Content uploaded by Christofer Edling
Author content
All content in this area was uploaded by Christofer Edling on Apr 18, 2016
Content may be subject to copyright.
Faith or Social Foci? Happiness, Religion, and
Social Networks in Sweden
Christofer Edling
1,2,
*, Jens Rydgren
3
and Love Bohman
3
Abstract: In this article, we study 19-year-olds in Sweden (n¼2,942) with and without an immigration
background (Iran, Yugoslavia, and Sweden). We follow-up on a recent study, which shows that religion
and happiness tend to be positively associated at the individual level only in countries with high
aggregate levels of religiosity and proposes that what affects happiness is not religiosity per se but
conformity to the standard in one’s country. We take these results a step further and study the
relationship between religion and happiness across immigrant groups that have significantly different
experiences of religion. Are we more likely to find a positive association between religion and happiness
among young Swedes with parents born in Iran and Yugoslavia than among those with two Sweden-
born parents? And do these associations depend on their sense of affiliation with Sweden? We argue
that there are strong theoretical reasons to assume that previous results also apply to the observed
association between religious networks and happiness, and we study to what extent previous results can
be generalized to societies like Sweden, which has a very low aggregate level of religiosity, and whether
that effect differs by immigration background. The results show that religion and religiousness per se
have little impact on happiness. In particular, we find that social networks tend to be positively associated
with happiness, and that this effect is driven by co-organizational membership among friends.
Introduction
During the past 20 years, academic interest in the
connection between religion and well-being has grown
substantially (Ellison, 1998;Cotton et al., 2006;Green
and Elliot, 2010).
1
Many studies demonstrate a positive
relationship between a high degree of religious partici-
pation and lower prevalence of symptoms of depression
and anxiety and higher scores on self-rating indicators of
good mental health and life satisfaction (Koenig, 2001;
Strawbridge et al., 2001;Sawatzky et al., 2005;Zullig
et al., 2006;Sternthal et al., 2010). A number of studies
have focused specifically on the relationship between
religion and religious activity, suggesting that religious
activity is a source of happiness (e.g. Borooah, 2006;
Inglehart, 2010) and well-being (e.g. Helliwell, 2003,
2006;Helliwell and Putnam, 2004). However, in a recent
study, Eichhorn (2012) demonstrated that religion and
happiness tend to be positively associated at the
individual level only in countries with high aggregate
levels of religiosity.
2
The proposed reason for this is that
what affects happiness is not religiosity per se but
conformity to the standard in one’s country. A question
not addressed by Eichhorn is the relationship between
religion and happiness across immigrant groups that
have significantly different experiences of religion. If we
follow this reasoning, it follows that the likelihood of
finding positive associations between religion and hap-
piness in countries with low aggregate levels of religiosity
is larger for immigrants originating from highly religious
contexts than for the native population. The reason is
that highly religious immigrants may have an alternative
reference group (or country) and thus entice happiness
from conforming to their country of origin (i.e. a
country with a high level of aggregate religiosity) rather
than to their country or residence (i.e. a country of low
level of aggregate religiosity).
In this article, we address these questions by using
unique data on 19-year-olds in Sweden who were
sampled according to immigration background (Iran,
Yugoslavia, and Sweden). Despite the fact that the State
and the Evangelical Lutheran Church did not officially
1
Department of Sociology, Lund University, Box 114, 22221 Lund, Sweden;
2
Stellenbosch Institute for Advanced
Study, 7600 Stellenbosch, South Africa;
3
Department of Sociology, Stockholm University, 106 91 Stockholm,
Sweden. *Corresponding author. Email: christofer.edling@soc.lu.se
European Sociological Review VOLUME 0 NUMBER 0 2014 1–12 1
DOI:10.1093/esr/jcu062, available online at www.esr.oxfordjournals.org
ßThe Author 2014. Published by Oxford University Press. All rights reserved.
For permissions, please e-mail: journals.permissions@oup.com. Submitted: June 2012; revised and accepted: July 2014.
European Sociological Review Advance Access published August 7, 2014
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
separate until January 2000, Sweden has one of the
lowest levels of aggregate religiosity in the world
(Zuckerman, 2008;Eichhorn, 2012), with less than a
quarter of the population claiming to believe there is a
God (Eurobarometer, 2005). At the same time, Swedes
report very high levels of life satisfaction, with 90 per
cent claiming they are ‘overall satisfied with life’
(Eurobarometer, 2005). Sweden also has a relatively
large presence of immigrants, many of whom come from
regions and countries where religion is of central
importance in public and private life, including the
two groups that will be analysed in this study.
3
Lim and Putnam (2010) recently suggested that the
association between religion and life satisfaction can be
explained by social networks within the congregation.
4
The argument is that social networks that build on
religious faith are more important for people’s life
satisfaction than other social network ties because people
tend to find social exchange more meaningful when it
comes from someone with whom they share a core set of
values. Assuming that the result regarding life satisfac-
tion can be extended to happiness (see footnote 1), and
drawing on Eichhorn’s (2012) findings, we argue that
there are strong reasons to assume that religious
networks and happiness are positively associated only
in countries with high aggregate levels of religiosity, such
as the United States. Thus, we study the extent to which
Lim and Putnam’s (2010) results can be generalized to
societies that have a very low aggregate level of
religiosity, such as Sweden. Moreover, there are reasons
to expect differences between individuals with and
without an immigrant background, and we expect that
social ties based in religion are more important for
immigrants coming from countries with higher levels of
aggregate religiosity.
Religion and Happiness
Two main explanations have been posited for the
observed correlation between religion and happiness,
one focusing on the private and subjective dimensions of
religion, and one focusing on religion’s role in facilitat-
ing access to social networks and support (see Lim and
Putnam, 2010). For the former set of explanations,
religion is believed to give inner peace and meaning and
a sense of belonging, which in turn increases happiness
(Greeley and Hout, 2006). Religion is also believed to
offer a framework for interpreting one’s surroundings
and for ascribing meaning and purpose to an unpre-
dictable world (Emmons et al., 1998;Inglehart, 2010).
The second set of explanations argues that religion
functions as a source of solidarity and source of mutual
identification in society (Durkheim, 1976 [1915]),
emphasizing that religious activity largely consists of,
and is a focus for, social relations (Simmel, 1905). From
this perspective, religion is positively linked to happiness
since religious organizations give people opportunities to
meet with and connect to like-minded people, thus both
catering to homophily preferences and furthering social
homophily (e.g. McPherson et al., 2001). Lim and
Putnam (2010) found that the number of friends a
person has in a congregation explains the link between
religious service attendance and life satisfaction. This link
could not be explained by considering the number of
friends per se, even if the number of friends marginally
reduced the effect of service attendance on life satisfac-
tion. When we take the number of friends within the
congregation into account, the association between
religious service attendance and life satisfaction was
reduced to almost zero. The authors concluded that
social networks that build on religious faith are distinct
from other social networks (Lim and Putnam, 2010).
The reason is that social networks that build on religious
faith link people with a shared sense of identity,
something that would explain the lack of support for
the idea of social networks as a mediating factor in
previous research, which did not distinguish between
religious and secular friendship ties (Ellison et al., 1989;
Greeley and Hout, 2006). Hence, the proposed mechan-
ism is that people tend to find social network ties more
meaningful when they link to someone with a shared
social identity (Ellison and George, 1994;Underwood,
2000;Lim and Putnam, 2010).
However, churches and other religious organizations
are not the only foci that can potentially generate strong
ties between individuals who share mutual identities.
Such contacts are also likely to be generated in
nonreligious voluntary organizations, and there are
strong reasons to believe that in countries with low
levels of aggregate religiosity such organizations are even
more important than religious ones. It is not likely that
the strong connection between religion and life satisfac-
tion mediated through congregational social networks
that Lim and Putnam (2010) find in the United States
can be generalized to societies such as Sweden, where the
overall level of organization is very high but largely
nonreligious (Rothstein, 2001;Putnam and Campbell,
2010).
5
Research on the relationship between religion and
happiness among immigrants is fairly limited. Still, there
are a few studies on this topic, and they show a positive
association between religious participation and well-
being among immigrants (Harker, 2001;Ellison et al.,
2009;Connor, 2012). Most of these studies analyse the
United States, although Connor (2012) studied the
relationship between religion and well-being of
2EDLING, RYDGREN AND BOHMAN
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
immigrants in Australia, Western Europe, and the
United States. His results indicate that immigrants
from different regions of origin experience different
levels of well-being. Religious involvement has a small
positive effect on well-being, at least on immigrants in
the United States and Western Europe, and immigrants’
involvement in non-religious group activities seems to
have a smaller effect on well-being than religious
involvement. For the United States, only religious
involvement has a positive effect on well-being, whereas
involvement in non-religious activities show a slight
negative association with well-being. These findings are
consistent with Lim and Putnam’s claim that in the
United States, congregations stand out as the most
dominant and widespread organizational foci. However,
for immigrants in Western Europe, involvement in both
religious and non-religious activities affects well-being
positively (Connor, 2012).
Data and Measurement
We use data from the Swedish survey Social Capital and
Labor Market Integration (SC09), in which a telephone
interview was conducted between October and December
2009 on a sample of 19-year-olds (n¼5,695). The SC09
sample was based on three different cohorts of Swedes
born in 1990: (i) all individuals with at least one parent
born in Iran; (ii) 50 per cent of all individuals with at
least one parent born in (former) Yugoslavia; and (iii) a
simple random sample of 2,500 individuals with two
Swedish-born parents. In total, 2,942 interviews were
completed, resulting in a response rate of 51.6 per cent.
The largest share of the non-response was the not-at-
home non-response with 37.6 per cent. The refusal rate
was 8.1 per cent.
6
Iran and the former Yugoslavia are
both major sources of immigration to Sweden.
Immigration from Iran consists primarily of political
refugees and other humanitarian migrants. Yugoslavia,
on the other hand, was the source of extensive labour
immigration in the 1960s, and more recently also of
refugees fleeing the Yugoslav wars. Since the sample was
selected based on the parents’ country of origin, the
sample includes first- and second-generation immigrants
from these regions of origin. Descriptive statistics are
summarized in Table 1, where we see that 47 per cent of
the respondents’ have both parents born in Sweden, 31.5
per cent have at least one parent born in former
Yugoslavia, and 21.5 per cent have at least one parent
born in Iran.
About 20 per cent of respondents with at least one
parent born abroad also have one parent born in Sweden.
Furthermore, 15.2 per cent of those with at least one
parent born in Iran were themselves born in Iran (and
25.8 per cent were born in Iran or Iraq), whereas as many
as 66.3 per cent of the respondents with at least one
parent born in Yugoslavia were themselves born in former
Yugoslavia. This reflects the fact that the wave of refugees
from Iran started during the late 1970s and early 1980s,
whereas the refugees from former Yugoslavia came to
Sweden during the early to mid-1990s. Most of the
respondents with immigrant backgrounds have parents
who emigrated from their countries of origin because of
conflicts that were largely religious in character: the
revolution in Iran, in 1979, which resulted in Ayatollah
Khomeini’s Islamist regime and the civil war in
Yugoslavia in the early 1990s, in which religious affiliation
was very much associated with ethnicity.
Dependent Variable
We measure happiness with the statement In general I
am happy, to which respondents could reply on a five-
level Likert-scale (‘strongly disagree’ to ‘strongly agree’)
that we treat as an ordinal scale variable. As seen in
Table 1, the mean score is 4.36, but the standard
deviation of 0.83 is non-negligible and there is sufficient
variance along the whole 5-point scale. The average
happiness score is high, as one would expect in a sample
of Swedes, but there are differences across the three
groups (not shown in Table 1); respondents with a
Yugoslav background tend to be happiest (4.43), fol-
lowed by respondents with a solely Swedish background
(4.39), and Iranian background (4.19).
7
Independent Variables
Respondents are first asked whether they are religious.
For respondents who say that they are at least somewhat
religious, religious belonging is coded in three broad
categories: Christians, Muslims, and Others. To these we
add the group ‘Not religious’, consisting of those
claiming not to be religious at all. The group ‘Others’
includes religious groups that in the sample are too small
to constitute categories of their own (e.g. Jews and
Buddhists) as well as respondents claiming to be
religious but who do not belong to any particular
religious tradition. This group contains about 3 per cent
of the respondents. The largest group—53 per cent—
consists of those who are not religious (i.e. they do not
belong to any religious tradition and they see themselves
as atheists or ‘not at all religious’), followed by
Christians (27 per cent) and Muslims (16 per cent). It
is not shown in Table 1, but among respondents with
two Swedish-born parents, 36 per cent are Christian and
62 per cent are atheists or not at all religious. The
Swedes and the Iranians are equally non-religious.
HAPPINESS, RELIGION, AND SOCIAL NETWORKS 3
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
Among those with at least one parent born in Iran, 61
per cent consider themselves to be non-religious, 21 per
cent as Muslims, 9 per cent as Christians, and just short
of 8 per cent belong to the ‘Others’ group. Those with at
least one parent born in Yugoslavia constitute the most
religious group, with only 35 per cent identifying as non-
religious. In this group, we also find the highest
percentage of Muslims (38 per cent), whereas 26 per
cent are Christian.
Immigration background is indicated by two variables,
one indicating the parents’ country of birth according to
the sampling strategy (Yugoslav and Iranian), and one
indicating whether the respondent is herself an immi-
grant to Sweden (27 per cent of the sample).
Religious activity is indicated by religious service
attendance, that is, how often respondents participate in
religious services at a church, mosque, synagogue, or
other house of worship. Previous studies have found
service attendance to be associated with happiness and
other forms of subjective well-being (Witter et al., 1985;
Cannella Jr and Lubatkin, 1993;Lim and Putnam, 2010).
In the survey, respondents could report the frequencies
‘each day’, ‘more than once a week’, ‘about once a week’,
‘more than once a month’, ‘more than once a month’,
‘more seldom’, or ‘never’. We code the answer into a
variable measuring weekly religious activity, ranging
from 1 (never) to 6 (every day).
8
Respondents were also asked a number of questions
about their self-identity, and we include an indicator
variable for religious identity that takes on the value 1 if
the respondent ranked religion as the most important
basis for identity, and 0 otherwise. We also measure
affinity with Swedish culture with the question ‘to what
extent do you feel an affinity with Swedish culture and
Swedish traditions?’ where answers were given on a 5-
level Likert scale, ranging from ‘no affinity’ (1) to ‘a
great deal of affinity’ (5). We coded answers 4 and 5 as
indicating strong affinity with Swedish culture. About 47
per cent of the respondents feel a strong affinity with
Swedish culture, and between-group differences are
surprisingly small and indicative of rather weak nation-
alist sentiments among Swedes.
9
The respondents’ social network is measured by
questions on friendship. The respondents were asked to
Table 1 Descriptive statistics
Variable nMedian Mean SD Min. Max.
Happiness
Self-reported happiness 2,870 5 4.36 0.83 1 5
Religion and identity
Muslim 2,942 0 0.16 0.37 0 1
Christian 2,942 0 0.27 0.44 0 1
Other 2,942 0 0.06 0.23 0 1
Service attendance 2,874 1 1.55 0.86 1 6
Religion important for identity 2,883 0 0.13 0.34 0 1
Affinity to Swedish culture 2,878 0 0.46 0.50 0 1
Immigrant background
At least one parent born in Yugoslavia 2,942 0 0.32 0.47 0 1
At least one parent born in Iran 2,942 0 0.22 0.41 0 1
Immigrant 2,942 0 0.26 0.44 0 1
Friendship
No. of close friends (max. 5) 2,942 5 4.07 1.19 0 5
No. of close friends religious 2,942 0 0.58 1.02 0 5
No. of close friends organizational members 2,942 1 1.19 1.32 0 5
Ln (No. of daily contacts) 2,873 3.21 3.33 1.00 0 6.21
Control variables
Sex (woman ¼1) 2,942 0 0.49 0.50 0 1
Ln (Family income, hundreds SEK) 2,928 8.38 8.26 0.64 3.69 12.04
Mark in ninth form 2,855 210 207.85 64.72 0 320
Parents education
Not finished high school 2,846 0 0.06 0.24 0 1
Finished high school 2,846 1 0.54 0.50 0 1
At least 2 years of post high school schooling 2,846 0 0.38 0.48 0 1
PhD 2,846 0 0.02 0.15 0 1
4EDLING, RYDGREN AND BOHMAN
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
name up to five close friends, and the number of friends
is used as a measure of the presence of a social network
of close friends. As seen in Table 1, the number of
friends ranges from 0 to 5, with a mean of 4.1 and a
standard deviation of 1.2. To gauge religious networks,
we used information on whether the respondents regard
their close friends as religious or not. For each alter, the
respondent was asked to what extent alter is religious.
The answer is given on a four-point scale, recoded
as non-religious (‘not at all religious’ and ‘not espe-
cially religious’) and religious (‘somewhat religious’ and
‘very religious’). On average, 0.58 out of 4.1 friends are
religious (SD ¼1.02). We also asked whether the
respondents are or have been members of the same
organization as the named alters.
10
The average respond-
ent was or had been a member of the same organization
with 1.19 of his or her friends, and there is considerable
variation here (SD ¼1.32). Finally, we include an
indicator of the number of persons whom the respond-
ent is ‘in touch with, on average, on an ordinary day’.
The raw score is 48.5 people per day (SD ¼70.7). In the
analyses, we use the log of the number of daily contacts
(Mean ¼3.33, SD ¼1).
In addition to the variables of main interests, we
include control variables measuring sex, (logged) dis-
posable family income,
11
parents’ education, and ninth-
grade marks. These variables are included since they may
interfere with the expected association between religion
and happiness (Morris, 1991). Personal income refers to
salaries and additional work-related incomes. Parents’
education is measured as the level of education for the
most educated parent, coded as a four dummy variables
(not finished high school; finished high school [‘gym-
nasium’]; at least 2 years of post-high school schooling;
PhD).
Assessing Cohort Limitations
Previous research has established a U-bend relationship
between age and happiness, such that people in middle-
age are the least happy (e.g. Blanchflower and Oswald,
2008;Stone et al., 2010; but see Wolbring et al., 2013).
The social factors that may explain this relationship
include employment, income, health, and social rela-
tionships. Because these factors vary across the life
course both in absolute and relative terms, happiness
levels can also vary across age groups. Huge variations
would be a clear indication that mechanisms and factors
at work on one age group might not impact happiness
the same way in another age group. Furthermore, some
previous studies report that older people attend religious
activities more frequently than younger (Iannaccone,
1992,1998). Other studies, however, have not been able
to establish such an association (Long and Settle, 1977).
In his review of the literature, Moberg (1972) concluded
that older people showed a decline in religious activities
outside the home but an increase in religious feelings.
The associations between age and religious activities
found in longitudinal studies (Blazer and Palmore, 1976)
seem to be due to the fact that people who seldom
attend religious services are more prone to drop out of
longitudinal studies (Markides et al., 1987). Some studies
find a U-shaped association between church attendance
and age, such that the young and the old were more
likely to attend religious services (Sullivan, 1985;Sawkins
et al., 1997). Still other studies found no association at
all (Cameron, 1999).
Since we study a cohort of 19-year-olds while still
aiming to make more general claims for the adult
population as a whole, we use the most recent World
Value Survey (WVS) data for Sweden, collected in 2006,
in order to roughly assess the impact of those age
differences in happiness and religious activity found in
previous research. Although the WVS does not contain
the rich relational information that we have access to in
SC09, it has the advantage of having been used
repeatedly in studies of both happiness and religion.
The WVS sample does not allow for analysis of single
age cohorts, so we split the WVS sample (n¼1,003) into
quartiles to assess differences between four age groups
(18–33, 34–47, 48–60, 61–85). Here we briefly present
our findings, referring the reader to the Supplementary
material for specific details on items results.
The mean happiness score is 1.62, and we find that the
age group 48–60 is somewhat less happy than the other
groups (1.68). The two youngest age groups do not
differ (both with a score of approximately 1.6), whereas
the oldest group (61–85) is the happiest (1.58), thus
suggesting a U-bend relationship. However, the differ-
ence is not significant and is driven by a disproportion-
ally larger share of ‘quite happy’ answers in the 48–60
group. For service attendance, the three youngest groups
have a mean score of roughly 6.1, whereas the oldest
groups (61–85) have a mean of 5.4, indicating a
significantly higher tendency to attend services among
the older population. To perceive oneself as a religious
person is clearly much less common among the youngest
(22 per cent) than among the oldest (48 per cent). This
tendency is linear and significant both substantially and
statistically.
However, more important is the fact that the mech-
anisms linking religion and happiness may differ across
age groups, i.e. the correlation between religiousness and
happiness might be contingent on age. We use the WVS
to assess also this challenge. Using the same age quartiles
and controlling for immigration background, the analysis
HAPPINESS, RELIGION, AND SOCIAL NETWORKS 5
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
(see Supplementary material) reveals that being religious
and attending church services does not impact happiness
for age groups 18–33, 34–47, and 48–60. However, for
the oldest quartile of the sample, age group 61–85,
service attendance substantially and significantly affects
happiness, whereas there is still no effect from being a
religious person. Keeping in mind that younger Swedes
seem less prone to attend religious services and to
consider themselves religious, the WVS analysis suggests
that results for our cohort could be extended to a
significant share of the adult population,
12
whereas there
is reason to believe that other mechanisms are at play in
the oldest quartile of the adult population.
Results
We estimated a range of models to test the association
between religion, religious activity, and happiness. In this
section, we present a selection of these results, focusing
on the weak relationship between religion and happiness
and the strong relationship between social networks and
happiness. The results are robust across a range of
specifications, including alternative binary coding of the
dependent variable and different constellations of inde-
pendent variables. In the following, we discuss only
results from Ordinal Logistic Regression models with
robust standard errors, as estimated with Stata 12.1
(StataCorp, 2011).
The first model (Model 1) gives the binary relation-
ship between happiness and the major religious faiths.
Both Christians and Muslims report a significantly
higher degree of happiness than non-religious individ-
uals. However, the link between religiousness and
happiness is not particularly robust. In the full model
(Model 6), the parameter estimates are no longer
statistically significant, and for Christians the size is
considerably smaller. In fact, even when we factor in
parents’ country of origin, the association between
religious affiliation and happiness becomes statistically
non-significant (Model 2). However, stating religious
affiliation is a relatively weak indicator of how important
religion is in a person’s life. The two variables ‘Service
attendance’ and ‘Religious identity’ are better indicators
of the importance of religion. As is evident from Models
3 and 4, service attendance has a significant positive
impact on happiness (that disappears when factoring in
social networks), whereas religious identity has no
significant effect on happiness. It is worth emphasizing
that these weak relationships are not driven by multi-
collinearity between religious affiliation and religious
activity. These null results are in line with Eichhorn
(2011). In a society like Sweden, religion plays a very
minor role. So what makes a Swede happy?
The full models (Models 5 and 6) are an indication
that happiness has a strong social dimension. The higher
the number of daily social contacts an individual has,
the greater her happiness. However, to have close friends
per se does not contribute to perceived happiness. It
turns out that close friends are important when they are
organized around a shared social focus, even if that
focus is not shared religious faith. The number of close
friends that the respondents perceive as being religious
has no impact on happiness. As it turns out, the
interaction between religious friends and religious iden-
tity is also non-significant (data not shown). Thus, the
results run counter to the assumption that religious
persons are happier when surrounded by a circle of
religious friends. Instead, we find a strong and robust
positive association between happiness and the number
of friends with whom one shares organizational
affiliation. Religious and non-religious respondents
alike are significantly happier when they have friends
belonging to the same organizations and clubs as
themselves. These results support the argument that
one’s friends have a stronger impact on happiness
if these are friends with whom one shares a sense of
social identity and belonging. However, this shared sense
of social identity does not have to be religiously
grounded.
Separate analyses of the three groups confirm the
results reported above with one particularly interesting
exception. The most notable difference between the three
groups relates to the impact of having friends with a
shared organizational affiliation. Figure 1 is a striking
illustration of this finding, showing the predicted
probabilities for happiness given the number of friends
with whom the respondent shares organizational mem-
bership. These plots are based on Model 6, estimated
separately for each group with all other variables held
constant at their mean.
As the number of friends with a shared organizational
membership increases so does the probability that the
respondent will declare that she or he is happy. The
difference between 0 and 5 co-organizational friends
amounts to an increase in the probability of happiness of
slightly more than 20 percentage points. Incidentally, this
is the same figure as Lim and Putnam (2010) reported
for congregational friendship networks. So there are
good reasons to believe that congregational friendship
networks and co-organizational friendship networks are
functionally equivalent, but that the former affect
happiness (and other forms of subjective well-being)
only in countries with high aggregate levels of religiosity,
in which congregations fulfill important personal and
social needs. As visualized in Figure 1, there is one very
important caveat to this finding: the strong linear
6EDLING, RYDGREN AND BOHMAN
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
relationship holds true only for respondents with both
parents born in Sweden.
Discussion
This article confirms results reported in Eichhorn (2011):
in a country with a low level of aggregate religiosity such
as Sweden, religion is not especially important for
happiness.
The few significant positive associations we found are
not robust and disappeared when we controlled for
parents’ country of origin. For respondents with two
parents born in Sweden, happiness is unaffected by
religion, whether measured as religious affiliation or
religious service attendance.
Unlike Lim and Putnam (2010), we failed to find an
association between the prevalence of co-religious friends
and happiness. These results did not differ significantly
between the native Swedes and the two immigrant
groups (Iranians and Yugoslavs) in our sample. Rather,
we found that friendship ties are more positively
associated with happiness if they are embedded in co-
organizational memberships. These co-organizational
memberships involve both religious and non-religious
organizations. However, in Sweden, the stock of mem-
berships in religious organizations is just a small fraction
of the total organizational involvement (Rothstein,
2001). Hence, our findings strongly indicate that it is
membership in non-religious, social organizations and
clubs that drive the positive effect between social
involvement and happiness in a country like Sweden
with a low aggregate level of religiosity, and not religious
networks. Thus, our study provides an important
addition to previous research by showing that mechan-
isms linking religion and happiness vary across contexts,
and that Eichhorn’s (2011) distinction between countries
with high and low levels of aggregate levels of religiosity
may be relevant also in this respect.
However, the finding that friendship ties are more
positively associated with happiness if they are embedded
in co-organizational memberships is context dependent.
We only found a positive association for respondents
whose parents were born in Sweden, and not for
respondents of Iranian or Yugoslavian background. We
can only speculate why this is the case. Sweden has a
tradition of a very high involvement in nonreligious
voluntary organizations (Rothstein, 2001), and one may
assume that individuals with two Swedish-born parents
are socialized into valuing organizational life more highly
compared to individuals who are more influenced by
non-Swedish traditions. Another potential explanation is
that kinship plays a more important role as a social foci
for respondents with an immigrant background, fulfilling
a function equivalent to civil organizations (and
.3 .4 .5 .6 .7 .8
0 1 2 3 4 5
Iranian born parent(s)
.3 .4 .5 .6 .7 .8
0 1 2 3 4 5
Swedish born parents
.3 .4 .5 .6 .7 .8
0 1 2 3 4 5
Yugoslavian born parent(s)
Pr(Happiness=5) given
the no. of Alters sharing
Ego’s organizational affiliation
Figure 1 Relationship between being very happy and number for friends sharing organizational membership conditional on
parents country of birth. (5 per cent confidence intervals)
HAPPINESS, RELIGION, AND SOCIAL NETWORKS 7
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
Table 2 Ordered logistic regression of happiness on religion and social networks
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Muslim 0.248* (0.109) 0.197 (0.131) 0.138 (0.134) 0.182 (0.137) 0.180 (0.141) 0.181 (0.142)
Christian 0.244** (0.087) 0.147 (0.090) 0.063 (0.098) 0.055 (0.098) 0.045 (0.100) 0.050 (0.100)
Other 0.270 (0.201) 0.162 (0.208) 0.201 (0.209) 0.185 (0.209) 0.171 (0.213) 0.156 (0.214)
Yugoslav 0.249* (0.118) 0.250* (0.117) 0.230 (0.118) 0.297* (0.121) 0.313** (0.121)
Iranian 0.402*** (0.107) 0.387*** (0.107) 0.405*** (0.107) 0.374*** (0.109) 0.317** (0.111)
Immigrant 0.116 (0.129) 0.121 (0.129) 0.092 (0.130) 0.130 (0.133) 0.133 (0.134)
Service attendance 0.104* (0.053) 0.110* (0.054) 0.095 (0.056) 0.087 (0.057)
Religious identity 0.121 (0.122) 0.088 (0.124) 0.083 (0.124)
Swedish affinity 0.293*** (0.076) 0.282*** (0.077) 0.271*** (0.078)
(ln) Daily contacts 0.120** (0.039) 0.110** (0.039)
# Friends 0.014 (0.037) 0.013 (0.037)
# Friends org. aff. 0.150*** (0.031) 0.123*** (0.032)
# Friends religious 0.024 (0.045) 0.019 (0.045)
Control variables
(sex, income, parents
education, and 9th
grade mark)
Yes
Log pseudolikelihood 2,823.48 2,807.83 2,805.78 2,797.67 2,777.38 2,767.40
n2,704 2,704 2,704 2,704 2,704 2,704
Unstandardized coefficients with standard errors in parenthesis.
*P < 0.05, **P < 0.01, ***P < 0.001
8EDLING, RYDGREN AND BOHMAN
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
religion), of embedding friendship ties in a context of
shared identity.
To conclude, we have found that in Sweden there is
no positive effect of religion on happiness. This is
equally true for respondents with both parents born in
Sweden as for individuals with an immigrant back-
ground. Hence, the expectation of finding a more
positive association between religion and happiness for
individuals originating from countries with a high
aggregate level of religiosity was not met. In particular,
there is no relationship between religious friendship
networks and happiness for any of the three groups.
However, for the group with two Swedish-born parents,
other forms of organized friendship compensate for the
lack of congregational networks that Lim and Putnam
(2010) identified as providing the positive influence of
religion on happiness. Thus, it seems to us that in a
society with a low aggregate level of religiosity, religious
belonging or service attendance will not improve hap-
piness in any way. If we tie this to Inglehart’s (2010)
conceptualization, we would say that Sweden provides a
demonstration that in modern (low-level religious)
society, traditional routes (religion) to happiness do
not work, but that some aspects of the modern way of
life (participation in nonreligious organizations) does
effectively work as a functional equivalent of traditional
behaviors and values.
Funding
This work was supported by the European Research
Council (grant number 263422) and the Swedish
Research Council for Health, Working Life, and
Welfare (FAS 2007-0806).
Supplementary Data
Supplementary data are available at ESR online.
Notes
1We study happiness empirically as measured by
agreement with the survey statement ‘In general I
am happy’. The research literature on religion that
we cite not only uses similar indicators to measure
happiness (e.g. Eichhorn, 2012), but also include
work on religion and its impact on ‘life-satisfaction’
(e.g. Lim and Putnam, 2010) and ‘well-being’ (e.g.
Connor, 2012), including ‘subjective well-being’ and
‘psychological well-being’. It can be argued that
subjective well-being encompasses both life satisfac-
tion and happiness, and that the former is a
cognitive and more inert emotion, whereas the
latter is an affective and more current emotion (see
e.g. Gamble and Ga
¨rling, 2012). Thus, there are
important conceptual hierarchies and distinctions to
consider. According to the most recent World Value
Survey data, the correlation between happiness and
life satisfaction lies consistently around 0.5 in the
developed world. In their meta-analysis of religion
and subjective well-being, Witter et al. (1985) found
that different measures (including happiness, life
satisfaction, well-being) produced consistent results
with no impact on effect size. We give the different
indicators used in previous studies equal weight in
our background discussion, being fully aware that
this is an oversimplification.
2Eichhorn (2011) analysed pooled World Value
Survey data (2000–2004 and 2005–2007 waves)
from 43 countries (Europe plus Australia, Canada,
New Zeeland, and USA; n¼74,703). For country
sample sizes, etc., see www.worldvaluessurvey.org.
Life satisfaction is measured on an ordinal scale (1
to 10). The data are analysed with a hierarchical
linear model, with individuals nested within coun-
tries, implemented in HLM6.
3In 2009, 14 per cent of the Swedish population was
born abroad and 11 per cent of those born in
Sweden had at least one parent who was born
outside Sweden (Statistics Sweden, 2010). Sweden is
the third largest recipient of asylum seekers in the
EU (Swedish Migration Board, 2010).
4Lin and Putnam (2010) analyse data from the Faith
Matters Study, a 2006 national US random sample
of people age 18 years and above targeting religious
issues. The survey had a panel structure, with a
second wave carried out in 2007. The response rate
was 53 per cent (n¼3,108) and 61.6 per cent in the
panel (n¼1,915). For further details on the survey,
and results, see Putnam and Campbell (2010). Life
satisfaction was measured on an ordinal scale (1 to
10). Lim and Putnam (2010) primarily used the
panel for robustness check, and analyzed 2006 data
by means of logistic regression, running STATA11.
5In comparison, membership in voluntary organiza-
tions in Sweden is at an internationally high level. In
1992, 92 per cent of Swedish adults belonged to a
voluntary organization, and 52 per cent considered
HAPPINESS, RELIGION, AND SOCIAL NETWORKS 9
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
themselves active members (Rothstein, 2001).
Although this share had decreased somewhat by
the year of 2000, a comparison of memberships in
voluntary organizations shows that Sweden is
ranked second out of 35 countries (Baer, 2007). At
the same time, however, only 1.8 per cent of the
population were active members of the (Lutheran)
State Church of Sweden, and the corresponding
figure for Christian free churches was less than 1 per
cent (Rothstein, 2001). Unfortunately, there are no
reliable figures for active membership in Muslim
organizations.
6The non-response analysis carried out by Statistics
Sweden suggests that there is a slight under-repre-
sentation of 19-year-olds with low grades and
poorly educated parents (Lo
¨fgren, 2010). All inter-
views were conducted in Swedish. Language did not
cause problems: interviewers reported very low
Swedish proficiency in only 0.1 per cent of the
interviews.
7A one-sample t-test reveals that the mean happiness
score for respondents with at least one
Yugoslavian or Iranian born parent differs signifi-
cantly from the grand mean. We do not believe this
is an indication of cultural differences in their
parents’ countries of origin (i.e. based on our own
calculations on 2005 World Value Survey data,
Slovenians and Serbs are significantly less happy
than Swedes).
8Service attendance differs between Christian and
Muslim respondents (
2
[5] ¼61.51) such that a
larger share of Muslims answer that they never
participate in religious service. There are no signifi-
cant differences in the tail of the distribution, i.e.
among those who take part in religious services once
a week or more often. The difference does not affect
our results.
9A total of 46 per cent of respondents with at least
one Yugoslavian born parent have a strong affinity
with Swedish culture, compared with 47 per cent
among respondents with two Swedish born parents,
and 48 per cent among respondents with at least
one Iranian born parent.
10 These co-organizational memberships involve both
religious and non-religious organizations. In
Sweden, however, membership in religious organ-
izations is just a small fraction of the total
organizational involvement (Rothstein, 2001).
11 We have tested for different income measures and not
found any that confounds the effect of our core
independent variables (personal income has no
significant effect on happiness, while disposable
income has a significant positive effect on happiness).
12 It should be noted that we cannot disentangle age
effects and cohort effects, since we use cross-sectional
data. Moreover, it should be noted that the
comparison between our 19-year-old interviewees
and the group of 18- to 33-year-olds in the World
Value Survey is far from perfect. However, con-
structing narrower age groups in the WVS data
would leave us with too few observations.
References
Baer, D. (2007). Voluntary association involvement in
comparative perspective. In Tra
¨ga
˚rdh, L. (Ed.), State
and Civil Society in Northern Europe: The Swedish
Model Reconsidered. New York: Berghahn Books.
Blanchflower, D. G. and Oswald, A. J. (2008). Is well-
being U-shaped over the life cycle? Social Science &
Medicine,66, 1733–1749.
Blazer, D. and Palmore, E. (1976). Religion and aging in
a longitudinal panel. The Gerontologist,16, 82–85.
Borooah, V. K. (2006). How much happiness is there in
the world? A cross-country study. Applied Economics
Letters,13, 483–488.
Cameron, S. (1999). Faith, frequency and the allocation
of time: a micro level study of religious capital and
participation. Journal of Socio-Economics,28,
439–457.
Cannella, A. A. Jr and Lubatkin, M. (1993). Succession
as a sociopolitical process: Internal impediments to
outsider selection. Academy of Management Journal,
36, 763–793.
Connor, P. (2012). Balm for the soul: immigrant religion
and emotional well-being. International Migration,
50, 130–157.
Cotton, S. et al. (2006). Religion/spirituality and adoles-
cent health outcomes: a review. Journal of Adolescent
Health,38, 472–480.
Durkheim, E. (1976 [1915]). The elementary forms of
the religious life. London: George Allen & Unwin.
Eichhorn, J. (2012). Happiness for believers?.
Contextualizing the effects of religiosity on life-
satisfaction. European Sociological Review,28,
583–593.
Ellison, C. G., Gay, D. A. and Glass, T. A. (1989). Does
religious commitment contribute to individual life
satisfaction. Social Forces,68, 100.
10 EDLING, RYDGREN AND BOHMAN
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
Ellison, C. G. et al. (2009). Religious involvement and
depressive symptoms among Mexican-origin adults
in California. Journal of Community Psychology,37,
171–193.
Ellison, C. G. and George, L. K. (1994). Religious
involvement, social ties, and social support in a
Southeastern community. Journal for the Scientific
Study of Religion,33, 46.
Ellison, C. G. and Levin, J. S. (1998). The religion-health
connection: evidence, theory, and future directions.
Health Education & Behavior,25, 700–720.
Emmons, R. A., Cheung, C. and Tehrani, K.
(1998). Assessing spirituality through personal
goals: implications for research on religion and
subjective well-being. Social Indicators Research,
45, 391–422.
Eurobarometer (2005). Eurobarometer 225: Social Values,
Science and Technology. Luxembourg: Office for
Official Publications of the European Communities.
Gamble, A. and Ga
¨rling, T. (2012). The relationship
between life satisfaction, happiness, and current
mood. Journal of Happiness Studies,13, 31–45.
Greeley, A. M. and Hout, M. (2006). Happiness and
lifestyle among conservative Christians. In
Greeley, A. M. and Hout, M. (Eds.), The Truth
About Conservative Christians: What They Think and
What They Believe. Chicago: University of Chicago
Press, pp. 150–161.
Green, M. and Elliott, M. (2010). Religion, health, and
psychological well-being. Journal of Religion and
Health,49, 149–163.
Harker, K. (2001). Immigrant generation, assimilation,
and adolescent psychological well-being. Social
Forces,79, 969–1004.
Helliwell, J. F. (2003). How’s life? Combining individual
and national variables to explain subjective well-
being. Economic Modelling,20, 331–360.
Helliwell, J. F. (2006). Well-being, social capital and
public policy: what’s new? Economic Journal,116,
C34–C45.
Helliwell, J. F. and Putnam, R. D. (2004). The social
context of well-being. Philosophical Transactions of
the Royal Society B: Biological Sciences,359, 1435.
Iannaccone, L. R. (1992). Sacrifice and stigma: reducing
free-riding in cults, communes, and other collectives.
Journal of Political Economy,100, 271–291.
Iannaccone, L. R. (1998). Introduction to the economics
of religion. Journal of Economic Literature,36,
1465–1495.
Inglehart, R. F. (2010). Faith and freedom: traditional
and modern ways to happiness. In Diener, E.,
Kahneman, D. and Helliwell, J. (Eds.), International
Differences in Well-Being. New York, NY: Oxford
University Press, pp. 351–397.
Koenig, H. G., McCullough, M. E. and Larson, D. B.
(2001). Handbook of Religion and Health. Oxford:
Oxford University Press.
Lim, C. and Putnam, R. D. (2010). Religion, social
networks, and life satisfaction. American Sociological
Review,75, 914–933.
Long, S. H. and Settle, R. F. (1977). Household
allocation of time and church attendance: some
additional evidence. The Journal of Political Economy,
85, 409–413.
Lo
¨fgren, J. (2010). Kalibreringsrapport, Bilaga 1. Social
Capital: Teknisk Rapport.O
¨rebro: Statistics Sweden.
Markides, K. S., Levin, J. S. and Ray, L. A. (1987).
Religion, aging, and life satisfaction: an eight-year,
three-wave longitudinal study. The Gerontologist,27,
660.
McPherson, M., Smith-Lovin, L. and Cook, J. M. (2001).
Birds of a feather: homophily in social networks.
Annual Review of Sociology,27, 415–444.
Moberg, D. O. (1972). Religion and the aging family.
Family Coordinator,21, 47–60.
Morris, D. C. (1991). Church attendance, religious
activities, and the life satisfaction of older adults in
Middletown, USA. Journal of Religious Gerontology,8,
83–96.
Olsson, T. and Ericson, B. (2008). Fra
˚n predikanter till
aktivister: fo
¨reningsliv & samha
¨lle i utveckling.
Borla
¨nge: Bjo
¨rnen.
Putnam, R. D. and Campbell, D. E. (2010). American
Grace: How Religion Divides and Unites us. New
York: Simon and Schuster.
Rothstein, B. (2001). Social capital in the social demo-
cratic welfare state. Politics & Society,29, 207.
Sawatzky, R., Ratner, P. A. and Chiu, L. (2005). A meta-
analysis of the relationship between spirituality and
quality of life. Social Indicators Research,72, 153–188.
Sawkins, J. W., Seaman, P. T. and Williams, H. C. S.
(1997). Church attendance in Great Britain: an
ordered logit approach. Applied Economics,29,
125–134.
Simmel, G. (1905). A contribution to the sociology of
religion. The American Journal of Sociology,11,
359–376.
StataCorp. (2011). Stata Statistical Software: Release 12.
College Station, TX: StataCorp LP.
Statistics Sweden. (2010). Fo
¨dda i Sverige - a
¨nda
˚olika?
Betydelsen av fo
¨ra
¨ldrarnas fo
¨delseland [Born in
Sweden - but still different? The significance of
parents’ country of birth]. Demographic Reports
2010:2. Stockholm: Statistics Sweden.
HAPPINESS, RELIGION, AND SOCIAL NETWORKS 11
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
Sternthal, M. J. et al. (2010). Depression, anxiety, and
religious life. Journal of Health and Social Behavior,
51, 343–359.
Strawbridge, W. et al. (2001). Religious attendance in-
creases survival by improving and maintaining good
health behaviors, mental health, and social relation-
ships. Annals of Behavioral Medicine,23, 68–74.
Stone, A. A. et al. (2010). A snapshot of the age
distribution of psychological well-being in the United
States. PNAS,107, 9985–9990.
Sullivan, D. H. (1985). Simultaneous determination of
church contributions and church attendance.
Economic Inquiry,23, 309–320.
Svanberg, I. and Westerlund, D. (Eds.), (2011). Religion i
Sverige. Stockholm: Dialogos.
Swedish Migration Board (2010). Migration 2000–2010,
Rapport fra
˚n Migrationsverket 2:2010 [Report from
the Swedish Migration Board]. Norrko
¨ping: Swedish
Migration Board.
Underwood, P. W. (2000). Social support: the promise
and reality. In Rice, B. H. (Ed.), Handbook of
Stress, Coping, and Health: Implications for
Nursing Research, Theory, and Practice. London:
Sage, pp. 367–391.
Witter, R. A. et al. (1985). Religion and subjective well-
being in adulthood: a quantitative synthesis. Review
of Religious Research,26, 332–342.
Wolbring, T., Keuschnigg, M. and Negele, E. (2013).
Needs, comparisons, and adaptation: the importance
of relative income for life satisfaction. European
Sociological Review,29, 86–104.
Zuckerman, P. (2008). Society Without God: What the
Least Religious Nations Can Tell Us about
Contentment. New York: New York University Press.
Zullig, K. J., Ward, R. M. and Horn, T. (2006). The
association between perceived spirituality, religiosity,
and life satisfaction: the mediating role of self-rated
health. Social Indicators Research,79, 255–274.
12 EDLING, RYDGREN AND BOHMAN
at Lunds Universitet on September 3, 2014http://esr.oxfordjournals.org/Downloaded from
