Upload
phamthuy
View
216
Download
1
Embed Size (px)
Citation preview
GHENT UNIVERSITY
Faculty of Medicine and Health Sciences
Academic year 2011-2012
THE INFLUENCE OF PSYCHOSOCIAL CONSTRUCTS ON THE RELATIONSHIP
BETWEEN SOCIAL CAPITAL AND MENTAL HEALTH
Part one: scientific article
Master thesis presented to obtain the degree of
Master of Health Education and Health Promotion
By Alison Taylor
Promotor: Prof. Dr. Sara Willems
Co-promotor: Drs. Veerle Vyncke
3
Putting the pieces together: Social support and resilience as intermediate variables
in the relationship between social capital and mental health.
Authors:
Vyncke, Veerlea (MA)
Taylor, Alisonb (MA)
Hardyns, Wimc (PhD.)
Pauwels, Lievend (Prof. Dr.)
Willems, Sarae
(Prof. Dr.)
abce University of Ghent, faculty Medicine and Health Science, department of General Medical Practice
and Primary Care, 1K3, De Pintelaan 185, 9000 Ghent, Belgium. ([email protected];
[email protected]; [email protected]; [email protected])
acResearch Foundation Flanders, Egmontstraat 5, 1000 Brussel.
d University of Ghent, faculty Study of Law, department of Criminal Law and Criminology,
Universiteitstraat 4, 9000 Ghent, Belgium. ([email protected])
Corresponding author: Vyncke Veerle, University of Ghent, faculty Medicine and
Health Science, department of General Medical Practice and Primary Care, 1K3, De
Pintelaan 185, 9000 Ghent, [email protected]; (tel) +3293326082;
(fax) +3293326082
Acknowledgements: The authors wish to thank the Research Foundation Flanders for the
financial support to the SWING-Survey. Furthermore, we would like to thank the City
of Ghent for the contribution to the sample selection, the students of University Ghent
for obtaining the interviews and the participants of Ghent.
5
Abstract
Social capital is one of the determinants that is believed to contribute to a better mental
health (Conrad, 2010; De Silva et al., 2005; De Silva, 2006). Although the relationship
between social capital and mental health has extensively been explored, the underlying
mechanisms largely remain unknown (Rostila, 2011; Thoits, 2011). This study
examines the role of social support and resilience as intermediate variables in the
relationship between social capital and mental health. Data from the SWING survey
2011, which collected data in 50 neighbourhoods in Ghent (Belgium) are used
(N=1024).
The Baron & Kenny method (1986) and the Sobel test (1982) are used to answer the
research question. The analyses showed that both social support and resilience mediate
the relationship between having many social ties in the low social class (high level of
number of accessed positions in low social class) and a better mental health. The Sobel
test (1982) shows that the association between social capital and mental health via
social support is significant for all social capital variables (number of accessed
positions, number of accessed positions in high social class, number of accessed
positions in low social class and occupational status of alters in people’s network).
In sum, this study mainly identifies social support as intermediating variable in the
relationship between indicators of social capital and mental health, whereas the results
concerning resilience are less unambiguous. Further research should explore the role of
additional determinants of mental health as intermediate variables in this association
(Fisher & Baum, 2010). Moreover, longitudinal studies are recommended to examine
the causal pathways between social capital and mental health (Thoits, 2011).
6
Further research is needed to inform policy makers and developers of health
interventions on the importance of psychosocial constructs for mental health and the
pathways linking social capital to mental health.
References:
Almedom A.M. (2005). Social capital and mental health. An interdisciplinary review of
primary evidence. Social Science & Medicine, 61, 943-964.
Baron, R.M., & Kenny, D.A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic and statistical considerations.
Journal of Personality & Social Psychology, 51, 1173-1182.
Conrad, D. (2010). Social capital and men’s mental health. In D. Conrad & A. White
(Eds.), Promoting men’s mental health, pp. 26-38.
De Silva, M.J. (2006). Systematic review of the methods used in studies of social
capital and mental health. In: McKenzie K., Harpham, T. (Eds). Social capital and
mental health (pp. 39-67).London: Jessica Kingsley.
De Silva, M.J., McKenzie, K., Harpham, T., & Huttly, S.R.A. (2005). Social capital and
mental illness: a systematic review. Journal of epidemiology and community health,
59(8), 619-627.
Fisher, M., & Baum, F. (2010). The social determinants of mental health: implications
for research and health promotion. Australian and New Zealand Journal of
Psychiatry, 44, 1057-1063.
Rostila, M. (2011). A resource-based theory of social capital for health research: can it
help us bridge the individual and collective facets of the concept? Social Theory &
Health, 9(2), 109-129.
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural
equation model. In S. Leinhardt (Ed.), Sociological Methodology, pp. 290-312.
Washington DC: American Sociological Association.
Thoits, P.A. (2011). Mechanisms linking social ties and support to physical and mental
health. Journal of Health and Social Behavior, 52(2), 145-161.
Keywords: social capital, social networks, mental health, psychosocial constructs,
social support, resilience, position generator, social resource
7
Introduction
Mental health problems are one of the top three causes of life-years lost to disability
worldwide (McKenzie & Harpham, 2006). According to the Belgian Health Interview
Survey of 2008, a quarter of the Belgian population experiences mental health
complaints (Gisle, 2008). In1946, the World Health Organisation has defined health as
“a state of complete physical, mental and social well-being and not merely the absence
of disease or infirmity” (WHO, 1946). This definition reflects the evolution towards a
holistic approach of health, which acknowledges mental health as a valuable aspect of
population health (Morrens, 2008) and underlines the role of psychosocial constructs as
determinants of health.
Psychosocial constructs are intrinsic and extrinsic elements that can enhance well-being
either directly or indirectly by reducing the negative effects of stressful incidents in
daily and crisis situations (Ensel & Lin, 1991).
Social capital is one of these psychosocial constructs. In the past decennia, researchers
have used the term ‘social capital’ to refer to a myriad of aspects of the social
environment. Although a clear definition of the concept is still lacking, most definitions
of social capital include one of the following three elements: norms of reciprocity, trust
and social networks (Ferlander, 2007). Social capital has repeatedly been associated
with different aspects of mental health (De Silva et al., 2005; De Silva, 2006; Yip et al.,
2007; Hamano et al., 2010; House & Kahn, 1985, in Harpham et al., 2002; Hawe &
Shiell, 2000 in Giordano & Lindström, 2011; Heaney & Israel, 2008; Kawachi &
Berkman, 2001) and is thought to protect against common mental disorders, such as
8
anxiety and depression (De Silva et al., 2007; Fujiwara & Kawachi, 2008; Stafford et
al., 2008; Webber & Huxley, 2007).
Although some researchers have tried to explain the relationship between social capital
and (mental) health (e.g. Cohen & Wills, 1985; Berkman et al., 2000), only a few
explore the theoretical pathways of the mechanisms involved (Rostila, 2011; Thoits,
2011).
This study aims to contribute to this missing evidence, by exploring the role of two
other psycho-social constructs, being social support and resilience, in the relationship
between social capital and mental health. The lack of clarity in the operationalisation of
social capital hinders researchers in drawing clear conclusions about the concept and the
way in which it operates (Derose & Varda, 2009). A social resource based
operationalisation of the concept has been suggested as a manner in which researchers
can focus on the core of social capital, namely resources in social networks (Lin, 2000;
Rostila, 2011; Van der Gaag, 2005). Within this operationalisation, social capital is
defined as “the resources embedded in one’s social networks, resources that can be
accessed or mobilized through ties in networks” (Lin, 2001, p.73)”. A social resource
based approach to social capital could be of particular interest for this study as it is
believed to enable the analysis of the mechanisms between social capital and health
(Rostila, 2011)
Social support
One of the mechanisms that might link social capital to mental health is social support
(Conrad, 2010; Song & Lin, 2010). Social support refers to the perceived and/or
received aid from network members (Song, 2010).
9
Although social capital and social support are both relationship-based determinants of
health and very much related (Song, 2010; Song & Lin, 2009), they are not identical
(Almedom, 2005). Research has shown that social capital is associated to health,
beyond the effect of social support (Song & Lin, 2009; Haines et al, 2011; Verhaeghe et
al, 2012). Following a resource based operationalisation of social capital, social capital
captures the resources possessed by alters in one’s network, available to the individual
through investment in his social network (Song, 2010). Thus, social capital can clearly
be distinguished from social support, which refers to different forms of aid individuals
receive or perceive from their network members (Berkman 1984; House 1981).
Social support is a fundamental determinant of self-reported health (Poortinga, 2006b),
physical health and mental health above all (e.g. Gazzangiga & Heatherton, 2003;
Thoits, 2011; Withley & McKenzie, 2005). The association between social support and
mental health is well-established. Social support provides emotional stability, which
contributes to mental health (Caplan, 1974 in Withley & McKenzie, 2005). Moreover,
literature shows that social support can buffer the negative effects of general life stress
on depression (Kim et al., 2005; Thoits, 2011). Some studies have also explored the
indirect relationship between aspects of social networks or social capital and mental
health via social support; social support has been included in theoretical models that
link aspects of social networks to health (Berkman et al., 2000) (Cohen & Wills, 1985).
Furthermore, a recent Belgian study showed that social support partially explains the
association between social capital and self-rated health (Verhaeghe et al., 2012).
10
Resilience
Social support is not the only pathway through which social capital might influence
mental health (Berkman et al. 2000; Conrad 2010). Resilience is another psychosocial
construct which might play a role in this relationship.
Antonovksy originally introduced the concept of resilience with the term ‘sense of
coherence’ (SOC) (1979, in Antonovsky & Sagy, 1986). He defined SOC as the degree
to which someone experiences his life as ordered, predictable and manageable.
Someone with a strong SOC is less inclined to experience stressful events as threatening
and frightening. Resilience is currently described as a set of attitudes and behaviours
associated with adaptive coping strategies (Windle et al., 2011). People with a high
level of resilience are characterized by a higher level of mastery, pro-social behaviour,
more self-esteem and a higher level of optimism (Rutter, 1985; Cederblad, 1996;
Lamand et al., 2008).
First of all, a high level of resilience has been related to better mental health (Connor et
al., 1999; Troy & Mauss, 2011), higher subjective well-being (Burns et al., 2011) and
more positive adaptive behaviours towards negative life experiences (e.g. coping with
stress) (Charney 2004; Aspinwall & MacNamara, 2005; Bonanno 2004, in Norris et al.,
2008; Butler et al., 2007, in Norris et al., 2008). A low degree of resilience is also
believed to increase the vulnerability to anxiety and depression (Burns et al., 2011).
Secondly, resilience has been linked to aspects of social networks, although the
evidence on this relationship is limited. Wolf et al. (2010) found that elderly with close
social networks possess higher levels of resilience. Resilience has also been associated
to social support: the support by family, friends and acquaintances helps individuals to
11
use effective coping strategies during stressful moments (Buikstra et al., 2010).
Unfortunately, to our knowledge, no studies are available that explicitly link social
capital to resilience.
This study will be among the first to explore the relationship between social capital and
resilience.
We hypothesize that social capital is positively associated to resilience (hypothesis 1).
As literature shows that resilience is negatively related to stress (Amirkhan & Greaves,
2003; Dumont & Provost, 1999; Rutter, 1985; Tugade & Fredrickson, 2004; Lee et al.,
2011), we further suggest that higher levels of resilience lead to better mental health
(hypothesis 2).
Furthermore, based on the literature review, we hypothesize that the relationship
between social capital and mental health runs through the intermediate variables social
support (hypothesis 3) and resilience (hypothesis 4).
13
Data and methods
This study is based on the Social ecology of Well-being In Neighbourhoods in Ghent
(SWING) Survey 2011. The SWING Survey 2011 was designed to provide information
on the role of social processes, at both the neighbourhood and the individual level, in
social disparities in health at both levels of analysis. The SWING Survey 2011 is part of
a four year research project financed by the Research Foundation - Flanders (FWO).
This study was approved by the Ethical Commission of Ghent University (reference
numbers 2011/458 and 2011/676).
Sampling
The SWING Survey 2011 reached a representative sample of 1.025 inhabitants in
Ghent. Ghent is the fifth-largest city of Belgium and located in Flanders, the northern
part of Belgium . The city covers 158 km² with a population of approximately 247.000
residents and is divided into 201 ‘statistical sectors’. A statistical sector, which is
comparable to the census tract level in the Anglo-Saxon system, is the smallest
administrative level on which objective administrative data (demographic, social and
economic indicators) are available. In the current study, statistical sectors are used to
operationalize neighborhoods.
Fifty of the 201 statistical sectors in Ghent are purposely selected, striving for a
representative sample of neighborhoods with regard to population density and
deprivation level. In each statistical sector, a representative sample of inhabitants is
selected from the National Population Register. This sample is stratified based on sex,
14
age and nationality (Belgian or not Belgian). To be included in the study, respondents had to
(1) be older than 18 years, (2) have sufficient knowledge of the Dutch language to complete the
questionnaire and (3) give informed consent. People living in a residential setting (e.g. home
for the elderly, prison, etc.) could not participate in the study.
The questionnaire was partly administered face-to-face. Questions that were too
sensitive and could possibly lead to higher non-respons when administred face-to-face
(e.g. questions on income and financial difficulties, alcohol and drug use) were gathered
using a short self-administered questionnaire, that was handed over to the respondents
after completion of the face-to-face partim.
Dependent variable
The dependent variable in this study is mental health. The 5-item index on symptoms of
depression and nervousness from the MOS 36-Item Short-Form Health Survey (SF-36)
(van der Zee & Sanderman, 2011) is used to measure mental health. The respondents
are asked to rate how they felt during the past four weeks, using a 6 - point Likert scale
to answer the following items: (1) ‘Have you been a very nervous person’, (2) ‘Have
you felt so down in the dumps that nothing could cheer you up?’, (3) ‘Have you felt
calm and peaceful?’, (4) ‘Have you felt downhearted and blue?’, (5) ‘Have you been a
happy person?’. The mental health scale has a good reliability ( =0.79), which is
supported by previous research The dependent variable mental health is included in the
analyses as a dummy variable. A low group (code 0), referring to a worse mental health,
and a high group (code 1), referring to a better mental health, are distinguished by using
the median of the original variable.
15
Independent variables
The Position Generator measures social capital defined as the resources embedded in
social networks (Lin & Erickson, 2008). The Position Generator in the current study is
based on the 20-item Position Generator used in the Social Cohesion Indicators in
Flanders (SCIF) study (Hooghe et al., 2009), with some minor adjustments based on
performed cognitive interviews. The instrument lists a number of occupations and asks
whether the respondent knows someone having these specific professions. The
occupations are purposely chosen to represent divers economic disciplines and cover the
whole socio-economic spectrum (M. Van der Gaag, 2005).
The Position Generator contains information on both the volume and the composition of
network resources (Van der Gaag, 2005; Lin & Erickson, 2008).
First, the volume of network resources is measured by calculating the total number of
accessed positions (M. Van der Gaag, 2005; M. P. J. Van der Gaag & Snijders, 2005).
For each occupational position, a dummy variable was made in this study that reflected
whether or not the respondent knows someone with this occupation. The number of
accessed positions, calculated by summing up the dummy variables that reflect the
availability of different occupational positions in the respondent’s personal network,
refers to people’s network size and shows to have a good reliability across two parallel
occupational lists (P. P. Verhaeghe et al., 2013).
Secondly, the position generator gives information on the composition of network
resources. The International Socio-Economic Index of occupational status (ISEI)
(Ganzenboom et al., 1992) is used to determine the occupational prestige linked to each
occupational group. The ISEI refers to the general income and educational level of
people in a certain occupational position: each occupational group is given an ISEI
16
score based on the result of a weighted sum of the average education and the average
income of occupations in a sample of over 70 000 males aged between 21-64 in 16
countries (Ganzenboom et al., 1992; P. P. Verhaeghe et al., 2013). The ISEI-score is
determined for each of the twenty occupations of the position generator. For the
categories ‘someone who is on welfare’ and ‘someone who is unemployed’, the
occupational prestige scores determined by the researchers of the SCIF-survey were
copied (0 and 5 respectively). The mean occupational position represents the general
level of resources that are available for the respondent through his/her social contacts
and is calculated by taking the mathematical mean of the ISEI-scores of all the positions
the respondent has access to via his social network. The ISEI classification can also be
used to distinguish between alters from different socio-economic classes. This enables
two class-based measures of social capital: the number of accessed positions in high
social class and the number of accessed positions in low social class. A categorisation
that distinguishes between two broad social classes is found to have a good reliability
across two versions of the Position Generator in a parallel-test experiment (P. P.
Verhaeghe et al., 2013).
The indicators from the Position Generator are included in the analyses as categorical
variables. First, the quartiles were determined for each indicator, leading to a variable
that distinguishes between a low (first quartile / reference category), middle (second and
third quartile) and high level (fourth quartile) of social capital.
Resilience and social support are used as intermediate variables to explore the indirect
effect between the independent and dependent variables.
17
Resilience is assessed using two items ( = .74) from the Brief Resilience Scale (Smith
et al., 2008). Respondents are asked to evaluate their ability to bounce back or recover
from stress on a five point Likert scale ( ‘totally not agreed’ to ‘totally agreed’ ). This
variable is included in the analyses as a dichotomous variable that distinguishes a low
(lower than the median of the variable; code 0) and a high level (higher than the median
of the variable; code 1) of resilience.
Social support and social influence measures the respondent’s perception of received
social support and ruling social norms in their network and is based on items from the
MOS Social Support Scale (Sherbourne & Stewart, 1991) and the Resource Generator
by Lannoo & Devos (unpublished work). The scale consists of 11 items (α=0.93), and
asks how many of the respondents’ friends, family members or acquaintances ‘(1)
understand your problems?’ ‘(2) would take you to the doctor/hospital when you are too
sick to go there yourself?’ ‘(3) would let you move into their house for a week if you
temporarily could not stay at your house?’ ‘(4) would help you with a little job you
couldn't do without help, e.g. moving heavy furniture in the house?’ ‘(5) would help
with daily chores if you were sick’ ‘(6) would be able to give advice on the invoice if
you would wonder why you had to pay so much at the doctor/dentist’ ‘(7) would be able
to give you legal advice (e.g. when you have conflicts with your landlord, your boss,
local authorities, etc.)’ ‘(8) would be able to give advice in case of a conflict within your
family’ ‘(9) would encourage you to exercise (e.g. walking, dancing, riding your bike,
doing sports’ ‘(10) would encourage you to eat healthy’ ‘(11) would encourage you to
go to the doctor if you experience health problems?’. The number of forms of support
the respondent believes he/she could access through his/her network is reflected in the a
general sum score. This variable is included in the analyses as a dichotomous variable
18
that distinguishes a low (lower than the median of the variable, code 0) and a high level
(higher than the median of the variable, code 1) of social support.
All analyses are controlled for three socio-demographic variables: sex, age and income
per capita. Gender is coded as zero for men and one for women. Respondents’ age in
years is calculated based on their birth year. Respondents were asked to estimate their
total net income (including wages, salaries, benefits, child support etc.) using 13
intervals of 500 euro, ranging from 0-499 € to 10.000 € or more. The income per capita,
a measure for income that is weighted based on the household size, was calculated to be
able to compare the income between the respondents. The OECD modified equivalent
scale was used to calculate this measure (Atkinson et al., 2001).
Statistics
The Baron & Kenny method (1986) was used to examine the intermediate effect of
social support and resilience (intermediate variables I) on the relationship between
social capital (independent variable X) and mental health (dependent variable Y). Using
the software program SPSS Statistics 19.0, three logistic regression models are built for
every indicator of social capital. The choice for logistic regression models can be
attributed to distributional problems in the intermediate and dependent variables, and
the non-linear relationship between social capital and mental health. The logistic
regression models follow the theory of Baron & Kenny (Baron & Kenny, 1986) (Figure
1), which tests mediation.
19
(Please insert Figure 1 here)
Model 1 describes the association between social capital and mental health (path c).
Model 2 examines the association between the intermediate variables (social support
and resilience) and mental health. Model 3 describes the association between the
intermediate variables and mental health while controlling for social capital on the one
hand, and between social capital and mental health while controlling for the
intermediate variables on the other hand (path c’). The Baron & Kenny method tests
indirectly estimates the relationship between social capital and mental health via the
intermediate variables based on estimates of different paths in three statistical models
(Hayes, 2009). Baron & Kenny state that a variable functions as a mediator if (a) there
is a significant association between the dependent and the independent variable on the
one hand, (b) between the intermediate variable and the dependent variable on the other
hand and (c) if the significant association between the independent and dependent
variable loses its significant when the influence of the intermediate variable is
controlled for (Baron & Kenny, 1986, p. 1176).
The Sobel test (Sobel, 1982, 1986) directly quantifies the relationship between the
independent and dependent variable via the intermediate variable, rather than inferring
its existence from a set of statistical models (Hayes, 2009). This test, also known as the
product of coefficients approach, compares the strength of the indirect effect of the
independent on the dependent variable via an intermediate variable to the null
hypothesis that this effect equals zero. The indirect effect is defined as the product of
path a (the path between the independent variable and intermediate variable) and path b
20
(path between the intermediate variable and the dependent variable) (ab). The Sobel
test divides this product by its standard error. This ratio is compared to the standard
normal distribution to test for significance (Preacher & Hayes, 2004, McKinnon, 2007).
The Sobel test was calculated using the spreadsheet by Herr (2006)1.
1Spreadsheet available at http://nrherr.bol.ucla.edu/Mediation/logmed.html
21
Results
Descriptives
Table 1 gives an overview of the characteristics of the sample. The sample consists of
1024 participants, with an equal share of men and women (48.2% and 51.8%
respectively). Age ranges from 18 to 92, with a mean of 47.
(Please insert Table 1 here)
Social support as a mediator in the relationship between social capital and health
following the Baron & Kenny procedure (1986)
The Baron & Kenny method is followed for each social capital indicator separately.
Table 2 gives an overview of the statistics resulting from these analyses.
(Please insert Table 2 here)
Having a high number of contacts (NAP) in the low social class (NAP low social class)
is the only social capital indicator that is significantly associated with mental health
(b=0.428, se=0.205, p=0.037). On the other hand, all social capital variables are
significantly associated to social support. There is a significant positive association
between social support and mental health, even after controlling for the different
indicators of social capital.
When the social capital variables and social support are jointly regressed on mental
health, the association between having a high number of contacts in the low social class
(NAP low social class) and mental health is rendered non-significant (b=0.393,
22
se=0.207, p=0.058). This implies that social support partially mediates the positive
association between have a high number of contacts in the low social class (NAP low
social class) and mental health. The association between the other indicators of social
capital and mental health remains non-significant after adding social support to the
model (Table 2).
Social support as an intermediate variable in the relationship between social capital
and health: the Sobel test
Table 4 reports the findings of the Sobel test. The Sobel test shows that the indirect
effect of social capital on mental health via social support is significant for almost all of
the social capital indicators. Only for the relationship between the middle level of
number of accessed positions in the low social class (NAP low social class) and mental
health, social support is not identified as a significant intermediate variable (z-value=
1.64).
Resilience as a mediator in the relationship between social capital and health following
the Baron & Kenny procedure (1986)
Having a high level of contacts in the low social class (NAP low social class) is the only
indicator of social capital that is significantly associated with mental health (b=0.428,
se=0.205, p=0.037). Some indicators of social capital are significantly associated to
resilience: the volume of available network resources (middle and high NAP)
(bm=0.409, sem= 0.182, pm=0.024; bh=0.738, seh=0.234, ph=0.002) and having a high
number of contacts in the high social class (High NAP high social class) (b=0.467,
se=0.234, p=0.046) are associated to higher levels of resilience. Resilience and mental
23
health are positively and significantly associated, even after controlling for the different
indicators of social capital (p<0.001) (Table 3).
When the social capital variables and resilience are jointly regressed on mental health,
the association between having more contact in the low social class (high NAP low
social class) and mental health is rendered non-significant (b=0.362, se=0.213,
p=0.090). This implies that the positive relation between have a high number of
contacts in the low social class (NAP low social class) and mental health partly runs via
higher levels of resilience. The association between the other indicators of social capital
and mental health remains non-significant after adding resilience to the model (Table 3).
(Please insert Table 3 here)
Resilience as an intermediate variable in the relationship between social capital and
health: the Sobel test
The Sobel test shows that the volume of network resources is the only indicator of
social capital for which the indirect effect of social capital on mental health via
resilience is significant (middle and high NAP) (z-valuem=2.17, z-valueh=2.96) (Table
4).
(Please insert Table 4 here)
25
Discussion
Although the relationship between social capital and mental health is well-established,
few researchers have explored the pathways of the mechanisms involved in this
association (Rostila, 2011; Thoits, 2011). Literature research suggests that both social
support and resilience might play a role as intermediate variables in this relationship,
however few studies directly test these pathways. Therefore, this study explores the role
of social support and resilience as intermediate variables in the relationship between
social capital and symptoms of depression and nervousness.
Aspects of social networks have been positively related to resilience, but evidence that
links social capital to resilience is to our knowledge not available. The first hypothesis
of this study therefor is that social capital is positively associated with resilience. The
data support this hypothesis: having access to a higher number of contacts (middle NAP
and high NAP) is related to having a higher level of resilience. This is in accordance
with earlier findings by Wolf and colleagues (2010) who suggest that strong social
networks contribute to a higher level of resilience in elderly during a heat wave.
Furthermore, having more contacts within the high social class (high NAP high social
class) is related to higher levels of resilience.
Secondly, the analyses show that higher levels of resilience are associated with a better
mental health (hypothesis 2). A possible explanation to this finding might be found in
the negative relation between resilience and stress that is reported in literature
(Amirkhan & Greaves, 2003; Dumont & Provost, 1999; Rutter, 1985; Tugade &
Fredrickson, 2004; Lee et al., 2011).
26
The main goal of this study was to explore the role of social support and resilience in
the relationship between social capital and mental health. The Baron & Kenny method
suggests that the positive relationship between a high number of contacts in the low
social class and mental health is partially mediated by both social support (hypothesis 3)
and resilience (hypothesis 4). This means that knowing a high number of people in the
low social class is related to a better mental health status, via higher levels of perceived
social support and resilience. This finding can be considered as an addition to earlier
Belgian research that identified social support as an intermediate variable in the
relationship between having more weak ties in the low social class and self-rated health,
although in that case the association between social capital and health was positive
(Verhaeghe et al., 2012). The finding of a positive association between volume of social
resources in the low social class and mental health, is somewhat in contrast to available
literature. People in a higher socio-economic position are believed to possess better
resources, to have access to more valued social resources and to have more knowledge
on how to obtain wanted resources (Lin, 2000). Therefore, the social resource based
approach to social capital states that having access to ties higher on the social ladder
enables better goal attainment in instrumental actions such as job attainment and social
mobility (Lin & Dumin, 1986; Lin, 2001). The present findings suggest that having
access to people in the lower socio-economic strata has an important value, as it seems
that these contacts lead to higher levels of social support.
Additional to the Baron & Kenny method, the Sobel test identified social support as an
intermediate variable in the relationship between having access to more social resources
in general, having a high level of accessed position in the low social class, having a high
27
and middle level of contacts in the high social class and a higher and middle level of
occupational status within one’s network on the one hand and mental health on the other
hand. Resilience was only identified as an intermediate variable in the relationship
between having access to more social resources and mental health.
In sum, the analyses are fairly unanimous on the role of social support as an
intermediate variable in the relationship between social capital and mental health. This
is in accordance to research findings of Verhaeghe et al. (2012) who found that social
support partially mediates the association between diverse indicators of social capital
and self-rated health. The findings regarding the role of resilience are more
contradicting.
Strengths & limitations
The SWING survey has several strengths. Firstly, the survey uses both international
(e.g. Resource Generator, Van der Gaag & Snijders, 2004; MOS Social Support Survey,
Sherbourne & Stewart, 1991) and national (e.g. Social Cohesion in Flanders Survey)
validated instruments. Secondly, an expert panel was consulted during the drafting of
the questionnaire. Additionally, the draft was pretested by means of 11 cognitive
interviews. Thirdly, bias was minimized by redirecting sensitive questions to a self-
administered questionnaire. Moreover, the sample selection and data collection
followed a strict protocol, promoting the representativeness of the sample.
However, some limitations in the study call for caution when interpreting the results.
For instance, the followed operationalisation of social capital represents just one of the
prevailing visions on the concept in literature. Furthermore, the cross-sectional design
28
of this study hinders insight in causal relationships (Kawachi & Berkman, 2001). It is
possible that the relationship between social capital and mental health goes from mental
health to social capital instead of the proposed way, a situation which is also referred to
as ‘reverse causality’ (Thoits, 2011). Literature suggests that the current mental health
state of people influences their social capital (Thoits, 2011), as mental health problems
often have a negative influence on social functioning, resulting in a downward ‘social
drift’ (Goldberg & Morris 1963; Jones et al. 1993; in Cullen & Whiteford, 2001).
Although the method of Baron & Kenny is the most widely used method to assess
mediation, this approach has been criticized (MacKinnon et al., 2007). Baron & Kenny
consider a significant association between the dependent and independent variable to be
a prerequisite for mediation (Preacher & Hayes, 2004). However, this is critiqued in
literature, as the requirement of a significant relationship between X and Y is believed
to reduce the power to detect mediation especially in the case of complete mediation
(MacKinnon et al., 2007). The Sobel-test, based on the product of coefficients, has been
suggested to have higher power to detect indirect effects (MacKinnon et al., 2002).
Nevertheless, the Sobel-test requires a normal approximation of the sample distribution
of the indirect effect (Hayes, 2009), while distributional problems are often encountered
in real research situations (Kenny, 2012; Bollen & Stine, 1990; Stone & Sobel, 1990 in
Hayes, 2009). The use of more advanced techniques such as building structural
equation models, including bootstrapping, to analyse the indirect effect would be a
valuable addition to the current analyses.
29
Recommendations for further research
The current study has explored the role of social support and resilience as intermediate
variables in the relationship between social capital and mental health. However, mental
health is explained by a complex whole of different determinants, such as genetic and
physiological factors, cultural processes and macro-level determinants (Berkman, et al.,
2000, Fisher & Baum, 2010), which could explain the rather limited explained variance
by the analysed models.
Future studies should explore the influence of diverse psychosocial factors.
Furthermore, analyses that include different determinants of mental health (for instance
psychosocial, economic and genetic factors) might be particularly interesting. These
studies enable to compare the impact of diverse health determinants, which might help
both policy makers and researchers to set priorities for further research and investments
(Egan et al., 2008, Almedom, 2005; Fisher & Baum, 2010).
In 2005, De Silva and colleagues concluded their literature review on social capital and
mental health with the statement that the “current evidence is inadequate to inform the
development of specific social capital interventions to combat mental illness.” (De Silva
et al., 2005). The authors claimed that further developments in the research on social
capital were needed, such as analyses on the pathways that link social capital to mental
health and the use of longitudinal studies. This call seems more timely than ever.
The current study finds that having access to more contacts within the low social class is
beneficial for mental health, via higher levels of perceived social support. This seems in
contradiction to the theoretical framework by Nan Lin and colleagues (Lin, 2000, Lin,
2001), which states that having access to occupations higher on the socioeconomic
ladder leads gives access to better social capital, and that better social capital enhances
30
goal attainment (Lin & Dumin, 1986, Lin, 2000, inequality in social capital). However,
this theoretical framework was originally developed in the context of instrumental
actions, actions aimed at gaining new resources (e.g. job attainment). In contrast,
maintaining healthy is seen as an expressive action (i.e. an action with as a goal the
protection of possessed resources) (Lin, 1999). Attention should be paid to expanding
the theoretical framework to specific and substantiated hypotheses on the role of social
resources for expressive actions. Furthermore, literature has suggested the possibility
that the influence of social capital differs for different groups of people (e.g. different
socio-economic groups, men and women,...), and that social capital therefore could
contribute to social inequities (Lin, 2000, inequity in social capital, Song & Lin, 2009).
Further research should try to take this into account.
Conclusion
A vast amount of the Belgian population (14%) suffers from mental health problems
(Gisle, 2008). Social capital is thought to be one of the determinants that contributes to
a better mental health (Conrad, 2010). Although the relationship between social capital
and mental health is already explored, the underlying mechanisms are largely unknown
(Rostila, 2011; Thoits, 2011).
Therefore, this study explored if other psychosocial constructs, i.c. social support and
resilience, influence the relationship between social capital and mental health (research
question).
31
The analyses mainly find evidence for social support as an intermediate variable
between social capital and mental health, but the findings are mixed and differ
according to the used analysis (the Baron & Kenny method (1986) versus the Sobel test
(1982)).
This study contributes to the existing literature, as it tries to unravel the pathways that
underlie the association between social capital and mental health, and includes
resilience in these analyses. In order to inform policy makers and developers of health
interventions on the importance of psychosocial constructs for mental health and the
way in which social capital can contribute to the mental health of individuals, further
research is needed.
33
References
Almedom A.M. (2005). Social capital and mental health. An interdisciplinary review of
primary evidence. Social Science & Medicine, 61, 943-964.
Amirkhan, J.H., & Greaves, H. (2003). Sense of coherence and stress: the mechanics of
a healthy disposition. Psychology and Health, 18(1), 31-62.
Antonovsky, H., & Sagy, S. (1986). The development of a sense of coherence and its
impact on responses to stress situations. Journal of Social Psychology, 126(2), 213-
225.
Aspinwall, L.G., & MacNamara, A. (2005). Taking positive changes seriously. Cancer
104, 2549-2556.
Atkinson, T., B. Cantillon, et al. (2001). Social Indicators. The EU and Social Inclusion.
Oxford: Oxford University Press.
Baron, R.M., & Kenny, D.A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic and statistical considerations.
Journal of Personality & Social Psychology, 51, 1173-1182.
Berkman, L.F. (1984). Assessing the physical health effects of social networks and
social support. Annual Review of Public Health, 5, 413-432.
Berkman, L.F., Glass, T., Brissette, I., Seeman, T.E. (2000). From social integration to
health: Durkheim in the new millennium. Social Science & Medecine, 51, 843-857.
Buikstra, E., Ross, H., Hegney, D., McLachlan, K., & Rogers-Clark, C. (2010). The
components of resilience: perceptions of an Australian rural community. Journal of
Community Psychology, 38(8), 975-991.
Burns, R.A., Anstey, K.J., & Windsor, T.D. (2011). Subjective well-being mediates the
effects of resilience and mastery on depression and anxiety in a large community
34
sample of young and middle-aged adults. Australian and New Zealand Journal of
Psychiatry, 45, 240-248.
Cederblad, M. (1996). Fifty years of epidemiologic studies in child and adolescent
psychiatry in Sweden. Nordic Journal of Psychiatry, 50, 55-66.
Charney, D.S. (2004). Psychobiological and vulnerability: implications for successful
adaptation to extreme stress. American Journal of Psychiatry, 161, 195-216.
Cohen, S, & Wills, T.A. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin, 98, 310-357.
Connor, K.M., Sutherland, S.M., Tupler, L.A., Malik., M.L., & Davidson, J.R.T.
(1999). Fluoxetine in post-traumatic stress disorder: randomised, double-blind
study. British Journal of Psychiatry, 175, 17-22.
Conrad, D. (2010). Social capital and men’s mental health. In D. Conrad & A. White
(Eds.), Promoting men’s mental health (pp. 26-38).
Cullen, M., & Whiteford, H. (2001). The interrelations of social capital with health and
mental health: discussion paper. Commonwealth of Australia: Canberra. Retrieved 4
October, 2011, from http://www.carers.health.gov.au/internet/main/publishing.nsf/
Content/63AC7D77E983E319CA257288000BCE0C/$File/intsoc.pdf
Derose, K.P. and Varda, D.M. (2009) Social capital and health care access: a systematic
review. Medical Care Research and Review, 66, 272-306.
De Silva, M.J. (2006). Systematic review of the methods used in studies of social
capital and mental health. In: McKenzie K., Harpham, T. (Eds). Social capital and
mental health (pp. 39-67).London: Jessica Kingsley.
35
De Silva, M.J., McKenzie, K., Harpham, T., & Huttly, S.R.A. (2005). Social capital and
mental illness: a systematic review. Journal of Epidemiology and Community
Health, 59(8), 619-627.
De Silva M.J., Huttly, S.R., Harpham, T., Kenward, M.G. (2007). Social capital and
mental health: a comparative analysis of four low income countries. Social Science
& Medicine, 64, 5-20.
Dumont, M., & Provost, M.A. (1999). Resilience in Adolescents: Protective role of
social support, coping strategies, self-esteem, and social activities on experience of
stress and depression. Journal of Youth and Adolescence, 28(3), 343-363.
Egan, M., Tannahill, C., Petticrew, M., & Thomas, S. (2008). Psychosocial risk factors
in home and community settings and their associations with population health and
health inequalities: a systematic meta-review. BMC Public Health, 8, 239.
Ensel, W.M., & Lin, N. (1991). The life stress paradigm and psychological distress.
Journal of Health and Social Behaviour, 32(4), 321-341.
Ferlander, S. Ferlander, S. (2007). The importance of different forms of social capital
for health. Acta Sociologica, 50(2), 115-128.
Fisher, M., & Baum, F. (2010). The social determinants of mental health: implications
for research and health promotion. Australian and New Zealand Journal of
Psychiatry, 44, 1057-1063.
Fujiwara, T., & Kawachi, I. (2008). A prospective study of individual-level social
capital and major depression in the United States. Journal of Epidemiology and
Community Health, 62, 627-633.
Ganzenboom, H.B.G., De Graaf, P.M., Treiman, D.J. (1992). A standard international
socio-economic index of occupational status. Social science research, 21(1), 1-56.
36
Gazzaniga, M.S., & Heatherton, T.F. (2003). Emotions, stress and coping. In
Psychological science: mind, brain and behaviour (pp. 310-343). New York: W.W.
Norton & Company.
Giordano, G. N., & Lindström, M. (2011). Social capital and change in psychological
health over time. Social Science & Medicine, 72, 1219-1227.
Gisle, L., (2008). Gezondheidsenquête, België 2008. Mentale gezondheid. Brussel:
Wetenschappelijk Instituut Volksgezondheid. Retrieved 20 October, 2011, from
http://www.health.belgium.be/internet2Prd/groups/public/@public/@dg1/@mentalc
are/documents/ie2divers/19073502_nl.pdf
Haines, V., Beggs, J., & Hurlbert, J. (2011). Neighbourhood disadvantage, network
social capital and depressive symptoms. Journal of health and social behaviour, 52,
58-73.
Hamano, T., Fujisawa, Y., Ishida, Y., Subramanian, S.V., Kawachi, I., & Shiwaku, K.
(2010). Social capital and mental health in Japan: a multilevel analysis. PLoS ONE,
5(10), 1-6.
Harpham, T., Grant, E. & Thomas, E. (2002). Measuring social capital within health
surveys: key issues. Health policy & planning, 17(1),106-111.
Hayes, A.F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new
millennium. Communication Monographs, 76 (4), 408-420.
Heaney, C.A., & Israel, B.A. (2008). Social networks and social support. In: K. Glanz,
B.K. Rimer, & K. Viswanath (Eds.), Health behaviour and health education:
theory, research and action, pp 185-209. San Francisco: Jossey-Bass.
37
Herr, N.R. (2006). Tools for automatically calculating logistic regression mediation,
spreadsheet. Retrieved 25 april, 2012, from http:// nrherr. bol.ucla. edu/ Mediation/
logmed.html
Hooghe, M. et al. (2009). Technical report for the Social Cohesion Survey Flanders
2009 (SCIF 2009). Leuven: KULeuven.
Kawachi, I., & Berkman, L.F. (2001). Social ties and mental health. Journal of Urban
Health,78(3), 458-467.
Kenny, D.A. (2012). Mediation. Retrieved 8 april, 2012, from http://davidakenny.net/c
m/mediate.htm
Kim, M. T., Han, H. R., Shin, H. S., Kim, K. B., & Lee, H. B. (2005). Factors
associated with depression experience of immigrant populations: A study of Korean
immigrants. Archives of Psychiatric Nursing, 19, 217-225.
Lamand, A.J., Depp, C.A., Allison, M., Langer, R., Reichstadt, J., Moore, D.J., Golshan
S, Ganiats, T.G., & Jeste, D.V. (2008). Measurement and predictors of resilience
among community-dwelling older women. Journal of Psychiatric Research, 43,
148-154.
Lannoo, S. & Devos, C. Evaluation of an Extended Version of the Resource Generator
for the Measurement of Accessed, Mobilized and Supplied Social Capital.
Lee, J.E.C., Sudom, K.A. McCreary, D.R. (2011). Higher-order model of resilience in
the Canadian forces Canadian. Journal of Behavioural Science, 43(3), 222-234.
Lin, N. (1982). Social Resources and Instrumental Action. In: P. V. Marsden & N. Lin,
Social Structure and Network Analysis (pp. 131-145). Beverly Hills, CA: Sage.
Lin, N. (1999). Building a network theory of social capital. Connections, 22 (1), 28-51.
38
Lin, N. (2001). Social Capital: A theory of social structure and action. Cambridge:
University Press.
Lin, N. (1995). Les resources sociales: une theorie du capital social. Revue francaise de
sociologie. XXXVII, 995, 685-705.
Lin, N., & Erickson B.H. (2008). Theory, measurement, and the research enterprise on
social capital. In N. Lin & B.H. Erickson (Eds.), Social capital: an international
research program, pp. 1-24. Oxford: University Press
Macinko, J., & Starfield, B. (2001). The utility of social capital in research on health
determinants. The Milbank Quarterly, 79(3), 387-427.
MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007). Mediation analysis. Annual
Review of Psychology, 58, 593- 614.
McKenzie, K., & Harpham, T. (2006). Social capital and mental health. London:
Jessica Kingsley Publishers.
Morrens, B. (2008). Sociaal kapitaal en gezondheid: een overzicht van de recente
onderzoeksliteratuur. Tijdschrift voor sociologie, 2(3), 138-157. Retrieved 13
March, 2011, from http://www.acco.be/download/nl/10204643/file/tvs 2008-jg29-2-
3 sociaal kapitaal en gezondheid een overzicht van de recente onderzoeksliteratuur.
Norris, F.H., Stevens, S.P., Pfefferbaum, B., Wyche, K.F., & Pfefferbaum, R.L. (2008).
Community resilience as a metaphor, theory, set of capacities, and strategy for
disaster readiness. American Journal of Community Psychology, 41, 127-150.
Poortinga, W. (2006b). Social relations or social capital? Individual and community
health effects of bonding social capital. Social Science & Medicine
63(1), 255-270.
39
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating
indirect effects in simple mediation models. Behavior Research Methods,
Instruments and Computers, 36,717-731.
Rostila, M. (2011). A resource-based theory of social capital for health research: can it
help us bridge the individual and collective facets of the concept? Social Theory &
Health, 9(2), 109-129.
Rutter, M. (1985). Resilience in the face of adversity: protective factors and resistance
to psychiatric disorders. British Journal of Psychiatry, 25, 173-180.
Sherbourne, C.D., & Stewart, A.L. (1991). The MOS social support survey. Social
Science and Medicine, 32(6), 705-714.
Smith, B.W., Dalen, J., Wiggins, K., Tooley, E., Christopher, P., & Bernard, J. (2008).
The Brief resilience scale: assessing the ability to bounce back. International
Journal of Behavioural Medicine, 15, 194-200.
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural
equation model. In S. Leinhardt (Ed.), Sociological Methodology (pp. 290-312).
Washington DC: American Sociological Association.
Sobel, M.E. (1986). Some new results on indirect effects and their standard errors in
covariance structure models. In N. Tuma (Ed.), Sociological Methodology (pp. 159-
186). Washington, DC: American Sociological Association.
Song, L. (2010). Social capital and psychological distress. Journal of Health and Social
Behaviour, 52 (4), 478-492.
Song, L., & Lin, N. (2009). Social capital and health inequality: evidence from Taiwan.
Journal of Health and Social Behavior, 50, 149-163.
40
Song, L. & Lin, N. (2010). Social capital and health: introduction. In N. Lin (Ed.),
Social capital. Critical concepts in the social sciences. Volume 3: Economy and
health, pp. 229-231. London: Routledge
Stafford, M., De Silva, M., Stansfeld, S., & Marmot, M. (2008). Neighbourhood social
capital and common mental disorder: Testing the link in a general population
sample. Health & Place, 14(3), 394-405.
Szreter, S., & Woolcock, M. (2004). Health by association? Social capital, social
theory, and the political economy of public health. International Journal of
Epidemioloyg, 33(4), 650-67 .
Thoits, P.A. (2011). Mechanisms linking social ties and support to physical and mental
health. Journal of Health and Social Behavior, 52(2), 145-161.
Troy, A.S., & Mauss, I.B. (2011). Resilience in the face of stress: emotion regulation as
a protective factor. In S.M. Southwick, D. Charney, & M. J. Friedman, Resilience
and mental health: challenges across the lifespan (pp. 30-44). Cambridge:
University Press.
Tugade, M.M., & Fredrickson, B.L. (2004). Resilient individuals use positive emotions
to bounce back from negative emotional experiences. Journal of Personality and
Social Psychology, 86 (2), 320-333.
Van der Gaag, M. (2005). The measurement of individual social capital. Psychological,
Pedagogical and Sociological Sciences. Groningen, University of Groningen. Ph.D
dissertation.
Van Der Gaag, M. & T. Snijders (2004). Proposals for the measurement of individual
social capital. In H. Flap & B. Volker, Creation and Returns of Social Capital (pp.
199-218). London: Routledge.
41
Van der Gaag, M.P.J. & Snijders, T.A.B. (2005). The resource generator: measurement
of individual social capital with concrete items. Social Networks, 27, 1-29.
Van der Zee, K.I., & Sanderman, R. (2011). Het meten van de algemene
gezondheidstoestand met de RAND-36, een handleiding. Noordelijk centrum voor
gezondheidsvraagstukken. Rijksuniversiteit Groningen.
Verhaeghe, P.P., Pattyn, E., Bracke, E., Verhaeghe, M., & Van De Putte, B. (2012). The
association between network social capital and self-rated health: Pouring old wine
in new bottles? Health & Place, 18, 358-365.
Verhaeghe, P.P., Van De Putte, B., & Roose, H. (2013). Reliability of position
generator measures across different occupational lists: a parallel-test experiment.
Field Methods, In press.
Webber, M., & Huxley, P. (2007). Measuring access to social capital: the validity and
reliability of the resource generator-UK and its association with common mental
disorder. Social Science & Medicine, 65, 481-492.
Whitley, R., & McKenzie, M.D. (2005). Social capital and psychiatry: review of the
literature. Harvard Review Psychiatry, 13, 71-87.
Windle, G., Bennett, K.M., & Noyes, J. (2011). A methodological review of resilience
measurement scales. Health and Quality of Life Outcomes, 9(8), 1-18.
Wolf, J., Adger, W.N., Lorenzoni, I., Abrahamson, V., & Raine, R. (2010). Social
capital, individual responses to heat waves and climate change adaptation: An
empirical study of two UK cities. Global Environmental Change, 20, 44-52.
World Health Organisation (1946). Preamble to the Constitution of the World Health
Organization as adopted by the International Health Conference, New York, 19-22
June, 1946; signed on 22 July 1946 by the representatives of 61 States (Official
42
Records of the World Health Organization, no. 2, p. 100) and entered into force on 7
April 1948. Retrieved 15 march, 2011, from http://www.who.int/about/definition/en
/print.html
Yip, W., Subramanian, S.V., Mitchell, A.D., Lee, D.T.S., Wang, & J., Kawachi, I.
(2007). Does social capital enhance health and well-being? Evidence from rural
China. Social Science & Medicine, 64, 35-49.
43
Tables
Table 1: Descriptives of all variables
N m sd Range
(min-max)
Independent variables
NAP 9.3 4.63 20 (0-20)
low level (reference category) 301
middle level 504
high level 212
NAP in high social class (NAPH) 4.9 2.66 9 (0-9)
low level (reference category) 333
middle level 488
high level 200
NAP in low social class (NAPL) 2.7 1.91 7 (0-7)
low level (reference category) 333
middle level 494
high level 192
Mean ISEI (mISEI) 47.7 9.50 79 (0-79)
low level (reference category) 253
middle level 502
high level 251
Intermediate variables
Social support 42.3 17.61 84 (0-84)
Resilience 7.9 1.58 8 (2-10)
Dependent variable
Mental health (score 100) 74.1 15.23 88 (12-100)
Confounders
Age (years) 47.1 18.34 75 (18-92)
Income per capita (euro) 1582.1 828.33 9868.42 (131.58
-10000)
Note: NAP= number of accessed positions, N=absolute number, m=mean, sd=standarddeviation
44
Table 2: Social support as a mediator in the relationship between social capital and
health following the Baron & Kenny procedure (1986)
Expl.var.(%) b se p
NAP
Model 1 (XY) 5.4
middle level of NAP -0.053 0.17 0.755
high level of NAP 0.252 0.209 0.227
Model 2 (XI) 21.6
middle level of NAP 0.906 0.184 <0.001
high level of NAP 1.470 0.225 <0.001
Model 3 (X,IY) 6.3
middle level of NAP -0.132 0.174 0.447
high level of NAP 0.126 0.214 0.557
social support 0.396 0.149 0.008
NAP high social class
Model 1 (XY) 5.2
middle level of NAPH 0.083 0.163 0.612
high level of NAPH 0.228 0.208 0.273
Model 2 (XI) 24.4
middle level of NAPH 0.990 0.176 <0.001
high level of NAPH 1.827 0.232 <0.001
Model 3 (X,IY) 6.1
middle level of NAPH -0.016 0.167 0.923
high level of NAPH 0.063 0.217 0.773
social support 0.403 0.150 0.007
NAP low social class
Model 1 (XY) 5.6
middle level of NAPL 0.060 0.158 0.704
high level of NAPL 0.428 0.205 0.037
Model 2 (XI) 17.1
middle level of NAPL 0.345 0.165 0.037
high level of NAPL 0.649 0.214 0.002
Model 3 (X,IY) 6.6
middle level of NAPL 0.028 0.159 0.858
high level of NAPL 0.393 0.207 0.058
social support 0.382 0.145 0.009
Mean ISEI
Model 1 (XY) 5.7
middle level of mISEI 0.295 0.172 0.087
45
high level of mISEI 0.277 0.204 0.175
Model 2 (XI) 17.5
middle level of mISEI 0.719 0.183 <0.001
high level of mISEI 0.875 0.217 <0.001
Model 3 (X,IY) 6.6
middle level of mISEI 0.026 0.175 0.196
high level of mISEI 0.185 0.207 0.372
social support 0.386 0.147 0.009
Note: X= mean ISEI, Y= mental health, I= social support, expl.var.= explained variance in Y,
b=regression coefficient, se=standard error, p=significance level
46
Table 3: Resilience as a mediator in the relationship between social capital and
health following the Baron & Kenny procedure (1986)
Expl.var.(%) b se p
NAP
Model 1 (XY) 5.4
middle level of NAP -0.053 0.170 0.755
high level of NAP 0.252 0.209 0.227
Model 2 (XI) 5.0
middle level of NAP 0.409 0.182 0.024
high level of NAP 0.738 0.234 0.002
Model 3 (X,IY) 16.2
middle level of NAP -0.183 0.180 0.309
high level of NAP 0.057 0.218 0.794
resilience 1.442 0.169 <0.001
NAP high social class
Model 1 (XY) 5.2
middle level of NAPH 0.083 0.163 0.612
high level of NAPH 0.228 0.208 0.273
Model 2 (XI) 4.0
middle level of NAPH 0.187 0.174 0.282
high level of NAPH 0.467 0.234 0.046
Model 3 (X,IY) 16.0
middle level of NAPH 0.038 0.171 0.823
high level of NAPH 0.112 0.217 0.605
resilience 1.433 0.168 <0.001
NAP low social class
Model 1 (XY) 5.6
middle level of NAPL 0.060 0.158 0.704
high level of NAPL 0.428 0.205 0.037
Model 2 (XI) 3.8
middle level of NAPL 0.097 0.171 0.571
high level of NAPL 0.374 0.228 0.100
Model 3 (X,IY) 16.4
middle level of NAPL 0.041 0.165 0.803
high level of NAPL 0.362 0.213 0.090
resilience 0.382 0.145 0.009
Mean ISEI
Model 1 (XY) 5.7
middle level of mISEI 0.295 0.172 0.087
47
high level of mISEI 0.277 0.204 0.175
Model 2 (XI) 4.0
middle level of mISEI 0.338 0.180 0.061
high level of mISEI 0.290 0.219 0.186
Model 3 (X,IY) 16.5
middle level of mISEI 0.224 0.181 0.215
high level of mISEI 0.207 0.214 0.333
resilience 1.430 0.170 <0.001
Note: X= mean ISEI, Y= mental health, I=resilience, expl.var.= explained variance in Y,
b=regression coefficient, se=standard error, p=significance level
48
Table 4: Social support as an intermediate variable in the relationship between
social capital and health: the Sobel-test
Middle group High group
Social Support
Number of accessed positions 2.339 2.462
Number of accessed positions in high social class 2.424 2.542
Number of accessed positions in low social class 1.638 1.989
Mean ISEI 2.183 2.200
Resilience
Number of accessed positions 2.173 2.958
Number of accessed positions in high social class 1.066 1.943
Number of accessed positions in low social class 0.566 1.611
Mean ISEI 1.833 1.308
Note: p<0.05(standard normal distribution, df=1)
49
Figures
X Y
I
X Y
Figure 1: The Baron & Kenny steps (1986). Note: X= independent
variable, Y=dependent variable, I= intermediate variable
a b
c
c’