Hoverd, W.J., & Sibley, C.G. (2013) Religion, deprivation and subjective wellbeing:
Testing a religious buffering hypothesis. International Journal of Wellbeing, 3(2), 182-196.
doi:10.5502/ijw.v3i2.5
William James Hoverd
Massey University
w.hoverd@massey.ac.nz
Copyright belongs to the author(s)
www.internationaljournalofwellbeing.org
182
ARTICLE
Religion, deprivation and subjective wellbeing:
Testing a religious buffering hypothesis
William James Hoverd · Chris G. Sibley
Abstract: The buffering effect of religion upon wellbeing has been well demonstrated; however,
the question of whether this buffering effect also extends to protecting the religious against
material hardship still requires investigation. We present a Bayesian linear regression model
showing that religious affiliation provides a protective buffer against the corrosive effects on
subjective wellbeing of living in impoverished conditions. Results from a national probability
sample tested with an objective indicator of the deprivation of participants’ local
neighbourhood, derived from census data, indicated that religious people living in deprived
neighbourhoods were higher in subjective wellbeing than their non-religious counterparts living
in those same neighbourhoods (N = 5,984 New Zealanders). It was in impoverished conditions
that the difference in wellbeing between religious and non-religious people was apparent; those
living in affluent neighbourhoods showed comparably high levels of subjective wellbeing
regardless of whether or not they were religious. Our results explore new ground by showing
for the first time that the buffering effect of religion is readily apparent within New Zealand.
Keywords: religion, subjective wellbeing, deprivation, poverty, New Zealand
1. Introduction
Religion was famously described by Karl Marx, as the ‚opium of the people‛ (Marx & Engels,
1964). Marx’s critique of capitalism implies that one modern function of religion is to protect
and insulate the religious believer’s wellbeing from some of the stark realities of capitalism. In
the World Happiness Report (Helliwell, Layard & Sachs, 2012), religion was modelled as an
external factor (community-based factor, rather than lifestyle) which can have a positive effect
upon wellbeing and provide a buffering function (protective mechanism for personal wellbeing)
in the face of stressful life events. Yet the World Happiness Report does not explicitly consider
the possibility that the link between religion and wellbeing may be moderated by deprivation
(relative material disadvantage within a society). Currently, in terms of method, most studies
have concentrated on relating religious affiliation with, either, solely wellbeing measures or
solely measures of deprivation. Indeed, most studies of deprivation and religious affiliation, or
of wellbeing and religion, do not statistically adjust for other demographics at a national level.
When deprivation is examined, it has often been constructed from a collected set of variables
involving income and education, rather than from robust national measures (Smith & Faris,
2005), such as the New Zealand Deprivation Index. For example, Dolan, Peasegood and White
(2008), in their review of wellbeing literature, indicate that religious affiliation is positively
related to wellbeing and then, in other sections of their article, note that deprivation relates to
lower wellbeing. The possible interactive effect of these variables, which is central to a
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 183
buffering hypothesis, is inferred by much of the literature; however, direct empirical tests of
this link remain scarce.
Diener, Tay and Myers (2011) documented a reliable relationship between religious
affiliation, wellbeing and deprivation across different nations and American states. They found
that people living in poorer nations or poorer American states tended to be more religious than
those living in affluent nations or affluent American states. Thus, in nations with higher
religiosity, this religiousness appeared to create forms of social support which buffered
wellbeing, indicating support for the argument that religious affiliation protects individual
wellbeing from deprivation within poorer countries (Diener et al., 2011). Diener et al. (2011)
offer a macro-level analysis of religious buffering occurring across poorer nations and a microlevel
analysis of the effect occurring within relatively impoverished American states. Diener et
al. (2011) argued that this effect occurs for three primary reasons: Firstly, religion offers social
support; secondly, religion offers an alternate explanatory frame which offers unique meaning
for one’s life; and thirdly, religion teaches respect for others regardless of personal
circumstances. Their model indicated that these three mechanisms combine to explain why
religious affiliation buffers wellbeing against material hardship; that is, it mediates the effect.
Diener et al. (2011) analysed wellbeing and religiosity compared across nations and across
American states. We seek to replicate the buffering effect that they find at the micro level in
poorer American states at the neighbourhood level in the nation of New Zealand.
1.1 Religious affiliation and wellbeing
The link between religion and wellbeing has been a topic of substantial sociological interest
over the past three decades. The link between religious group affiliation and greater wellbeing
has been demonstrated in numerous studies, across a variety of contexts (e.g., Blaine &
Crocker, 1995; Bradshaw, Ellison & Marcum, 2010; Ellison, 1991; Fabricatore, Handal, Rubio &
Gilner 2004; Fiorito & Ryan, 2007; Hackney & Sanders, 2003). Most of the research in the field
indicates that religious affiliation has a positive effect upon an individual’s wellbeing. The
seminal work of Ellison (1991, p. 80), in his review of the pre-1990s’ literature on religion and
wellbeing, found that religion may affect wellbeing in four ways: ‚1) through social integration
and support; 2) through the establishment of personal relationships with a divine other; 3)
through the provision of systems of meaning and existential coherence; and 4) through the
promotion of more specific patterns of religious organization and personal lifestyle.‛
From the 1990s, this relationship between religious affiliation and wellbeing has been
extensively documented as occuring across a variety of Western Christian groups (Cohen-Zada
& Sander, 2011; Ellison, 1991; Mochon, Norton & Ariely, 2008; Strawbridge, Shema, Cohen,
Roberts & Kaplan, 1998). However, the exact nature of how the relationship between religious
affiliation and wellbeing functions is not yet fully understood (Mochon, Norton & Ariely, 2011;
Pargament, 2002). There exists minor variation between certain Christian denominations (Tix &
Frazier, 1998). There are also clear gender differences in this relationship (Chamberlain & Zika,
1988; Cohen-Zada & Sander, 2011; Steenwyk, Atkins, Bedics & Whitley Jr, 2010). In addition,
the relationship between religion and subjective wellbeing includes both positive and negative
effects upon wellbeing which still require further exploration (Ano & Vasconcelles 2005;
Ellison, Bradshaw, Kuyel & Marcum, 2012; Strawbridge et al., 1998).
The positive relationship seems to occur cross-culturally among Christians (Clarke &
Lelkes, 2009) and also other religious groups (Ellison & Taylor, 1996), with the effect recently
being found amongst studies of various Muslim groups (Ahmad & Pevalin 2012; Ahmed, 2010;
Abolfathi, Yadollah & Nuriza, 2009). It has also been argued that a positive relationship
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 184
between religious affiliation and wellbeing can be found across most religious groups
worldwide (Diener et al., 2011; Dolan et al., 2008; Kim, 2003), as long as there is adjustment for
the various contextual influences, and specific theological worldviews of individual religious
groups.
The positive benefits of the relationship between religious affiliation and subjective
wellbeing extend to mental coping and health (Bradshaw et al., 2010; Fabricatore et al., 2004;
Fiorito & Ryan 2007; Hackney & Sanders, 2003; Smith, McCullough & Poll, 2002). In addition,
evidence exists showing that general physical health (Homan & Boyatzis, 2010) and exercise
rates (Mochon et al., 2008) can be positively affected by the positive relationship between
religion and wellbeing. The wellbeing benefits derived from religious affiliation stem from
belief giving hope, structure and meaning to people’s everyday lives (Wnuk & Marcinkowski,
2012).
In sum, the extant research indicates that there is a general, positive relationship for
wellbeing which exists for those affiliated with a religious group, and that, on average,
religious people have greater personal wellbeing than non-religious people. Moreover, the
literature suggests that this relationship is cross-cultural and can be found across a variety of
religious groups. Thus, this literature suggests that religious affiliation, which is positively
related to wellbeing, offers a coping mechanism or a buffering effect to protect believers against
both physical and social hardship.
1.2 Deprivation and wellbeing
Decades of research by economists and psychologists suggests that the relationship between
deprivation and wellbeing is far from straightforward (see Diener & Seligman, 2004; Fischer &
Boer, 2011, for discussion). In a massive representative US sample, Kahneman and Deaton
(2010) reported that earning more was linked to higher life-satisfaction (general positive
evaluations of one’s life overall). However, they also found that once a certain level of income
was reached (approximately USD 75,000), no further gains in emotional happiness accrued.
Conversely, a study of a large sample of US and European populations (Alesina, Di Tella &
McCullough, 2004) has found that social inequality, in both America and Europe, is correlated
with lower rates of happiness. The general literature indicates that deprivation is reliably
related to decreased wellbeing (Dolan et al., 2008; Helliwell, Layard & Sachs 2012).
This general body of research shows conclusively that poverty does predict decreased
wellbeing, and that, conversely, money does ‘buy’ happiness, but only up to a point. Similar
effects have also been observed in national probability samples conducted in New Zealand. For
instance, Sengupta et al. (2012) analyzed data from more than 5,000 New Zealanders, and
reported that those people who earned less tended to be less happy, less satisfied with their
lives overall and more stressed in general. Thus, we can conclude that, on average, increased
deprivation is likely to lead to decreases in personal wellbeing.
2. Overview of the present study
We model the extent to which religious affiliation (the comparison of religious and nonreligious
people) buffers the effect of poverty on subjective wellbeing, using nationally
representative data in one Western nation: New Zealand. New Zealand functions as a superb
testing ground for our hypothesis because of its small size, its religious diversity, and because it
provides the opportunity to integrate nationally representative data regarding wellbeing,
deprivation, and religious affiliation. Moreover, New Zealand offers a fresh, relatively
understudied Western population to test for the existence of the relationship between religion,
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 185
wellbeing and deprivation. Outside of the work of Waldegrave (2009; 2010), who offers
evidence to suggest that a general positive relationship between religion and wellbeing exists
within older groups of New Zealanders, there has been no study of religious affiliation and
wellbeing in the New Zealand population. There has been only a single study of the
relationship between deprivation and religious affiliation in New Zealand (Hoverd, Bulbulia &
Sibley, In Press), which found that religious affiliation did not correlate with deprivation.
New Zealand’s population has been historically predominantly Christian (Oliver, 1966).
Since the 1960s, immigration from Southeast Asia, the Pacific and the Middle East has changed
the ethnic and religious demographics of the nation (Crothers, 2005; Wilson, 1993). New
Zealand’s population should be understood to be religiously plural including various Christian
affiliations, Muslims, Buddhists, Hindus and many other smaller religious groups (Hoverd,
2008; Statistics NZ, 2006.) Increasingly, the New Zealand Government (Morris, 2007) and the
scholarly community (Hoverd & Sibley, 2010; Nachowitz, 2007; Vaccarino, Kavan & Gendall,
2011) discuss New Zealand religious affiliation by employing the term diverse to explain the
variety of religious groups within the nation.
Because of this religious diversity and the generally accepted cross-cultural applicability of
the relationship between religious affiliation and wellbeing, we measure our data across all
religious people regardless of religious group differences. We employ the measure of religious
affiliation. Our study utilizes the measures found in Hoverd, Atkinson and Sibley (2012), where
self-reported affiliations with a religious group are called religious affiliation. This clarification of
terminology is essential because other studies of religious demographics (Kosmin & Keysar,
2009; Kosmin, Mayer & Keysar, 2001) use the term ‘religious identification’ to describe the
number of individuals who affiliate with various religious groups. Following Diener et al.
(2011), we theorised that religious affiliation should provide a protective buffer against the
potentially deleterious effects on subjective wellbeing and social capital from living in
impoverished conditions.
2.1 Hypothesis
We outline our hypothesis as follows: Religious people living in deprived neighbourhoods
should be higher in subjective wellbeing than their non-religious counterparts living in the
same neighbourhoods. Moreover, it will be in impoverished conditions that the difference in
wellbeing between religious and non-religious people will be most apparent; whereas, those
living in affluent neighbourhoods should show comparatively high levels of subjective
wellbeing regardless of whether or not they are religious.
3. Method
We present a Bayesian regression model testing whether being religious insulates one’s
wellbeing against economic hardship in a large-scale New Zealand national probability sample.
We test this enduring hypothesis using data on religious affiliation drawn from the nationally
representative and longitudinal New Zealand Attitudes and Values Survey (NZAVS). The
NZAVS also utilizes a measure of subjective wellbeing, the Personal Wellbeing Index, originally
developed by Cummins, Eckersley, Pallant, Van Vugt, & Misajon (2003). (2003). In addition, we
capitalize on a novel and highly robust measure of regional deprivation in New Zealand: the
New Zealand Deprivation Index 2006 (NZDep; Salmond, Crampton & Atkinson, 2007).
The NZDep2006 is a census-based measure of deprivation derived from a principle
components analysis of nine variables linked with socio-economic status, which are calibrated
for each neighbourhood/small area unit of New Zealand in which participants are located. We
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 186
used the percentile deprivation index, which gives an ordinal score from 1 (most affluent) to 10
(most deprived) for each mesh block area unit based on 2006 census data. The index measures
relative deprivation within the nation. As a consequence, when we employ the term
‚deprivation,‛ ‚impoverished,‛ and/or ‚poverty‛ within this manuscript we are referring to
those people who are rated as living in the most deprived geographic meshblock units of New
Zealand.1 Conversely, when we use the terms ‚affluent‛ or ‚wealthy‛ we are referring to those
people who are rated as living in the least deprived geographic meshblock units of New
Zealand. Thus, we formally model the extent to which religious affiliation (the comparison of
religious and non-religious people) acts as an ‘opiate’ which buffers the effects of deprivation
on subjective wellbeing through comparison of NZAVS religious affiliation data with the
Personal Wellbeing Index and the New Zealand Deprivation Index 2006.
This research offers important insight because it documents the hypothesized buffering
effect of religion in a large national sample (N = 5,984) and uses a robust measure of religious
affiliation, deprivation and wellbeing. It offers the first analysis of its type conducted in the
New Zealand context, and the first of which we are aware to employ Bayesian estimation
methods and statistically control for a wide range of other demographic covariates. This
research also offers strong evidence for the existence of a relationship between the three
variables of religious affiliation, wellbeing and deprivation within a single nation.
3.1 Sampling procedure
The NZAVS-2009 questionnaire was posted to 40,500 participants randomly selected from the
2009 NZ electoral roll. The publicly available version of the roll contained 2,986,546 registered
voters. This represented all citizens over 18 years of age who were eligible to vote regardless of
whether or not they chose to vote, barring people who had their contact details removed due to
specific case-by-case concerns about privacy. In sum, roughly 1.36% of all people registered to
vote in New Zealand were contacted and invited to participate.
3.2 Participant details
The NZAVS-2009 sampled a total of 6,518 participants. The overall response rate (adjusting for
address accuracy of the electoral roll and including anonymous responses) was 16.6%. The low
response rate was likely due to people’s unwillingness to sign up to a 20-year study. Complete
data for the measures analyzed here were provided by 5,984 participants (the majority of
missing data was for residential location, which means that regional deprivation indices could
not be assigned to those participants).
Of the 5,984 participants analyzed here, 2,657 (44.4%) of the sample identified as religious,
and the remaining 3,327 (55.6%) were not religious. The sample comprised 3,574 (59.7%)
women and 2,410 (40.3%) men. The mean age of the sample was 47.71 years (SD = 15.53). In
terms of ethnicity, 4,305 (71.8%) of participants identified as New Zealand European and the
remaining 28.2% identified with another ethnic group (primarily Maori, Pacific, and Asian
peoples). The sample was broadly comparable to the population proportion of different ethnic
1 The Index ratings are discussed in greater detail later in the methodology section. This index is based on census
data including the level of poverty and affluence, access to transport, and household crowding, of different small
neighbourhood units (around 100 persons on average) across the entire country. As such, it is an index of
deprivation that is not based solely on reported household income or education, but is arguably more robust as an
indicator of the level of poverty in each participant’s immediate residential location (Salmond, Crampton &
Atkinson, 2007.)
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 187
groups based on 2006 census data; however, the NZAVS did oversample women relative to
men.
In terms of education, 1342 (22.4%) had no formal qualification, 1757 (29.4%) had a
secondary school qualification, 974 (16.3%) had a tertiary diploma or trade certificate, 1359
(22.7%) had an undergraduate university degree or were studying toward one, and 552 (9.2%)
had a post-graduate qualification. In terms of relationship status, 4296 (71.8%) of participants
were in a romantic relationship or married, and the remaining 1688 (28.2%) were single. In
terms of employment, 4499 (75.2%) of participants were in full or part-time employment, and
the remaining 1485 (24.8%) were not.
3.3 Deprivation index
Participants provided their postal address, and we used this information to identify the level of
economic deprivation (NZDep2006) of the immediate area in which each participant resided
(for a full discussion of NZDep2006 see: Salmond et al., 2007). The NZDep2006 allocates a
deprivation score to each meshblock based on a principal components analysis of nine
variables using census data. These are (in weighted order): proportion of adults received a
means-tested Government supplied welfare benefit, household income, proportion not owning
their own home, proportion single-parent families, proportion unemployed, proportion lacking
qualifications, proportion household crowding, proportion with no telephone access, and
proportion with no car access. The NZDep2006 thus reflects the average level of deprivation of
different small neighbourhoods or community areas across the country (Salmond et al., 2007).
We used the percentile deprivation index, which gives an ordinal score from 1 (most affluent)
to 10 (most deprived) for each mesh block area unit based on 2006 census data. The mean
percentile NZDep2006 score for our sample was 5.00 (SD = 2.84). The NZDep2006 is commonly
used in medical-related studies of the New Zealand population in order to investigate potential
associations between certain illnesses, health risks and deprivation (e.g., McKenzie, EllisonLoschmann
& Jefferys 2010; Salmond, Crampton, King & Waldegrave 2006; Utter et al., 2010).
NZDep2006 is also used as a tool to develop national social policy for the New Zealand
Government (Salmond & Crampton, 2012). As such, NZDep2006 offers a nationally consistent
measure of deprivation for the New Zealand context.
3.4 Questionnaire measures
Consistent with the method used in the NZ Census, we assessed religious affiliation by asking
‚Do you identify with a religion and/or spiritual group?‛ The 2,657 participants who answered
‘yes’ were asked to complete an open-ended field specifying their religious denomination or
group membership.
Personal Wellbeing was measured using four items from the Personal Wellbeing Index
(PWI) developed by Cummins et al. (2003). The PWI measures satisfaction with the following
aspects of life: standard of living, health, what you are achieving in life, personal relationships, safety,
feeling part of a community and future security. Participants rated their level of satisfaction with
four aspects of their life on a scale from 0 (completely dissatisfied) to 10 (completely satisfied),
and ratings were then averaged to give an overall mean score. The mean scale score was 6.61
(SD = 1.79), and the scale had good internal reliability (Cronbach’s alpha = .73). These items,
and this item format, have been previously validated for use in the New Zealand context with
different populations and provide a good general measure of social capital (Manuela & Sibley,
in press; Sibley, Harré, Hoverd & Houkamau, 2011). The four items included in the scale were:
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 188
 ‚Your standard of living.‛
 ‚Your health.‛
 ‚Your future security.‛
 ‚Your personal relationships.‛
3.5 Data analysis
To test our predictions, we conducted a Bayesian regression model predicting subjective
wellbeing using data from a national probability sample conducted in New Zealand, and
matched these data for each participant with census indicators of the level of deprivation
versus affluence of each participant’s immediate neighbourhood. Our measure of deprivation is
thus highly reliable, as it is based on census data from the immediate community region (see
the method for further details). Our model tested regression-based interactions that examined
the extent to which (a) religious affiliation (religious or not religious) moderated the link
between deprivation and subjective wellbeing. Moreover, we adjusted other demographic
factors that might also relate to subjective wellbeing. These other demographic factors were not
the focus of the present study, and were included as control variables in order to adjust for
their main effect on wellbeing.
Table 1. Descriptive statistics and frequencies for all measures and demographics.
N % M SD Min Max
Meshblock deprivation index 5984 5.00 2.84 1 10
Age 5984 47.71 15.53 18 95
Subjective wellbeing 5984 6.61 1.79 0 10
Gender
Men 2410 40.3%
Women 3574 59.7%
Religious affiliation
Religious 2657 44.4%
Non-Religious 3327 55.6%
Romantic involvement
Involved 4296 71.8%
Single 1688 28.2%
Employment
Employed 4499 75.2%
Unemployed 1688 24.8%
Ethnicity
Majority group 4305 71.7%
Minority group 1679 28.3%
Education
No formal qualification 1342 22.4%
Secondary school qualification 1757 29.4%
Tertiary diploma or trade certificate 974 16.3%
Undergraduate university degree 1359 22.7%
Post-graduate qualification 552 9.2%
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 189
A Bayesian linear regression model tested whether deprivation moderated the association
between religion and subjective wellbeing. We tested for the significant interaction between
deprivation and affiliation by creating product (interaction) terms between religious affiliation
and centered scores on the New Zealand Deprivation Index. The model used diffuse priors and
was conducted in Mplus 7.0. Bayesian analysis is superior to frequentist estimation methods for
a number of reasons (see Gelman, Carlin, Stern & Rubin, 2003). For our purposes, this approach
allowed us to calculate the probability that our effects were significant based on credibility
intervals. Yuan and MacKinnon (2009) put it nicely when discussing the difference between
Bayesian credibility intervals relative to more well-known frequentist intervals when they
commented that ‚Bayesian credible intervals have more natural probability interpretations
than CIs. A 95% credible interval means that there is a 95% chance that the credible interval
contains the true value of the parameter on the basis of the observed data‛ (p. 304). In terms of
interpretation, b coefficients from our analysis may be interpreted in much the same way as in
a standard regression model, in that they represent the line of best fit for that parameter within
the model.2
4. Results
As reported in Table 2, our regression model indicated that those who were religious were
significantly higher in personal wellbeing relative to those who were non-religious (b = .144).
Table 2: Bayesian linear regression model testing the effect of deprivation on subjective
wellbeing depending upon religious affiliation (non-religious versus religious).
95% Credibility
Interval
b
Posterior
SD
β
Lower
2.5%
Upper
2.5%
p
Constant 5.468 .111
Religious affiliation (0 no, 1 yes) .144 .046 .041 .053 .234 .001
Meshblock deprivation index (1-10) -.103 .011 -.167 -.124 -.082 .000
Deprivation x affiliation .035 .016 .036 .004 .065 .014
Gender (0 women, 1 men) -.138 .046 -.038 -.227 -.049 .001
Age .006 .002 .050 .003 .009 .000
Romantic involvement (0 no, 1 yes) .817 .050 .205 .718 .915 .000
Employment (0 no, 1 yes) .133 .056 .031 .024 .243 .009
Ethnicity (0 minority, 1 majority) .237 .061 .050 .116 .356 .000
Education (ordinal coded) .095 .018 .068 .060 .130 .000
Note: N = 5984. Model used diffuse priors; p-values were one-tailed and give the proportion of the posterior
distribution below or above 0 (for positive and negative estimates, respectively). 95% Confidence Interval for
difference between observed and the replicated χ2 values = [-13.096, 13.356]. Bayesian Information Criterion =
23437.03, Posterior Predictive P-value = .496.
People who lived in more deprived conditions also tended to be significantly lower in personal
wellbeing (b = -.103). Critically, these two effects interacted. This is shown in the significant
effect for the religion by deprivation interaction term (b = .035). This indicates that neither the
effect of religion nor that of deprivation can be considered in isolation. The difference between
2
All results were comparable when a more widely used, frequentist-based Maximum Likelihood method of
estimation was used.
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 190
religious and non-religious people in personal wellbeing depended upon people’s relative level
of deprivation. Moreover, as can also be seen in Table 2, this interaction occurred when
adjusting for the main effects of various other demographic factors, including gender, age,
relationship status, employment status, ethnicity and education.
Differences in wellbeing for religious and non-religious people living in low, moderate and
high levels of deprivation are presented in Figure 1.
Figure 1. Estimated mean levels of subjective wellbeing (measured using the Personal
Wellbeing Index) for religious and non-religious people who were low, moderate and high
in deprivation (measured using the NZ Deprivation Index).
Note: Error bars represent the posterior SD of the estimate.
As can be seen, the difference in personal wellbeing was between religious and non-religious
people who were high in deprivation—the poorer people in society. We tested this regression
interaction by calculating predicted mean values at low, moderate and high values of
deprivation (values of NZDep = 2, 5, and 8, respectively) for religious and non-religious people
based on the parameters in our Bayesian regression model. Point difference tests indicated that
people who lived in affluent regions (NZDep = 2) showed similar, and high, levels of personal
wellbeing regardless of whether or not they were religious; they did not differ significantly
from one another (b = .039, posterior SD = .066, 95% CIs = [-.089, .199], p = .066). A different
pattern emerged, however, for those high in deprivation (NZDep = 8). Here, people who were
religious and poor, fared significantly better than those who were non-religious and poor. This
difference was significant (b = .248, posterior SD = .066, 95% CIs = [.119, .376], p < .001). These
findings indicate that religious people living in deprived neighbourhoods were higher in
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 191
subjective wellbeing than their non-religious counterparts living in those same
neighbourhoods.
5. Discussion
Religion has long been recognised as having a positive link with psychological wellbeing
(Ellison, 1991). However, robust empirical evidence has been lacking when it comes to
considering whether the relationship between religion and wellbeing could be extended to
buffering against the effects of economic deprivation. We modelled the proposed associations
using data from a large nationally representative New Zealand sample using Bayesian linear
regression. As hypothesised, our analyses indicated that religious people living in deprived
neighbourhoods were higher in subjective wellbeing than their non-religious counterparts
living in the same neighbourhoods. These effects were reasonably subtle and fairly small in
effect size, but they were reliable and held when adjusting for a number of demographic
covariates, including gender, age, and education. For those people who are religious and living
in impoverished neighbourhoods, religion seems to provide a protective buffer upon
wellbeing.
A key strength of our study is that we estimated our models using a Bayesian approach,
which arguably provides more reliable, and practically interpretable, probability estimates as
compared to more well-known frequentist estimation methods used in many regression models
(see Gelman et al., 2003). Another key strength of our research is that it used a highly robust
measure of deprivation based on neighbourhood characteristics and derived from census data
on the level of average income, household crowding, access to transport and other key features
at the neighbourhood-unit level. Our findings regarding religious buffering show that the effect
occurs when one indexes poverty based on the characteristics of people’s immediate location
(in roughly 100-person units).
Diener et al. (2011) noted that in wealthy nations non-religious people tend to have higher
subjective wellbeing than they do in poorer nations. Our findings replicate this result but in a
different manner. We find a religious buffering effect for religious people living in the deprived
neighbourhoods of New Zealand, in that religious people have higher subjective wellbeing
than do non-religious people. Thus, if you are non-religious in New Zealand your subjective
wellbeing is likely to be related to your level of personal wealth (Sengupta et al., 2012).
Currently, one of the limitations of these results is that while they find evidence for the effect,
they do not explain why this buffering function exists. Diener et al. (2011) explain that religious
affiliation provides this buffering effect because it offers social support, an alternative
explanatory framework which offers unique meaning to life and that religion teaches respect
for others regardless of personal circumstances. In future periods of data collection, the NZAVS
will be able to explore and unpack the unique roles that social support, life satisfaction and
community belonging might each contribute to this buffering effect.
This study is also important in the New Zealand context, as it provides the first nationally
representative evidence that the general effects of the relationship between religious affiliation
and wellbeing exist within the nation. While other studies investigated income by religious
group (Hoverd & Sibley, 2010) or provided a measure of the relationship between religion and
wellbeing in older population segments (Waldegrave, 2009; 2010), this study supplies a muchneeded
representative measure of the relationship between religious affiliation, wellbeing and
deprivation within and across a nationally representative New Zealand sample. Our results
confirm that a beneficial relationship between religious affiliation and wellbeing exists, at a
general level, in the religiously diverse New Zealand context. And when extended to
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 192
deprivation, the paper provides important practical and social outcomes by showing that
religious New Zealanders are better psychologically buffered against economic stress than are
non-religious people. Interestingly, it is the New Zealand Churches with growing membership
that tend to have congregations with lower average incomes, than churches experiencing a
decline in affiliation (Hoverd & Sibley, 2010). We expect to be able to explore and publish
further results investigating these relationships as we collect data and add measures to the
NZAVS. Since our dataset is longitudinal, eventually the NZAVS will allow us to measure
changes over time in affiliation, wellbeing and deprivation in New Zealand. Elsewhere (Sibley
& Bulbulia, 2012), the NZAVS dataset has been able to successfully measure the longitudinal
changes in perceived health amongst religious people in Christchurch, New Zealand, both
before and after the 2011 earthquake which killed over one hundred people and left thousands
of people homeless.
When it comes to the global data which relates to religious affiliation, wellbeing and
deprivation, together, our study sampled a religiously diverse (New Zealand) population,
whereas much of the existing research relating deprivation to strength of religious
identification has centred upon North American, Christian or Muslim populations, or research
across nations. When it comes to developing a working hypothesis or theory about the
relationship between religious affiliation, wellbeing and deprivation, we want to stress that
measuring deprived material conditions and contexts in relation to wealthy contexts is central
for isolating this effect. Micro-level neighbourhood comparisons within a nation are needed to
supplement cross-national macro data. The macro-level global research comparisons offered by
Diener et al. (2011) and a recent Gallup poll (Crabtree, 2010) show the effect amongst poorer
nations.
Using a national neighbourhood deprivation measure, we show that the buffering effect of
religion on subjective wellbeing also occurs at the micro level within deprived neighbourhoods
in wealthy nations, amongst religious people regardless of their specific religious affiliation.
Thus, the comparison between poor and wealthy nations made by Diener et al. (2011) produces
evidence for an effect which our results indicate also holds in comparisons between wealthy
and poor neighbourhoods within New Zealand. As a consequence, the effect that the subjective
wellbeing of religious people is buffered against material hardship holds both at a macro level
among nations and at a micro level within nations. Finally, our neighbourhood-based data
supports Diener et al.’s (2011) conclusion that, in less religious societies with favourable
circumstances, higher subjective wellbeing is achieved by most people whether they be
religious or non-religious. Understandably, more research will be required to further pursue
the applicability of such an argument. Further analysis of cross-national and national datasets
will be required. Moreover, it is likely that there is variation to be found among different
religious groups and within particular subsets of these groups when it comes to future studies
of the relationship between religion, deprivation and wellbeing.
6. Conclusion
We present a Bayesian linear regression model showing that religious affiliation provides a
protective buffer against the corrosive effects on subjective wellbeing and social capital of
living in impoverished conditions. Results from a national probability sample of N=5984 New
Zealanders indicated that religious people living in deprived neighbourhoods were higher in
subjective wellbeing than their non-religious counterparts living in the same neighbourhoods.
It is in impoverished conditions that the difference in wellbeing between religious and nonreligious
people is apparent; those living in affluent neighbourhoods showed comparably high
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 193
levels of subjective wellbeing regardless of whether or not they were religious. Our results
demonstrate that the effect by which religious people who live in impoverished conditions are
buffered against material hardship occurs at the neighbourhood level. We show that the
religious buffering effect occurs both at a macro (across nations) and a micro level (within a
nation’s neighbourhoods). Our results also forge new ground by showing for the first time that
this effect is readily apparent in New Zealand.
To return to Marx then, in the New Zealand context, our data indicate that the so-called
‚opium of the people‛ has its effects primarily when people are hurting due to poverty and
deprivation, and that simply belonging to or affiliating with a religious group is enough to
experience the palliative effect of religion. Moreover, our data supplement existing crossnational
research to suggest that this opiate effect occurs both at the macro level across nations
and at the micro level within nations.
Acknowledgements
Collection of the New Zealand Attitudes and Values Study 2009 (NZAVS-09) data analyzed in this paper
was funded by University of Auckland FRDF (#3624435/9853) and ECREA (#3626075) grants awarded to
Chris Sibley. Preparation of this manuscript was partially funded by a Government of Canada, DFAIT
Postdoctoral Research Fellowship at the MCRI Religion and Diversity Project at University of Ottawa
awarded to William James Hoverd in the period 2011/2012. We thank Victoria University of Wellington’s
Religious Studies Department for their support during manuscript preparation. We thank Lara Greaves
for helpful comments on an earlier version of this manuscript.
Authors
William James Hoverd
Massey University
w.hoverd@massey.ac.nz
Chris G. Sibley
University of Auckland
Publishing Timeline
Received 22 January 2013
Accepted 15 July 2013
Published 7 October 2013
References
Abolfathi, M., Yadollah, T.A.H., & Nurizan, Y. (2009). The role of religiosity on the relationship between
chronic health problems and psychological wellbeing among Malay Muslim older persons. Research
Journal of Medical Sciences, 3(6), 188-193.
Ahmad, S.M., & Pevalin, D.J. (2012). An Arabic version of the spiritual well-being scale. International
Journal for the Psychology of Religion, 22(2), 119-134. http://dx.doi.org/10.1080/10508619.2011.638592
Ahmed, A.K. (2010). Religiosity, subjective well-being, and neuroticism. Mental Health, Religion & Culture,
13(1), 67-79. http://dx.doi.org/10.1080/13674670903154167
Alesina, A., Di Tella, R., & McCullough, R. (2004). Inequality and happiness: are Europeans and
Americans different? Journal of Public Economics, 88(9-10), 2009-2042.
http://dx.doi.org/10.1016/j.jpubeco.2003.07.006
Ano, G.G., & Vasconcelles, E.B. (2005). Religious coping and psychological adjustment to stress: A metaanalysis.
Journal of Clinical Psychology, 61(4), 461-480. http://dx.doi.org/10.1002/jclp.20049
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 194
Blaine, B., & Crocker, J. (1995). Religiousness, race and psychological wellbeing: Exploring social
psychological mediators. Personality and Social Psychology Bulletin, 21(10), 1031-1041.
http://dx.doi.org/10.1177/01461672952110004
Bradshaw, M., Ellison, C.G., & Marcum, J.P. (2010). Attachment to God, images of God, and
psychological distress in a nationwide sample of Presbyterians. International Journal for the Psychology
of Religion, 20(2), 130-147. http://dx.doi.org/10.1080/10508611003608049
Chamberlain, K., & Zika, S. (1988). Religiosity, life meaning and wellbeing: Some relationships in a
sample of women. Journal for the Scientific Study of Religion, 27(3), 411-420.
http://dx.doi.org/10.2307/1387379
Clarke, A.E., & Lelkes, O. (2009). "Let us pray: Religious interactions in life satisfaction." (Working Paper:
Paris School of Economics).
Cohen-Zada, D., & Sander, W. (2011). Religious participation versus shopping: What makes people
happier? Journal of Law and Economics, 54(4), 889-906. http://dx.doi.org/10.1086/658862
Crabtree, S. (2010). Religiosity highest in the world's poorest nations. Gallup Global Report.
http://www.gallup.com/poll/142727/religiosity-highest-world-poorest-nations.aspx
Crothers, C. (2005). New Zealand religious affiliations in the new millennium. Stimulus, 13(4), 31-49.
Cummins, R.A., Eckersley, R., Pallant, J., Van Vugt, J., & Misajon, R. (2003) Developing a national index
of subjective wellbeing: The Australian Unity Wellbeing Index. Social Indicators Research 64: 159–190.
http://dx.doi.org/10.1023/A:1024704320683
Diener, E., & Seligman, M.E.P. (2004). Toward an economy of well-being. Psychological Science in the
Public Interest, 5(1), 1-31. http://dx.doi.org/10.1111/j.0963-7214.2004.00501001.x
Diener, E., Tay, L., & Myers, D. (2011). The religion paradox: If religion makes people happy, why are so
many dropping out? Journal of Personality and Social Psychology, 101(6), 1278-1290.
http://dx.doi.org/10.1037/a0024402
Dolan, P., Peasgood. T., & White, M. (2008). Do we really know what makes us happy? A review of the
economic literature on the factors associated with subjective well-being. Journal of Economic
Psychology, 29(1), 94-122. http://dx.doi.org/10.1016/j.joep.2007.09.001
Ellison, C.G., Bradshaw, M., Kuyel, N., & Marcum, J.P. (2012). Attachment to God, stressful life events,
and changes in psychological distress. Review of Religious Research, 53(4), 493-511.
http://dx.doi.org/10.1007/s13644-011-0023-4
Ellison, C.G., & Taylor, R.G. (1996). Turning to prayer: Social and situational antecedents of religious
coping among African Americans. Review of Religious Research, 38(2), 111-131.
http://dx.doi.org/10.2307/3512336
Ellison, C.G. (1991). Religious involvement and subjective well-being. Journal of Health and Social Behavior,
32(1), 80-99. http://dx.doi.org/10.2307/2136801
Fabricatore, A.N., Handal, P.J., Rubio, D.M., & Gilner, F.H. (2004). Stress, religion, and mental health:
Religious coping in mediating and moderating roles. International Journal for the Psychology of Religion,
14(2), 91-108. http://dx.doi.org/10.1207/s15327582ijpr1402_2
Fiorito, B., & Ryan, K. (2007). Spirituality and psychological well-being: A mediator-moderator study.
Review of Religious Research, 48(4), 341-368.
Fischer, R., & Boer, D. (2011). What is more important for national well-being: Money or autonomy? A
meta-analysis of well-being, burnout and anxiety across 63 societies. Journal of Personality and Social
Psychology, 101(1), 164-184. http://dx.doi.org/10.1037/a0023663
Gelman, A., Carlin, J.B., Stern, H.S., & Rubin, D.B. (2003). Bayesian data analysis (2nd ed.). New York, NY:
Chapman & Hall/CRC.
Hackney, C.H., & Sanders, G.S. (2003). Religiosity and mental health: A meta-analysis of recent studies.
Journal for the Scientific Study of Religion, 42 (1), 43-55. http://dx.doi.org/10.1111/1468-5906.t01-1-00160
Helliwell, J.F., Layard, R., & Sachs, J. (eds.). (2012). World happiness report. New York: The Earth Institute,
Columbia University. http://issuu.com/earthinstitute/docs/world-happiness-report
Homan, K.J., & Boyatzis, C.J. (2010). Religiosity, sense of meaning, and health behavior in older adults.
International Journal for the Psychology of Religion, 20(3), 173-186.
http://dx.doi.org/10.1080/10508619.2010.481225
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 195
Hoverd, W.J. (2008). No longer a Christian country? Religious demographic change in New Zealand
1966-2006. New Zealand Sociology Journal, 23(1), 41-63.
Hoverd, W.J., Atkinson, Q.D., & Sibley, C.G. (2012). Group size and the trajectory of religious
identification. Journal for the Scientific Study of Religion, 51(2), 286-303. http://dx.doi.org/10.1111/j.1468-
5906.2012.01643.x
Hoverd, W.J., Bulbulia, J., & Sibley, C.G. (In Press). Does poverty predict religion? Religion, Brain &
Behavior, 2013.
Hoverd, W.J., & Sibley. C.G., (2010). Religious and denominational diversity in New Zealand 2009. New
Zealand Sociology Journal, 25(2), 62-92.
Kahneman, D., & Deaton, A. (2010). High income improves evaluation of life but not emotional wellbeing.
Proceedings of the National Academy of Sciences of the United States of America, 107(38), 16489-
16493. http://dx.doi.org/10.1073/pnas.1011492107
Kim, A.E. (2003). Religious influences on personal and societal wellbeing. Social Indicators Research, 63(1-
3), 149-170.
Kosmin, B., & Keysar, A. (2009). American religious identification survey 2008. Hartford: Trinty College.
Kosmin, B., Meyer, E., & Keysar, A. (2001). American religious identification survey 2001. New York: City
University of New York.
Manuela, S., & Sibley, C.G. (in press). Why do Pacific people with multiple ethnic affiliations have poorer
subjective wellbeing? Negative ingroup affect mediates the identity tension effect. Social Indicators
Research.
Marx, K., & Engels, F. (1964). On religion. New York: Schoken Books. PMCid:PMC1515419
McKenzie, F., Ellison-Loschmann L., & Jefferys, M. (2010). Investigating reasons for socioeconomic
inequalities in breast cancer survival in New Zealand. Cancer Epidemiology, 34(6), 702-708.
http://dx.doi.org/10.1016/j.canep.2010.07.007
Mochon, D., Norton, M.I., & Ariely, D. (2008). Getting off the hedonic treadmill, one step at a time: The
impact of regular religious practice and exercise on well-being. Journal of Economic Psychology, 29(5),
632-642. http://dx.doi.org/10.1016/j.joep.2007.10.004
Mochon, D., Norton, M.I., & Ariely, D. (2011). Who benefits from religion? Social Indicators Research,
101(1), 1-15. http://dx.doi.org/10.1007/s11205-010-9637-0
Morris, P. (2007). Religious diversity in New Zealand. Statement on religious diversity. UNESCO Publications.
Nachowitz, T. (2007). New Zealand as a multireligious society: Recent census figures and some relevant
implications. Aotearoa Ethnic Network Journal, 2(2), 1-8.
Oliver, W.H. (1966). Christianity among the New Zealanders. Landfall, 77, 4-20.
Pargament, K. (2002) The bitter and the sweet: An evaluation of the costs and benefits of religiousness.
Psychological Inquiry, 13(3), 168-181. http://dx.doi.org/10.1207/S15327965PLI1303_02
Salmond, C., & Crampton P. (2012). Development of New Zealand's Deprivation Index (NZDep) and its
uptake as a national policy tool. Canadian Journal of Public Health, 103(8), 57-61.
Salmond, C., Crampton. P., & Atkinson, J. (2007). NZDep2006 index of deprivation. Wellington:
Department of Public Health.
Salmond, C., Crampton, P., King, P., & Waldegrave, C. (2006). NZDep: A New Zealand index of
socioeconomic deprivation for individuals. Social Science & Medicine, 62(6), 1474-1485.
http://dx.doi.org/10.1016/j.socscimed.2005.08.008
Sengupta, N.K., Osborne, D., Houkamau, C.A., Hoverd, W.J., Wilson, M.S., Greaves, L.M., < & Sibley,
C.G. (2012). How much happiness does money buy? Income and subjective wellbeing in New
Zealand. New Zealand Journal of Psychology, 41(2), 21-34.
Sibley, C.G., & Bulbulia, J. (2012). Faith after an earthquake: A longitudinal study of religion and
perceived health before and after the 2011 Christchurch New Zealand earthquake. PLoS ONE, 7(12),
e49648. http://dx.doi.org/10.1371/journal.pone.0049648
Sibley, C.G., Harré, N., Hoverd, W J., & Houkamau, C. (2011). The gap in the subjective wellbeing of
Māori and New Zealand Europeans widened between 2005 and 2009. Social Indicators Research, 104,
103-115. http://dx.doi.org/10.1007/s11205-010-9729-x
Religion, Deprivation and Subjective Wellbeing
Hoverd & Sibley
www.internationaljournalofwellbeing.org 196
Smith, C., & Faris, R. (2005). Socioeconomic inequality in the American religious system: An update and
assessment. Journal for the Scientific Study of Religion, 44(1), 95-104. http://dx.doi.org/10.1111/j.1468-
5906.2005.00267.x
Smith, T.B., McCullough, M.E., & Poll, J. (2003). Religiousness and depression: Evidence for a main effect
and the moderating influence of stressful life events. Psychological Bulletin, 129(4), 614-636.
http://dx.doi.org/10.1037/0033-2909.129.4.614
Statistics New Zealand. (2006). New Zealand religious affiliation 2006. Statistics New Zealand.
http://www.stats.govt.nz/Census/2006CensusHomePage/QuickStats/quickstats-about-a-
subject/culture-and-identity.aspx
Steenwyk, S.A.M., Atkins, D.C., Bedics, J.D., & Whitley Jr, B.E. (2010). Images of God as they relate to life
satisfaction and hopelessness. International Journal for the Psychology of Religion, 20(2), 85-96.
http://dx.doi.org/10.1080/10508611003607942
Strawbridge, W.J., Shema, S.J., Cohen, R.D., Roberts, R.E., & Kaplan, G.A. (1998). Religiosity buffers
effects of some stressors on depression but exacerbates others. Journal of Gerontology, 53(3), 118-126.
Tix, A., & Frazier, P.A. (1998). The use of religious coping during stressful life events: Main effects,
moderation, and mediation. Journal of Consulting and Clinical Psychology, 66(2), 411-422.
http://dx.doi.org/10.1037/0022-006X.66.2.411
Utter J., Denny S., Crengle., S., Ameratunga S., Robinson E., Clark T.,< & Maddison R. (2010).
Overweight among New Zealand adolescents: Associations with ethnicity and deprivation.
International Journal of Pediatric Obesity, 5(6), 461-466. http://dx.doi.org/10.3109/17477160903568439
Vaccarino F., Kavan, H., & Gendall, P. (2011). Spirituality and religion in the lives of New Zealanders.
The International Journal of Religion and Spirituality in Society, 1(2), 85-96.
Waldegrave, C. (2009). Culture and religion. In P. Koopman-Boyden & C. Waldegrave, (Eds). Enhancing
wellbeing in an aging society: 65-84 year old New Zealanders in 2007 (EWAS) Monograph No. 1. (pp 191-
208). Published by Family Centre Social Policy Research Unit, Lower Hutt and the Population
Studies Centre, University of Waikato, Hamilton.
Waldegrave, C. (2010). Culture and religion among midlife New Zealanders. In Waldegrave, C., &
Koopman-Boyden, P. (Eds). Enhancing wellbeing in an aging society (EWAS) Monograph No. 2. (pp 197-
214). Published by Family Centre Social Policy Research Unit, Lower Hutt and the Population
Studies Centre, University of Waikato, Hamilton.
Wilson, J. (1993). Religious affiliation—Change over time: An analysis of intercensal change in religious
affiliation of 5 yearly age-cohorts in recent New Zealand census. New Zealand Sociology Journal, 8(1),
60-94.
Wnuk, M., & Marcinkowski, J.T. (2012). Do existential variables mediate between religious-spiritual
facets of functionality and psychological wellbeing? Journal of Religion and Health. Published online:
24th March 2012. http://dx.doi.org/10.1007/s10943-012-9597-6
Yuan, Y., & MacKinnon, D.P. (2009). Bayesian mediation analysis. Psychological Methods, 14(4), 301-322.
http://dx.doi.org/10.1037/a0016972