subjective Wellbeing in Australia (AuSNet Survey 2014), 2013–2016. Institute for Social Science Research, The University of Queensland.

63.

Wilkinson

(2002) Liberty, fraternity, equality. International Journal of Epidemiology 31: 538–543.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.13 MB