Explaining religiosity: towards a unified theoretical
model
Jörg Stolz
Abstract
The article presents a unified theoretical model, explaining differences in Christian
and ‘alternative’ religiosity at individual and collective levels. The model
reconstructs and integrates the most important theories explaining religiosity
(deprivation, regulation, socialization, cultural production, and ethnicity) as
complementary causal mechanisms in a rational-action based framework. It is
maintained that the mechanisms of the various theories are not exclusive, but
complementary, and that integration into the general model is both theoretically
and empirically beneficial. The model is tested on representative data from
Switzerland. Substantively, I find for the Swiss case that Christian religiosity can be
best explained by a religious socialization mechanism.The most important mechanisms
accounting for alternative religiosity involve deprivation, gender, and age.
Keywords: Christian religiosity; alternative religiosity; explanation; rational action;
deprivation
1. Introduction1
The question as to how we can explain religiosity sociologically gives rise to a
variety of answers. Modernization theorists claim that religiosity changes with
the level of rationalization and differentiation (Wallis and Bruce 1995;
Dobbelaere 2002). Rational-choice theorists want to convince us that what
really influences religiosity is regulation of the ‘religious market’ (Iannaccone
1991). Still others argue forcefully that socialization is most important (Kelley
and De Graaf 1997; Voas 2003; Voas and Crockett 2005). Recently, Pippa
Norris and Ronald Inglehart (2004) have claimed that religiosity is really
explained by two things only: existential insecurity (a form of deprivation) and
religious culture. The discussion is lively, to say the least, with opponents
accusing each other of serious errors of fact, revisionism, incompetence,
Stolz (Observatoive des Religions en Suisse, Universite de Lausanne) (Corresponding author email: joerg.stolz@unil.ch)
© London School of Economics and Political Science 2009 ISSN 0007-1315 print/1468-4446 online.
Published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden,
MA 02148, USA on behalf of the LSE. DOI: 10.1111/j.1468-4446.2009.01234.x
The British Journal of Sociology 2009 Volume 60 Issue 2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
1920
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
untruthfulness and sometimes even going so far as to wish for the death of
their adversaries’ theoretical constructions (Stark and Finke 2000: 79; Bruce
1999: 2).
In this multiplicity of different opinions, many researchers seem to think
that one or the other theory ‘must be right’, and that one mechanism should
be able to explain everything. While this would certainly be ‘elegant’ from a
theoretical point of view, it is in my view highly unlikely. The nature of social
and historical reality is such that many different economical, social, technical,
and historical mechanisms are at work in different times and places. We will
not find the one cause of religiosity even if we do surveys for another fifty
years.
Given this state of affairs, it seems sensible to set up a list of the most
important theoretical mechanisms known to date, and to integrate them into
a common theoretical framework. In doing so, we will get a larger picture of
possible causalities and interactions and will be able to use them jointly in
order to explain specific socio-historic settings. This is what I try to accomplish
in this paper. In order to do this, I use a general model of action
advanced by German scholar Hartmut Esser as a common theoretical
denominator, extract the (implicit or explicit) mechanisms that different
theories and empirical studies propose, give them a properly explanatory
structure, cleanse them of unnecessary assumptions, and build them into a
common theoretical framework. I maintain that theories that have been
transformed into causal mechanisms in the general model will usually
become both theoretically more interesting and easier to operationalize and
test. In what follows, I
– define the terms ‘religiosity’ and ‘explanation’ and sketch the abstract
explanatory model by Hartmut Esser (section 2)
– outline five of the most important theories that aim to account for religiosity
and integrate them as complementary causal mechanisms into the
general model by Esser (section 3)
– describe the theoretical advantages that have been gained by building the
mechanisms into the model (section 4)
– describe a case study that applies the model to representative data in
Switzerland, showing empirical pay-offs (section 5)
2. The problem: explaining religiosity
The goal of this article is to construct a unified model for the explanation of
religiosity. It is therefore useful to define the central terms of this problem and
to discuss the main theoretical background we will use.
346 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
2.1 Religiosity and religion
I define religiosity as individual preferences, emotions, beliefs, and actions that
refer to an existing (or self-made) religion. ‘Religion’ then denotes the whole
of cultural symbol-systems that respond to problems of meaning and contingency
by alluding to a transcendent reality which influences everyday life but
cannot be directly controlled.2
Religious symbol-systems incorporate mythical,
ethical and ritual elements as well as ‘salvation goods’. Note that – following
these definitions – religiosity is an individual and religion a cultural
phenomenon. If an individual prays, sacrifices, believes, loves or fears his god –
then this is ‘religiosity’. ‘Christianity’, ‘Islam’, ‘Christian Science’ or ‘Raelianism’,
on the other hand, are religious symbol-systems, that is, ‘religions’. In the
example below we will be measuring and explaining two distinct religiosities:
Christian and alternative. While the first religiosity is made up of preferences,
emotions, beliefs, and actions referring to the Christian symbol system, alternative
religiosity concerns preferences, emotions, beliefs and actions that are
concerned with ‘New Age’ or ‘Alternative Spirituality’ (Stolz 2005; Bloch
1998).3
2.2 Sociological explanation
Explanations are not just descriptions, typologies or ‘conceptual frameworks’,
but very concrete answers to ‘why-questions’. In a very general way, a phenomenon
can be said to be explained if one can show how it results from a
set of initial conditions and a generative (and therefore causal) mechanism
(Hedström 2005). More specifically, a sociological explanation is given if we
can show how an initial situation (macro) leads individuals to react to this
situation (micro) and how, through aggregation, these individual reactions
combine to form a new social outcome (macro), that is, the phenomenon to be
explained (Coleman 1990).The ‘social causality’ or the ‘mechanism’ therefore
incorporates
(a) opportunities, norms, and cultural resources in a situation,
(b) ‘rational’ action, based on beliefs and preferences by various individuals
reacting to these situations and
(c) the fact that these reactions may have various intended and unintended
effects.4
Important authors in the field of explanatory sociology are, for example,
Boudon (1998), Coleman (1990), or Goldthorpe (2000).A new formulation of
the model that summarizes and integrates the advances of recent decades
appears in the books by Hartmut Esser (1999, 2000). ‘Explaining religiosity’
with such a mechanism-based approach requires us therefore to show how
Explaining religiosity 347
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
attributes of initial situations (concerning opportunities, norms and culture)
will lead individuals to adapt rationally by choosing more (or less) religious
actions, thereby creating, when the individual actions are aggregated, the differences
in religiosities we wish to explain (Stolz 2009a). Note that such a
unified theory is not restricted exclusively to one type of religiosity, but should
be applicable to religiosity in general.5
2.3 An explanatory model of action (Esser)
The Esser model is outlined (in a simplified way) in Figure I. In what follows,
I explain the most important elements of this model by beginning on the top
right:
1. The model explains macro-phenomena.The goal is not to account for the
actions of a single individual, but to explain differences between social
groups (e.g. differing levels of religiosity, rates of suicide, etc.).
2. The macro-phenomenon that is to be explained results from the aggregation
of individual actions. A rate of suicide comes into being through
the aggregation of the different individual suicidal actions.
3. Individual actions are assumed to be rational. An action is said to be
rational, if it chooses from all possible options the one that promises the
greatest utility.We opt for a model of‘bounded rationality’ (Simon 1983);
Figure I: The explanatory model of action
348 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
this means that we acknowledge the influence of institutional and cultural
factors as well as the fact that individuals only have a limited faculty
of calculation.
4. Individual actions take place in situations.Here,Esser distinguishes external
and internal conditions of the situation. External conditions of the
situation are again broken down into opportunities,institutional rules,and
cultural framing. Opportunities consist of factual options faced by individuals,
that is, the means an actor controls in a situation in order to reach
his or her goals (Esser 1999: 52).This is where scarcity plays an important
role, which is why economists focus (often exclusively) on this level. A
second element of external conditions of the situation consists of institutional
rules,e.g.societal norms,roles or constitutions,which are backed up
by positive or negative sanctions (Esser 1999: 53). The third element of
external conditions is cultural frames. These are ‘models’ for typical
situations and processes such as ‘frames’ or ‘scripts’, as well as symbol
systems (like languages or ideologies).6
5. The internal conditions of the situation of individuals consist of beliefs,
preferences and identity. Beliefs include cognitions, the stock of different
types of expertise, and the expectations that an individual has at its
disposal. Preferences are ordered evaluations by means of which an
individual assesses its environment. Personal identity is the sum of selfdescriptions
and self-evaluations of an individual (including descriptions
and evaluations of him- or herself to the environment).
6. In our model, the external conditions of the situation are influenced by
‘Macro attributes of the society’.These can be societal structures (e.g. the
existence of a welfare state or democracy), societal processes (e.g. inflation
or industrialization) or events (e.g. natural catastrophes or the
soccer world championship). Certain societal attributes influence individual
situations necessarily.This is the case for compulsory school attendance
or military service. Other societal attributes have the effect that
individuals will be subjected to certain phenomena with a certain probability
only (e.g. poverty, illness, alternative culture, and religious
socialization). Still other societal attributes apply to all individuals in the
abstract, but may become important only in specific situations (e.g. juridical
norms).
7. The internal conditions of the situation are influenced in an important
way by (primary and secondary) socialization of the individual that has
taken place in the past. In childhood, during so-called ‘primary socialization’,
the individual learns beliefs and preferences which are internalized
into the personality of the individual to an important extent (Esser 2000:
371). Fundamental belief- and preference-structures often remain
Explaining religiosity 349
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
relatively stable during adulthood and constitute an ‘internal environment’
of individual action.
8. Interestingly, however, in this model internal conditions of the situation
are influenced not only by (former) socialization, but also by current
external conditions of the situation, specifically by institutional and cultural
parameters (Esser 1999: 75 ff.).The reason for this is that individuals
know that their most elementary needs (especially physical well-being
and social status) can only be fulfilled if they adapt in fundamental ways to
the respective society,culture,and institutions.Every society,historical era
or social group prescribes in a different way what goals or goods should be
striven for and by which means one should strive for them in a socially
accepted way.In science,for example,a central‘goal’ is‘reputation’,which
is to be attained with publications of high quality. In professional sports,
success in tournaments is the central goal;the accepted means is‘training’
(and the forbidden means:doping).The fact that we use means in order to
produce goods, which in turn provide utility (all of which is embedded in
institutions and culture) has been called ‘social production function’
(Lindenberg 1989). With the help of social production functions, we can
suddenly explain why individuals will often strive ‘from their very heart’
exactly for those goals that are the primarily valued goals and goods in a
given society or social group. They also help to explain sudden shifts in
individual preferences. For example, in the GDR, before 1989, to be a
member of the SED provided societal status. Many individuals therefore
developed preferences for membership and high positions in the party.
After the ‘Wende’, party membership was suddenly an important stigma
and individual preferences changed dramatically.
Note that with the help of this model a long-standing debate is overcome.
We do not address the question if individuals either react to opportunities or
follow norms and culture. Rather, the model supposes that individuals react (in
a limited sense) rationally both to opportunities and to institutions and culture,
and that their beliefs and preferences are shaped by opportunities, institutions,
and culture. In my view, one of the great strengths of this type of model is that
it succeeds in linking the different levels of social and individual reality in a
straightforward way. We thus begin to understand how macro-, meso- and
microlevels are interacting in order to produce social reality.7
3. A unified model for the explanation of religiosity
This model presented above shall now serve us as a ‘baseline structure’ in
order to integrate the most important elements of previous explanations of
religiosity. The central idea is very simple. The five most prominent theories
350 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
that explain religiosity (deprivation-, market-, modernization-, ethnicity-, and
socialization-theory) can all be reformulated as macro-micro-macro mechanisms
inside the framework of the Esser model. Seen in this way, the theories
are not mutually exclusive, but complementary (and themselves causally
linked in various ways). It depends on the given historic-geographical context
if and how strongly one or the other of the mechanisms will actually operate.
A central point of the rational action reconstruction is that I systematically
build in a micro foundation as well as the possibility to choose between
religious and secular alternatives into every one of the mechanisms. In what
follows, I describe the five reconstructed theories.
3.1 Macro attributes of society and external conditions of the situation
Deprivation
The first mechanism we introduce argues that deprivation of different kinds
leads to or reinforces religiosity (Weber (1978 (1920): 299ff; Stark and Bainbridge
1985; Gill and Lundsgaarde 2004; Norris and Inglehart 2004; Sherkat
and Ellison 1999). Deprivation may be defined as a situation in which an
individual is not able to satisfy one or several needs. Without claiming to be
exhaustive, I name the following societal attributes that express presence or
absence of deprivation: a high or low level of the supply of material goods,
absence or presence of a welfare state, (non)existence of high standard
medical care and of political/juridical security. When Individuals find themselves
in these and other depriving circumstances, they will seek a solution to
their problems. It may therefore be rational to turn to a religion that offers
help in various forms. Religions can help by (a) referring to myths and thus
provide ‘meaning’ to deprivations, e.g. theodicies; (b) embedding suffering into
ritual actions and strict rules of conduct (e.g. prayers, sacrifice, religious service,
ethics); (c) rendering suffering bearable through a specific ‘habitus’, e.g. being
one of the ‘chosen few’; (d) promising hope and good outcomes for the future.
Apart from these rather intangible goods, religious groups may also offer very
concrete help with housing, food, comfort, finances, medical care, etc.Thinking
back to our general theoretical framework we see that the given attributes of
society will influence the external conditions of the situation mainly through
opportunities.Our rational action reconstruction of deprivation theory emphasizes
especially two points. First, a low level of the societal deprivation
attributes will lead to an overall reduced probability of basic deprivations on
the individual level. Individuals simply do not face many of the problems that
are completely normal in other societies (high mortality of children, famine,
cholera, etc.); accordingly, they do not need corresponding (religious or
secular) coping strategies. Second, for remaining deprivations, individuals face,
besides religious answers, very powerful (sometimes even obligatory) secular
Explaining religiosity 351
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
coping strategies. Examples are unemployment insurance, health insurance,
centers for psychological help, etc. If such secular solutions to deprivations
exist, demand for religious coping strategies will often decrease, since secular
solutions will often be perceived as more efficient and reliable (Bruce 1999: 17;
Stolz 2008).8
Regulation
A second mechanism may be termed regulation. I define regulation very
broadly as the ways in which the state or social groups influence individual or
collective actors through enforceable rules and norms. Relating this mechanism
to our general framework, we see that here it is mainly norms that lead
actors to various types of rational adaptations. Two main theoretical
approaches to regulation can be distinguished, one concerning ‘supply’, the
other one ‘demand’.
The supply-side regulation mechanism, proposed by ‘rational-choicetheorists’
is as follows (Iannaccone 1991; Finke and Stark 1992; Stark and
Iannaccone 1994). The state regulates the supply of religion by applying rules
and norms, thus favouring some religious groups and discriminating others. In
extreme cases, it establishes one state religion and bans all or other religions
(as is the case in current Saudi Arabia) or it bans or hinders religion altogether
(as was the case in former East Germany). In moderate cases, such as modern
Germany, Italy or Sweden, some religious groups with ‘official recognition’
enjoy various advantages. Other unrecognized groups are therefore
disadvantaged. Such regulations, rational-choicers say, prevent the religious
market from working properly; it will be rational for ‘recognized’ religious
groups and church leaders to become lazy and they will supply – as all
monopolists do – a product which is too expensive and insufficiently attuned to
customer needs. As a result, customers do not find the religious product they
are looking for and – as a rational adaptation – will not ‘consume’ as much as
they would have done in a free market. Hence, overall religiosity goes down.
While this mechanism of market-like inter-religious competition seems to be
at work in some socio-historic cases (Finke and Stark 1992), a very large
number of regularities cannot be explained with its help (Bruce 1999, 2002;
Chaves and Gorski 2001).9
In order to integrate this mechanism into our
general framework, I have to drop the (theoretically very unfortunate) supposition
of ‘stable religious demand’ dear to the rational choicers. In contrast,
I suppose that demand for religious goods may change due to changing social
production functions and changing relative attractiveness of religious and
secular goods. Regulation of supply will therefore not only influence innerreligious
competition, but also (and in modern societies more importantly)
competition between the religious and the secular.
352 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
The demand-regulation story is very different and told less often. The state
and social groups can regulate individuals’ demand for religion: positive or
negative sanctions may apply to religious membership, participation, belief (or
at least absence of visible disbelief) may be positively or negatively sanctioned.
Rational individuals will now follow the rules and show or abstain from
showing religiosity in order to gain social approval and/or avoid punishment.
Such a regulation of demand may be a deliberate state policy (Borowik 2002;
Froese and Pfaff 2005); it may also be the result of anonymous social control in
small (e.g. rural) communities or generally in social groups with strong and/or
many social ties (McLeod 1998; Sherkat 1997; Olson 1999) Especially if we
look at historical evidence, there can be no doubt, that ‘regulation of demand’
has been a very important element of the religiosity, say, in Europe up until the
twentieth century.The normative pressure on individuals to belong to a certain
confession, to follow the prescribed rites and, hide disbelief and to abstain
from seeking out religious or magical alternatives was very high indeed (e.g.
Poos 1995; Swanson 1990).10
Production of culture
Societal ‘production of culture’ also influences the level of religiosity of a given
society (Weber 1978[1920]; Wilson 1982; Wallis and Bruce 1995). This fact
seems to almost have fallen into oblivion in the current discussion. The key
point here is that individuals will find religious beliefs and ‘truths’ more convincing
to the extent that they are surrounded by religious culture, be this
through products of societal subsystems (e.g. the media) or individual
interaction.There are various societal attributes that touch upon this point, the
most important being probably the following. First, the extent of development
of an independent and secular industry of leisure, a media system, and modern
science. From a certain level of societal differentiation and further, these
domains produce cultural goods (activities, information, interpretations, and
knowledge) that have freed themselves from religious connotations and follow
a systemic logic of their own (Dobbelaere 2002; Luhmann 1982).11
Second, the
size and effervescence of religious groups in a given society (Kelley and De
Graaf 1997; Norris and Inglehart 2004). If there are large and powerful religious
organizations in a society, they may run their proper production of media
and leisure facilities and thus make it more likely that individuals will come
into contact with religious culture. However, culture is not only transmitted by
media or leisure products, but also, and perhaps more importantly, by everyday
interactions among individuals. Thus, the size and effervescence of religious
groups also influences the likelihood with which individuals will come into
contact with religious or secular business partners, school mates and teachers,
potential spouses, neighbours etc.
Explaining religiosity 353
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Relating this mechanism to our general theoretical framework, we note that
the most important influence operates (independently from concrete decisions)
through cultural frames. In a society with a high level of secular
culture production (through societal subsystems or individual interactions),
individuals face an enormous amount of non-religious cultural interpretations
for whatever they encounter (Berger 1990[1969]: 107–8). Certainly, religious
interpretations of the foundation of the state of Israel, the terrorist attack on
the World Trade Center, or a stock market crash remain possible. However, the
dominant cultural patterns of interpretation explain these phenomena by
pointing to historical, sociological, psychological, economic, and other causes.
Rational individuals will therefore weigh the ‘evidence’ made available to
them and, by and large, adopt secular interpretations of what happens. Conversely,in
a society with a high level of religious culture production,individuals
will have a high probability of coming into contact with religious interpretations
of events.They will therefore have a much higher probability of adopting
religious beliefs and interpretations themselves. Cultural frames aside, the
leisure industry and the media system also have effects on the opportunities of
individuals. As a matter of fact, individuals often have a choice between religious
and secular options. Secular and religious newspapers,TV-stations, sport
clubs, get-togethers etc. compete for attention, time and money. Depending on
the quality and cost of the religious and secular options, rational individuals
will then ‘consume’ more or less religious or secular culture products and as a
consequence ‘become’ more or less religious.
Ethnic and cultural assertion of identity
A further mechanism may be called ‘ethnic and cultural assertion of identity’
(Bruce 1999).Identity may be defined as the set of descriptions that individuals
or a social group construct about themselves and their relations to their
environment. These descriptions and hypotheses find support on distinctions
(about what one is not), including norms and values that may be used in power
struggles (Esser 1999). Identity may be considered ethnic, if it supposes a
common descent and rests on a common language, culture, nationality or
religion. The basic idea is that religiosity and religion become strong where
they can be used as resources, in order to conserve and defend ethnic and
cultural identities. The societal attributes that are important in this case are a
rapid social and cultural change, the social and cultural threatening of all or
parts of the population as well as the level of immigration. These attributes
affect especially the cultural frames and opportunities of two types of actors.
First, we have individuals negatively affected by rapid social and cultural
change and/or political or cultural repression or conflict.These individuals see
their acquired human, social and cultural capital being devalued by rapid
social change or by oppression. It therefore becomes rational for them to fight
354 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
for the importance of their ascribed identity markers, of their cultural-ethnic
and religious identity. In this way, they can regain social worth. Well-known
examples are Polish or Quebec Catholics as well as Protestants and Catholics
in Ireland. Second: immigrants will often gather in the host country and build
social networks and communities,in order to conserve their cultural (and often
religious) identity and to receive help with integration. This may lead to a
special emphasis on religion and religiosity in ‘diaspora groups’ (e.g. Baumann
and Salentin 2006; Warner 1998).
Socialization
Yet another mechanism works through socialization. Socialization may be
defined as an interactive, both voluntary and involuntary, process of transmission
and learning in which individuals teach and learn norms, values, behaviour,
expertise, meanings and identity, and in which socialized individuals may
integrate these elements into their personality by internalization (compare to
Esser 2000: 371). Since religiosity refers to religious symbol systems, individuals
have to learn ‘religious content’ before they can even become religious. (De
Roos, Iedema and Miedema 2004). Thinking back to our general theoretical
framework, we see that socialization is one of the most important ways of
influencing the beliefs and preferences of children. When we look at how
children are influenced by socialization, we see that often little rationality is
involved. Thus, parents are likely to transmit their personal religious preferences,
beliefs and practices to their children – even if they do not consciously
try to do so. Conversely, children may learn consciously or unconsciously by
listening to, imitating, or just watching their parents. Not surprisingly, high
parental religiosity therefore leads on average to higher religiosity of children.
This mechanism may be described as ‘social learning’ (De Roos, Iedema and
Miedema 2004) ‘sedimenting of beliefs’ (Bruce 1999) or ‘transmission of religious
human capital’ (Iannaccone 1990). Empirical studies show that the
power of this transmission is influenced by various contextual factors and by
the type of values,beliefs or practices (Hoge,Petrillo and Smith 1982;De Roos,
Iedema and Miedema 2004; Lindner and Moore 2002). Especially important
is the question of whether the couple is religiously homogamous or heterogamous,
the former being much stronger in transmitting religious membership,
behaviour, and belief to children (Voas 2003; Need and De Graaf 1996;
Iannaccone 1990).12
This, however, is not the whole story. Theoretically more interesting is the
question under what conditions parents will want to socialize their offspring
religiously. What norms and opportunities of the situation of parents will lead
them to search for or refrain from a religious upbringing of their children? This
is where rationality strongly comes into play. Important societal attributes are
(without asserting completeness): The importance of religiosity as a ‘religious
Explaining religiosity 355
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
capital’ in the society; the relative size and type of religious group, and the
freedom of religious groups to produce and run various services and facilities.
These societal attributes translate into situations of both parents and children
in the following ways. Generally, parents will teach religious values to their
children if they believe that these values are ‘important’ and ‘useful’ in society
(often independently of whether they themselves are personally religious or
not). The central factor here is societal norms. If ‘being religious’ is important
in order to be considered a valuable member of the community or society (i.e.
is an important ‘social capital’), parents will put a lot of effort into the religious
education of their children. A study by David Voas (2003), for example, suggests
that baptism of children was considered an important social convention
if not a religious duty in England before the 1940s, so that even children of
mixed marriages were very likely to be baptized. After the 1940s, however,
these societal norms withered, and since then children in mixed marriages
have a very low probability of being baptized. This mechanism also explains
the fact that individuals who become parents will often increase their religious
practice in order to set an example for their children (Sherkat and Ellison
1999).A relatively low general ‘social value’ given to religion as a social capital
has therefore the effect on most parents that they will – rationally – bring up
their children less religiously. However, it has a different effect on a few
parents who – for individual reasons or because they are members of a highly
religious group – still think that religion is very important. This minority of
devout parents now has to put – again: rationally – much more personal effort
into socializing their children, if they want the socialization to be successful.
For now, they must not only teach the religious symbols and practices to their
children, but also must control their children’s surroundings from secularizing
influences in order to better inculcate religious beliefs and preferences. They
will therefore try to choose schools, youth groups, friends, television programs,
and music styles for their children, in order to let them live in a more or less
protected religious world. This has the overall effect that religious upbringing
becomes an extremely important explanatory concept especially in relatively
secular societies (Kelley and De Graaf 1997). In this respect certain opportunities
are equally important. Whether devout parents can find religious structures
for their children indeed depends on societal structures such as the
relative size and type of religious group and the freedom of religious groups to
produce various goods and services (such as television shows, radio programmes,
kindergartens, schools, universities, hospitals, therapy programmes).
If a society has powerful and relatively free religious groups that provide such
structures (as in the USA), it is possible for parents to protect their children
from secularizing influences quite easily (Bruce 2002). If a society does not
include religious groups with a lot of facilities or if religious groups are not
allowed to run a lot of facilities themselves, then socialization will become
much more difficult.
356 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
3.2 Internal conditions of the situation
The internal conditions of the situation consist of the beliefs, the preferences,
and the identity of the individual. These phenomena are, as already noted,
strongly influenced by individual socialization (especially through parents and
peer groups).They are thus to some extent not rationally chosen.Another part
of internal conditions of the situation, however, can be assumed to follow our
rationality assumptions. Let us first look at beliefs. Individuals – we assume –
will rationally try to get information on how they can solve their concrete
problems of life.They will adopt beliefs that are supported by ‘strong reasons’,
given the evidence (Boudon 1998). This means that their beliefs will be to a
large extent influenced by their ‘information environment’ concerning religious
or secular possibilities, e.g. how, in their society, one routinely copes with
deprivations in a religious or secular manner, what norms concerning religion
are currently valid etc. (Stolz 2009a). Preferences may equally – in part – be
chosen rationally. Following the theory of social production functions, we
assume that the way in which physical well-being and high social status may be
produced varies from society to society; every society or social group fixes
certain instrumental goals and acceptable means in order to reach important
final goals. Individuals have to adapt to these social production functions if
they want to meet their basic needs. Therefore, individuals will develop preferences
for religious or secular actions depending on how important these
actions are as a means of reaching final goals highly valued in their society or
social group (Stolz 2009b). For example, if no or insufficient possibilities of
secular coping with deprivation are available, individuals will develop strong
preferences for religious coping strategies. Participation in religious rites and
help through transcendent forces are now thought to be of primary
importance. Conversely, if powerful possibilities of secular coping exist, many
individuals will suddenly feel that religion is ‘not so important’ for them.
Religion is now seen to be responsible only for remnants of insecurity and
contingency (e.g. in the case of death). Or take the example of the regulation
mechanism: In societies with a high regulation of religious demand, individuals
will build up preferences for or against religious behavior as a means of
avoiding sanctions and of producing social esteem (‘extrinsic motivation’). In
societies with low regulation, on the other hand, individuals will develop
largely ‘intrinsic motivations’ for or against religiosity, etc.
3.3 Rational religious action
Individuals will react – following our model – to external and internal conditions
of the situation in a rational way. Given their knowledge and their
preferences and based on their evaluation of opportunities, institutions and
cultural frames, individuals will choose the action that seems – from their
Explaining religiosity 357
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
subjective point of view – to provide the greatest utility (Elster 1986; Stolz
2009a). If individuals pray to God for healing or alleviation of poverty, as may
be the case in many current agrarian societies, if they fulfill their religious
duties in order to be seen as respected members of the community, as in many
European societies until the end of the nineteenth century, if they go to Mass
as a sign of protest against a political regime, as in Poland before 1989, or if
they take their children out of religious instruction courses at school because
they deem religious knowledge ‘not important’ – in every one of these cases
they decide with the help of (bounded) rationality. On a very abstract level, the
model assumes that individuals will have a higher probability of choosing a
religious action, the more they are deprived (and the poorer the available
secular coping strategies), the more they are forced to exhibit religiosity by
norms and sanctions,the lower the secular production of culture,the more they
feel their ethnic or cultural identity threatened by social or cultural change,
and the more they have been socialized religiously. Individuals will have a
higher probability of giving their children a religious socialization, the more
religion is seen to be an important ‘cultural capital’ in society.
Open and problematical alternatives
At this point, I have to treat the way in which secular and religious alternatives
present themselves to the individual in everyday life. The entire model builds
on the idea that individuals (at least in certain societies) can ‘choose’ between
religious and secular options. One might now object that such decisions (e.g.
going to church or sleeping in, choosing either a religious youth group or a
football club) are empirically rare. However, this only seems to be the case,
because, once taken, decisions are routinized. In the words of Schütz and
Luckmann (1974 : 208ff.), when individuals routinize their behaviour, they
transform ‘problematic alternatives’ into ‘open alternatives’ that are not really
perceived as situations of decision anymore. Once I have decided to sleep in on
a Sunday morning instead of going to church, I do so routinely and do not
make a ‘conscious decision’ of this behaviour every Sunday. Conversely, under
certain conditions the ‘open alternatives’ may be transformed back into ‘problematic
alternatives’.
4. Theoretical advantages of the unified model
A central claim of this paper is that the reconstruction in the macro-micromacro
framework is not just a way of classifying possible mechanisms, but
leads to the following theoretical advantages:
1. The model presents a guide to possible causal mechanisms in a given
socio-historical setting. It provides researchers with a convenient tool in
358 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
order to think through the possible causalities and interactions and to set
these in relation with the historical parameters of the case under
investigation.
2. Since we routinely build a theoretical micro-foundation into the mechanisms,
they all become explanatory. Instead of just describing trends, they
can now predict variation of religiosity.The mechanisms implicit in modernization
theory and socialization theory are especially in dire need of
such micro-foundations.
3. Unnecessary and theoretically harmful assumptions are eliminated.
Examples are the idea of stable religious demand, the idea of an incompatibility
of socialization and rational action or the notion that one could
not react rationally to changing of norms.
4. The mechanisms become clearly distinctive, operationalizable, and
empirically testable. It becomes clear that the mechanisms are – on a
theoretical level – complementary. We therefore have to assess empirically
if they operate in a given socio-historical context, and if they do,
what is their relative importance.
5. An empirical application: Switzerland
Since – as the English like to say – the proof of the pudding is in the eating, I
will now try to show that the proposed model has not only theoretical but also
empirical advantages, by applying it to representative data from Switzerland.
As the model has been formulated in the abstract and for general purposes, it
has to be adapted to my special case. Decisions have to be taken as to what
types of religiosity to investigate, what ‘levels’ to consider, what variables to
measure at what levels and so on. In this example, my goal is to explain
Christian and alternative religiosity by specifying all the theoretical mechanisms
mentioned above, and using a two-level model with cantons on the
aggregate and persons on the individual level. Switzerland provides an interesting
case for testing different mechanisms producing religiosity, since it
includes extremely rural and modernized cantons, cantons with a virtually
absolute separation of church and state, others where church and state are
closely linked and three regions with different language and ethnic cultures.
5.1 Data
The data stem from a joint file of two surveys. First is the survey ‘Religion
et lien social’ (Campiche 2004). This study is based on telephone interviews
carried out in 1999. The population includes all individuals living in
Explaining religiosity 359
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Switzerland aged between 16 and 75. Sampling was done in a two-stage,
random way, first sampling communes, and then individuals inside the
communes. Response rate was 54 per cent; the number of interpretable interviews
was 1562. Second was the ISSP-study in Switzerland in 1999.This survey
was a posted questionnaire to people who had already participated in the
survey ‘Religion et lien social’.The additional questionnaire was completed by
1212 individuals. This leads to the fact that for some items only a restricted
sample of individuals is available. Analysis below will take this into account
very carefully.13
Tests show an overall good representativity for standard
demographic variables. Specifics can be found in Campiche (2004). For the
purposes of this paper, individuals belonging to non-Christian religions had to
be excluded from analysis, since non-Christian religions in Switzerland are so
small that they appeared in the sample with too modest frequency to allow
valid analysis.
5.2 Operationalizing Christian and alternative religiosity
I use four indicators in order to measure Christian religiosity: importance of
religion in general (7-point scale), frequency of prayer (5-point scale), frequency
of Christian religious service (5-point scale), belief that God exists and
that he has shown himself in Jesus Christ (5-point scale). These indicators tap
belief, preferences, individual and collective practice and are highly correlated.
I use four indicators in order to measure alternative religiosity: Belief in
astrology (yes/no), belief that some fortune tellers can really see in the future
(yes/no), belief that good luck charms really sometimes work (yes/no) and
belief in reincarnation (5-point scale).These items only tap beliefs.With factor
analysis techniques I built two scales for the two types of religiosities respectively
(see appendix, Table AI). These two variables will be the ‘dependent’
variables of further analyses. They measure two clearly different types of
religiosities and are empirically not correlated.
5.3 Operationalizing the mechanisms
Deprivation
‘Objective’ deprivation was operationalized by two variables: income (11point
scale) and education (7-point scale). These two indicators were combined
to form a ‘structural deprivation index’. ‘Subjective’ deprivation by the
variables of self-reported class-membership (5-point scale) and happiness
(4-point scale).
Socialization
Religious socialization was captured by five indicators, two tapping primary
and three secondary socialization.A summated index of parents’ church going
360 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
when the respondent was 12 years old measured parents’ religious practice in
the formative years of the respondent (11-point scale). A dummy variable
indicated if parents are religiously homogeneous, that is, if mother and father
belong to the same denomination (Catholic, Reformed etc.). A dichotomous
variable indicated if the respondent has been ‘confirmed’ a rite which
presupposes a period of catechism. An interval scale measured years of religious
education in school. And a further dichotomous question asked if the
respondent had attended a religious youth group at the age between 16 and 20.
Using factor analysis techniques, these variables were combined to form a
‘religious socialization index’,
Regulation of demand
In our theoretical section we have argued that religious demand may be
regulated by social norms. We assume that these norms will be stronger in
small villages, rural contexts and where norms of religious practice and belonging
have not visibly been broken. Regulation of demand is measured in this
study by three indicators. One is an index of community size (less than 3,000;
3,000–9,999; 10,000–100,000; more than 100,000 inhabitants). A second indicator
is a dichotomous variable distinguishing urban and rural living contexts in
Switzerland created by the FOS (federal office of statistics) in Switzerland.14
A
third indicator is the percentage of individuals without religion in the community
where individuals live, calculated on the basis of the national census from
1990. Using factor analysis techniques, these variables were combined to form
a ‘regulation of demand index’.
Ethnicity
Nationality was captured by using a dummy variable for ‘naturalized Swiss’
(the official term for a person who has acquired Swiss nationality after having
had a different nationality before) and ‘foreigner’, Swiss being the reference
category. Mother language was likewise captured by creating a dummy variable
for French-speaking and one for Italian-speaking indivduals (Germanspeaking
being the reference category).
Regulation of supply
As already noted, the 26 Swiss cantons all have adopted different ways degrees
of regulating religion, leading to very weak regulation in, for example, Geneva
and Neuchâtel and very strong regulation in cantons like Zurich or Berne.
Regulation of supply was measured with an adapted version of the well-known
scale by Chaves and Cann (1992). I adapted the scale in order to capture the
Explaining religiosity 361
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
differences in regulation between Swiss cantons as precisely as possible. This
led to a scale which ranges from 0 – 9 points.15
Production of culture
Production of culture was operationalized concerning leisure opportunities,
closeness to the scientific world and religious culture. Leisure opportunities
were operationalized as the mean state expenditure for culture in different
cantons per capita from 1990–98. While this is only an indirect measure for
total leisure possibilities, it nevertheless captures some of the leisure opportunity
differences between cantons. Cantons with few large leisure facilities such
as theatres, museums, opera will have lower state expenditure for culture per
capita.16
Closeness to the ‘scientific world’ was coded on the basis of the
information found in the Statistisches Jahrbuch der Schweiz 1998. Specifically,
we operationalized it as the percentage of state expenditure on education
related to total state expenditure per canton (including communes). Religious
culture was coded with two dummies, one for Reformed culture, one for mixed
(Reformed and Catholic) culture, Catholic being the reference category.
Individual-level and social-level variables
The theoretical model presented above gives possible independent variables
on social and individual levels. In order to keep my model reasonably simple,
I have entered on a collective level variables where cantons are known to
differ in important manners17
and/or where the variable does only make sense
on a social level (regulation, religious tradition, production of secular culture).
All other variables are measured on individual levels.18
Problems of this kind
have to be addressed with multi-level models (Hox 2002; Snijders and Bosker
2000). In our case, we use a two-stage-model, in which one level concerns the
cantons, while the other level concerns the individuals.19
5.4 Hypotheses
Table I gives on the ‘predicted’ side to the left an overview over our hypotheses,
distinguishing 7 causal mechanisms and two types of religiosity. These
hypotheses imply a knowledge of how the mechanisms have been
operationalized. Hypothesis (1) expects that religious regulation of supply will
diminish both Christian and alternative religiosity. Hypothesis (2) expects that
secular culture will diminish both Christian and alternative religiosity. Hypothesis
(3) expects that traditional Catholic culture will lead to more Christian
religiosity (compared to Protestant culture). Hypothesis (4) expects that
(Christian) religious socialization will further Christian, but not alternative
religiosity. Hypothesis (5) expects deprivation to raise both types of religiosity.
362 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Hypothesis (6) expects that the stronger the social norms (e.g. the smaller the
community), the higher Christian religiosity but the lower alternative
religiosity. Hypothesis (7) expects that foreigners or naturalized Swiss as well
as French- or Italian speaking individuals will show more Christian religiosity
than individuals with the Swiss nationality and german-speaking Swiss.
Drawing on previous research (Campiche 2004), we can hypothesize that
women will show more Christian and alternative religiosity than men and that
older individuals will show more Christian but less alternative religiosity than
younger individuals.
5.5 Data analysis
Data analysis was conducted with the statistical packages SPSS for data exploration
and -preparation and HLM6 for the multi-level modelling.All variables
were standardized. As a result, the regression coefficients given in the tables
below are standardized coefficients. We can thus compare the relative importance
of different causal factors.
5.6 Findings
Explaining Christian religiosity
What theoretical mechanisms do explain religiosity in Switzerland? Let us first
look at Christian religiosity (first column in Table II).20
The coefficients presented
are the standardized regression coefficients in a multi-level multiple
regression. These coefficients measure the ‘influence’ of each indicator on the
dependent variable ‘Christian religiosity’ while controlling for all other variables
in the model.21
Table I: Hypotheses: predictions and findings
Predicted Actual
Christian
religiosity
Alternative
religiosity
Christian
religiosity
Alternative
religiosity
Collective
(1) Regulation of supply - - n.s. n.s.
(2) Secular culture - - n.s. n.s.
(3) Religious (Catholic) culture + n.s. + n.s.
Individual
(4) (Christian) socialization + - ++ n.s.
(5) Deprivation + + + +
(6) Regulation of demand (norms) + - + n.s.
(7) Ethnicity + n.s. + n.s.
Individual (control)
(8) Gender (woman) + + + +
(9) Age + - + Notes:
- ≡ negative relationship ; n.s. ≡ not significant; + ≡ positive relationship; ++ ≡ strong positive
relationship.
Explaining religiosity 363
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
For Christian religiosity the results are as follows: State regulation of religious
supply (measured by the regulation index) and secular culture (measured
by state expenditure on culture and education) have no significant effect on
Christian religiosity. Here, we are on the collective level, that is, we look only
at how well the explanatory indicators explain the difference in mean Christian
religiosity between cantons. The theories expect that cantons with high
regulation and strong state expenditures on secular culture will show low
Christian religiosity. Clearly, we find no evidence for these theoretical
mechanisms. In contrast, ‘religious tradition of the canton’ has a significant
effect on Christian religiosity. Cantons in Switzerland are traditionally either
Protestant, Catholic or mixed (Protestant and Catholic).The variables used in
this model are dummy-variables, the base category being ‘Protestant’. The
Table II: Multi-level multiple regressions for Christian and alternative religiosity
Christian religiosity Alternative religiosity
Fixed
Regulation rel. supply
State regulation 0.031 0.016
Secular culture
State expend. on culture 0.101 -0.004
State expend. on educ. 0.014 -0.010
Religious tradition
Traditionally mixed 0.123 0.061
Traditionally catholic 0.145* 0.007
Socialization
Religious socialization index 0.339** -0.070
Deprivation
Income&Education Index -0.108** -0.168**
Happiness 0.051 -0.032
Regulation rel. demand
Regulation of demand index -0.080* 0.024
Ethnicity
Naturalized Swiss 0.068* -0.052
Foreigner 0.050 -0.054
French-speaking 0.010 -0.079
Italian-speaking 0.046 -0.063
Control
Gender (woman) -0.025 0.298**
Age 0.102* -0.098
Gender * Age 0.173* -0.186**
Full time -0.061 0.076
Part time -0.092** 0.044
Roman/Christ Catholic -0.037 0.102*
Other Christian 0.187** -0.084*
No Religion -0.163** -0.030
Random
s2
= var (Rij) (level 1 variance) 0.649 0.871
r0
2
= var (U0j) (level 2 variance) 0.007 0.000
N (level 1, individuals) 755 683
N (level 2, cantons) 21 21
R1
2
(level 1) 34.8% 13.2%
R2
2
(level 2) 55.2% 30.6%
Deviance 1903.066 1918.437
364 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
regression coefficients thus show the difference of the respective category to
the ‘Protestant’-category. The hypothesis was that cantons with a Catholic
tradition would show stronger religiosity than Protestant cantons (as has been
often shown in various studies on the level of countries in Western Europe).
This turns out to be correct (beta = 0.145*). The fact that regulation does not
explain anything, while religious tradition is the one important factor which
explains inter-cantonal differences in Switzerland is shown graphically in
Figure II and III.
The socialization mechanism finds strong support in the data (beta =
0.339**). The higher the religious socialization by parents, school and peers,
the higher the Christian religiosity of children when they have become adults.
Socialization turns out to be – by far – the most important mechanism explaining
Christian religiosity in our specific case. The deprivation mechanism finds
some, but not unequivocal support in the data. The structural deprivation
index is significantly related to Christian religiosity: the lower an individual
finds itself concerning education and income (i.e. the higher the ‘deprivation’),
the higher religiosity (-0.108**). However, the subjective indicator of
deprivation – self-reported happiness – is not related significantly to the
prediction of the theory.This may be due to the fact that causalities are difficult
to assess when looking at the link between religiosity and (un-)happiness.
Figure II: Christian religiosity and regulation (level of cantons)27
Notes: GE = Genève,AG = Aargau,AR = Appenzell Ausserrhoden, BS = Basel Stadt, NE=
Neuchâtel, OW = Obwalden, GR = Graubünden, SG = Sankt Gallen, SO = Solothurn,
LU = Luzern, SZ = Schwyz, GL = Glarus, SH = Schaffhausen, VS = Valais, TI = Ticino,
FR = Fribourg, UR = Uri, JU = Jura,VD = Vaud, BL = Basel Land, BE = Bern, ZH = Zürich
ZH
BE
BL
VD
JU
VS
FR
TI
UR
OW
GR
SG
SO
TG
LU
NW
SZ
ZG
GL
SH
AG
AR
BS
NE
GE
0
0.5
1
1.5
2
2.5
3
3.5
4
9
Regulation
0 1 2 3 4 5 6 7 8
Explaining religiosity 365
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
For (un-)happiness may be the cause of religiosity just as well as its effect.Thus
deprived persons may become religious and due to religions’ compensatory
power may then see themselves as rather happy. Effects may then cancel out in
the aggregate. The mechanism involving regulation of demand finds a certain,
although limited, support in this model. The theory predicts that individuals
have a lower probability of being (in a Christian way) religious if they live in
larger communities, if they live in an urban area and if the percentage of
individuals without religious affiliation in their village or city is higher. This
turns out to be correct (beta = -0.080*);the effect is,however,rather small.The
ethnicity mechanism is only very slightly supported by the data. The theory
would have predicted a higher Christian religiosity for immigrants and language
minorities.In the model,only the effect of being a‘naturalized Swiss’ has
a (small) positive effect: naturalized Swiss show a slightly higher Christian
religiosity. Concerning our control variables, we see a strong interaction
between gender and age. While there is little difference in religiosity between
old and young men, age makes a lot of difference for women: Older women
show a lot more Christian religiosity than men. This interaction is shown
graphically in Figure IV.We also see that individuals who work part time have
a slightly lower probability of Christian religiosity than individuals who are not
working (base-line group). Furthermore, other Christians (e.g. Evangelicals,
Orthodox) have a clearly higher probability and individuals with no official
religiosity a lower probability of showing Christian religiosity. This model
explains 34.8 per cent of the variance of Christian religiosity on the individual
level and 55.2 per cent on the collective level.22
Figure III: Christian religiosity and religious tradition (level of cantons)
Notes: see Notes for Figure II
OW
VS
JU
FR
TI
SO
LU
BE
AR
BS
SH
VD
ZH
BL
GE
NE
GL
AG
TG
SG
GR
ZG
SZ
NW
UR
0
0.5
1
1.5
2
2.5
3
3.5
4
Protestant Parity Catholic
366 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Explaining alternative religiosity
We now turn to the explanation of alternative religiosity (second column in
Table II).As we can see,the only theoretical mechanism explaining anything in
our case is the deprivation mechanism: individuals with higher levels of structural
deprivation (lower levels of education and income) show higher levels of
alternative religiosity.All the other mechanisms, be they regulation of demand
and supply, secular culture, religious tradition, socialization, or ethnicity fail to
explain this particular kind of religiosity.An important effect is however to be
found concerning our control variables. First, women show much more alternative
religiosity than men. Again, we find an interaction between age and
gender: younger women find alternative religiosity to be significantly more
interesting than older women, while the effect is not significant for men.
Furthermore, we find that Roman Catholics have a slightly higher and Other
Christians a slightly lower probability to be alternatively religious. In this
model we are able to explain 13.2 per cent of alternative religiosity on the
individual level and 30.6 per cent on the collective level.
All in all, explained variance is much lower for alternative religiosity than
for Christian religiosity both on the individual and the collective level. This
may be due to mainly two factors. First, alternative religiosity has probably not
been as reliably measured as Christian religiosity. Better items might lead to
higher explained variance. Second, there might be additional explanatory variables
that have been omitted for the case of alternative religiosity. Note,
specifically, that religious socialization (the most important predictor of
Figure IV: Christian religiosity, age and gender
0
0.5
1
1.5
2
2.5
Age
Male
Female
16-20 21-30 31-40 41-50 51-60 61-70 71-75
Explaining religiosity 367
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Christian religiosity) is only geared to Christian, but not alternative religious
socialization.23
6. Summary and discussion
This study has tried to make both a substantive and methodological contribution
by presenting the first comprehensive application of a new unified theoretical
framework for the explanation of religiosity to a special, namely the
Swiss, case. This new framework postulates seven distinct ‘generative mechanisms’
that ‘produce’ religiosity and are all integrated on a common ‘rational
action’ basis. Such a theoretical integration has several advantages. It provides
researchers with a guide to possible causal mechanisms in a given sociohistorical
setting, it renders the mechanisms explanatory and operationalizable,
cleanses them from unnecessary theoretical assumptions and leads to
more systematic empirical research. When applying the framework, I have
used multi-level multiple regression models in order to explain Christian and
alternative religiosity in Switzerland. Since I started with strong theory and
clear hypotheses, the results are interesting even when we do not find significant
relationships. Table I summarizes the findings by comparing our predictions
to the findings.
Substantively, the study throws new light on the case of Switzerland. Put
succinctly, we see that, in Switzerland, Christian religiosity could be best
explained by religious socialization mechanisms. Individuals had a much
higher probability of showing Christian religiosity if they had been strongly
socialized by their parents, if they came from a mono-religious household and
if they had Christian peer-socialization. Deprivation, social control, religious
tradition of the canton as well as gender and age also played a certain role.
Individuals with less education and income, living in smaller and rural communities
and in traditionally Catholic (or mixed) cantons had a higher likelihood
of showing Christian religiosity. The latter finding is very similar to that
by Norris and Inglehart (2004) who were also able to show that religious
culture was important on a collective (national) level. Furthermore, the study
shows that especially older women had a much higher probability than
younger women to be religious, while age was much less important for men.
The mechanisms linked to state regulation, ethnicity or the secular culture
of the canton had no explanatory power in our specific case. Although the
Swiss cantons show tremendous differences in regulation (going from almost
complete separation to the churches as state agencies), state regulation did not
explain any variation of Christian religiosity. These findings are additional
evidence for the view that regulation mechanism are not of primary importance
for the explanation of religiosity at least in western democracies (Chaves
and Gorski 2001). The predictions by the secular culture mechanism did not
368 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
fare any better.This mechanism supposes that the production of secular leisure
opportunities, secular media and modern science may ‘crowd out’ religious
products and interpretations. Again, the indicators used did not explain any
variance of Christian religiosity. It is probable that Swiss Cantons are uniformly
on a rather high level of secular production of culture and that remaining
variance is not able to influence religiosity. On the other hand, we have to
acknowledge that operationalization of secular culture is not entirely satisfactory
and could be much improved in future research. With a maximum of
explanatory variables I was able to explain 34.8 per cent (level1) and 55.2 per
cent (level 2) of the variance of Christian religiosity. Alternative religiosity
could be explained to a lesser extent than Christian religiosity (13.2 per cent on
level 1 and 30.6 per cent on level 2). The most important mechanisms were
linked to deprivation, gender and age. Individuals with lower education and
lower income as well as women and younger individuals had a higher probability
of being alternatively religious. All in all, it seems as if alternative
religiosity is especially important for individuals searching for ‘empowerment’
in the face of deprivation and/or social barriers and obstacles. Somewhat
similar findings are reported by Houtman and Mascini (2002) and Mears and
Ellison (2000).24
All other mechanisms (regulation of supply, secular culture,
religious tradition, socialization, regulation of demand, ethnicity) had no substantial
explanatory value for alternative religiosity.As was the case for Christian
religiosity, the ‘market mechanisms’ did not have any effect on alternative
religiosity in our specific case. We did not find more alternative religiosity
where regulation was lower; nor was alternative religiosity stronger where
Christian religiosity was waning. Since previous accounts concerning the Swiss
case (e.g. the contributions in Campiche and Dubach 1992) did not use the
systematic approach advocated here (using a list of possible causal mechanisms
and two clearly distinguishable religiosities), they have overlooked both
the central part played by socialization for Christian religiosity and of deprivation
for alternative religiosity.
As has been mentioned above, the present approach does not suppose that
the postulated mechanisms have to exhibit the same importance everywhere.In
fact, in Switzerland, the market, culture – as or ethnicity mechanisms do not
seem to operate at all – even though they are clearly very important in other
socio-historical contexts.This leads us to the question of why this is the case. In
principle, we would now have to explain due to what ‘initial conditions’ certain
mechanisms are operable or not in a given context.I have tried to do this above,
althouth these explanations have had to remain ‘post hoc’. In the future,
comparative research will have to try to include variables which explain the
presence or absence of certain mechanisms in given countries in the model itself.
Concerning methods, this article has tried to make advances in two respects.
First, not just one but two types of religiosity have been operationalized and
‘explained’: Christian and alternative religiosity.25
A lot of the literature looks
Explaining religiosity 369
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
only at one type of religiosity – mostly Christian, often operationalized by
frequency of church-attendance and/or strength of belief in god or selfdescription
as being more or less ‘religious’. As this contribution shows,
however, alternative religiosity is in our case not correlated to Christian religiosity
or to the importance one attributes to ‘religion’ in general (see Appendix,
Table A1). Furthermore, the structure of the generative mechanisms are
very different for alternative religiosity than for Christian religiosity. Second,
we have used a multi-level model in order to investigate the various generative
mechanisms which may ‘take effect’ on different levels of social reality. The
literature so far has mostly just looked at either the individual level or the
collective level (mostly: at nations), neglecting other levels.As is well known in
the methodological literature, such procedures are in danger of producing
artefacts (e.g. ecological or atomist fallacy). For example, if we looked in our
study only at the collective level and inferred from there directly to the
individuals, we would be inclined to say that religious culture is the one
important factor explaining individual religiosity in Switzerland.When looking
at the individual level, however, we note that other factors, such as socialization
or deprivation, are much more important.
While the proposed specific application of the unified theoretical framework
certainly has its limits concerning operationalization and generalizability26
, I
hope, nevertheless, to have been able to show the basic strength of my general
approach: The proposed unified framework produces testable hypotheses and
its different parts can be verified or falsified for different types of religiosities
and in different socio-historical contexts. Theory and empirical research are
thus brought in a closer relationship than before. If this proposal will be used
in future cross-national comparative research, an improved state of of the art
may be expected.
(Date accepted: January 2009)
Appendix
Table AI: Factor analysis for two types of religiosity
Factor 1 Factor 2
Christian religiosity Alternative religiosity
Importance of religion 0.829 0.037
Frequency of prayer 0.716 0.111
Frequency of religious service 0.662 0.055
God and Jesus 0.652 -0.211
Belief in astrology 0.035 0.632
Belief in good luck charms -0.048 0.627
Belief in fortune tellers -0.040 0.567
Belief in reincarnation 0.033 0.516
Notes: Principal Axis Factoring, Rotation = Oblimin; Factor correlation r = -0.007. The factor
analysis was also carried out on the polychoral correlations (instead of the Pearson Correlations)
and yielded essentially similar results. This analysis was carried out in R.
370 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Notes
1. I thank Steve Bruce, Hartmut Esser,
Philippe Gilbert, Stefan Huber, Denise
Hafner Stolz, Julian Höbsch, Jeanne Rey,
Mallory Schneuwly-Purdie, David Voas, the
editors, and four anonymous reviewers for
their helpful comments on previous versions
of this paper. Jean-Philippe Antonietti has
given statistical advice and Christine Rhone
has corrected the text.The usual disclaimers
apply.
2. See for similar definitions Geertz
(1993), Theissen (2000), Stolz (2004). Some
Table AII: Multi-level multiple regressions for Christian and alternative religiosity (higher N)
Christian religiosity Alternative religiosity
Fixed
Regulation rel. supply
State regulation 0.040 -0.009
Secular culture
State expend. on culture 0.039 -0.054
State expend. on educ. -0.026 0.043
Religious tradition
Traditionally mixed 0.108* 0.016
Traditionally catholic 0.088* -0.061
Socialization
Religious socialization index 0.326** -0.073*
Deprivation
Income&Education Index -0.043 -0.151**
Regulation rel. demand
Regulation of demand index -0.077* 0.017
Ethnicity
Naturalized Swiss 0.027 -0.045
Foreigner 0.067** -0.085
French-speaking -0.002 -0.072
Italian-speaking 0.032 -0.020
Control
Gender (woman) -0.002 0.275**
Age 0.066* -0.122**
Gender * Age 0.200** -0.161**
Full time -0.040 0.093
Part time -0.064** 0.068
Roman/Christ Catholic -0.009 0.108**
Other Christian 0.140** -0.096**
No Religion -0.195** -0.030
Random
s2 =
var (Rij) (level 1 variance) 0.684 0.876
r0
2 =
var (U0j) (level 2 variance) 0.001 0.008
N (level 1, individuals) 1,472 1,007
N (level 2, cantons) 21 21
R1
2
(level 1)ix
31.8% 11.9%
R2
2
(level 2)ix
68.0% 17.3%
Deviance 3712.853 2810.017
Notes: This is essentially the same analysis as that presented in the text above in that indicators
from all explanatory theories as well as control variables are included. The one difference is that
we exclude variables for which only limited N were available. This means that we use only three
variables for our religious socialization index (excluding number of years of religious education),
that we use only the variable ‘education’ in order to measure ‘structural deprivation’ and that we
exclude the variable ‘happiness’. In this way we can run the analysis on a much larger data-set
(N = 1,472 instead of N= 755 and N = 1,007 instead of N = 638). The findings are largely similar
which gives our analysis additional support.
Explaining religiosity 371
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
scholars doubt if we can define ‘religion’ in a
general way, arguing that ‘religion’ is a
Christo- or Eurocentric concept. In my view,
however, we do seem to need some general
term (be it ‘religion’ or some other expression)
and even scholars rejecting the term
do not seem to be able to do without it.
Given this state of affairs, it seems to be
preferable to define the term than to use an
implicit definition. Note also that ‘religion’
and ‘religiosity’ do not cover everything that
might interest a sociologist of religion. Especially,
there are ‘religious institutions’ (like
congregations, denominations, churches,
roles, etc.) that can be analytically distinguished
from both religiosity and religions
as cultural symbol-systems.
3. Alternative spirituality is not just an
‘individual phenomenon’ as some researchers
seem to think. Rather, it has – just as
Christianity or Islam – both an individual
side (= religiosity or spirituality) and a cultural
side (= symbol system, language).
Alternative spirituality systems are made
up of a large number of symbols coming
from various traditions; there exist,
however, shared ways to apprehend and
‘mix’ these symbols. This has been
described with terms such as individualism,
holism, pluralism, or reflexivity (Bloch
1998, Besecke 2001). Still by far the best
description of alternative spirituality in
Switzerland is Mayer (1993).
4. The social phenomenon, which is the
outcome of a macro-micro-macro-explanation,
can then itself be seen as the starting
point of a new macro-micro-macro element
(Esser 2000). Since the theory links microand
macro levels (involving meso-levels, if
warranted), there is no ‘causal priority’ of
one or the other level. Any macro-situation
that is the initial ‘cause’of a phenomenon to
be explained is itself the outcome of microlevel
actions and vice-versa.
5. It is clear, however, that we restrict ourselves
to sociological factors that can enhance
or diminish religiosity in individuals.
6. The distinction between opportunities,
institutions, and cultural frames is an analytical
one. In concrete empirical cases, the
three ‘domains’ are very often linked in
various ways (Esser 1999: 51).
7. Because of space requirements, the
meso-level remains less explicit in this
article than would be warranted from the
point of view of the theory. Stolz (2009b) is
an attempt to show how this framework
might explain the success and failure of religious
organizations (i.e. on the meso-level).
8. Gill and Lundsgaarde (2004) convincingly
show how churches may be ‘crowded
out’ by the state when it comes to the production
of ‘welfare state goods’ and thus
how aggregate religiosity turns out to be
weaker in countries without a welfare state
than in countries with one.
9. See for a harsh critique: Bruce 1999, for
a sympathetic overview: Warner 1993, and
for an attempt to typologize salvation goods
and show the limits of the religious market
concept: Stolz 2006.
10. Note that states may (and do) use
regulation of demand and supply simultaneously,
which can lead to opposite effects
on religiosity.
11. Historically, the possibilities of
secular leisure and media consumption
appeared first in the cities (McLeod 1998:
14).
12. Research shows that various structural
barriers exist for heterogamous
couples to transmit their faith (e.g. they have
to decide which religion to transmit, the religious
officials may not consent to ‘mixed
transmission’ etc.). It has to be noted,
though, that individuals in religious heterogamous
couples have not made religious
membership an important criterion for the
choice of their partner; it is therefore likely
that they do not put much emphasis on religion
or the religious socialization of their
children in the first place.
13. Two points are noteworthy. First,
explanation of Christian religiosity can rely
on more data than explanation of alternative
religiosity, since important items for the
latter are only given in the ISSP survey.
Second, the more predictors a model
includes, the lower the N we can rely on,
since the probability of missings increases.
372 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Concerning missing values, I have (1)
refrained from ‘imputing’ data in case of
missing values. (2) in case of missing values
in items that belong to a scale, I have calculated
values for the scale on the basis of
existing data (3) calculated models for both
Christian and alternative religiosity in which
missing data are minimized. See Table AII
in the Appendix.
14. ‘Urban’ contexts are made of cities
and agglomerations. A city has more then
10,000 inhabitants; an agglomeration
includes more than 20,000 individuals, integrates
different communities and disposes of
a ‘central zone’which is made up of different
community centres.
15. I thank René Pahud de Mortanges for
help with questions of cantonal church law.
The specifics of the operationalization and
coding can be obtained from the author.
16. A problem with this operationalization
is that it focuses on high culture,
whereas popular culture might be also
relevant. Unfortunately, I have not been
able to track down indicators for popular
culture by canton.
17. For example, it would not be sensible
to measure existence of a highly developed
welfare state on the cantonal level in Switzerland,
since this is a constant.
18. This choice of what variables have to
be measured at what levels would be made
quite differently if one wanted, say, to
compare nations on a world-wide basis. It
would then be mandatory to measure GDP
and the existence of a welfare state on the
aggregate level.
19. The alternative to a hierarchical linear
model would be a covariance model with the
group levels as fixed factors. Such a model,
however, seems to be less useful for my purposes
since I want to assess the effect of
different variables measured at the group
level. Unexplained variance on the group
level is in my case not to be seen as due to
sampling error (since I include all Swiss
cantons), but to unmeasured variables and
residual error.See Snijders and Bosker 2000:
43. I thank Joop Hox and Jean-Philippe
Antonietti for discussions on this point.
20. The zero-order correlations can be
obtained from the author.
21. After inspection of the data, I present
a random intercept model (Snijders and
Bosker 2000: 38 ff). Estimating different
slopes for different parameters in different
cantons does not improve the fit of the
model substantially.
22. We calculate explained variances for
level 1 and level 2 with formulas given by
Snijders and Bosker 2000: 102.
23. Due to many missings in some variables,
the N’s in these models are somewhat
reduced. In Table AII in the Appendix I
show the same two models while excluding
the variables with many missings.
24. In an interesting paper, Woodhead
(2007: 123) explains the preponderance of
women in alternative spirituality by pointing
to the fact that women in late modernity are
‘more likely to suffer from more complex
and far-reaching conflicts of role and identity
than men’. Further empirical research
has to show if this mechanism is correctly
identified and stable.
25. While other types of religiosity exist
and might be included, the two types investigated
here can be reasonably well established
in the data and have led to a
considerable body of literature.
26. e.g. in some instances, operationalizations
could have been more fortunate (especially
concerning secular culture, regulation
of demand and alternative religiosity).
27. The variable Christian religiosity is
standardized and has therefore mean = 0
and sd = 1. In Figures II–IV, I have performed
a linear transformation on the variable
for better visibility.This has no effect on
the associations between variables.
Bibliography
Baumann, M. and Salentin, K. 2006
‘Migrant Religiousness and Social
Incorporation: Tamil Hindus from
Sri Lanka in Germany’, Journal of
Explaining religiosity 373
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
Contemporary Religion 21(3): 297–
323.
Berger, P.L. 1990 The Sacred Canopy: Elements
of a Sociological Theory of Religion.
New York: Anchor Books [first published
in 1967].
Besecke, K. 2001 ‘Speaking of Meaning in
Modernity: Reflexive Spirituality as a Cultural
Resource’, Sociology of Religion 62(3):
365–82.
Bloch, J.P. 1998 New Spirituality, Self, and
Belonging. How New Agers and NeoPagans
Talk about Themselves, Westport:
Praeger.
Borowik, I. 2002 ‘The Roman Catholic
Church in the Process of Democratic Transformation:
The Case of Poland’, Social
Compass 49(2): 239–52.
Boudon, R. 1998 ‘Social Mechanisms
Without Black Boxes’, in P. Hedström and
R. Swedberg (eds) Social Mechanisms. An
Analytical Approach to Social Theory,
Cambridge: Cambridge University Press.
Bruce, S. 1999 Choice and Religion: A Critique
of Rational Choice Theory, Oxford:
University Press.
Bruce, S. 2002 God is Dead: Secularization
in the West, Oxford: Blackwell.
Campiche, R. and Dubach, A. 1992 Croire
en Suisse(s), Lausanne: éditions l’Age
d’Homme.
Campiche, R.J. 2004 Les deux visages de la
religion. Fascination et désenchantement,
Genève: Labor et Fides.
Chaves, M. and Cann, D.E. 1992 ‘Regulation,
Pluralism, and Religious Market
Structure’, Rationality and Society 4(3): 272–
290.
Chaves, M. and Gorski, P.S. 2001 ‘Religious
Pluralism and Religion Participation’,
Annual Review of Sociology 27: 261–81.
Coleman, J.S. 1990 Foundations of Social
Theory, Cambridge: The Belknap Press of
Harvard University Press.
De Roos, S., Iedema, J. and Miedema, S.
2004 ‘Influence of Maternal Denomination,
God Concepts, and Child-Rearing Practices
on Young Children’s God Concepts’,
Journal for the Scientific Study of Religion
43(4): 519–35.
Dobbelaere, K. 2002 Secularization: An
Analysis at Three Levels, Brussels: Peter
Lang.
Elster, J. 1986 ‘Introduction’ in J. Elster
(ed.) Rational Choice, Oxford: Basil
Blackwell.
Esser, H. 1999 Soziologie. Spezielle Grundlagen,
Band 1: Situationslogik und Handeln.
Frankfurt: Campus.
Esser, H. 2000 Soziologie. Spezielle Grundlagen,
Band 6: Sinn und Kultur. Frankfurt:
Campus.
Finke, R. and Stark, R. 1992 The Churching
of America 1776–1990: Winners and Losers
in Our Religious Economy, New Brunswick,
NJ: Rutgers University Press.
Froese, P. and Pfaff, S. 2005 ‘Explaining a
Religious Anomaly:A Historical Analysis of
Secularization in Eastern Germany’, Journal
for the Scientific Study of Religion 44: 439–
22.
Geertz, C. 1993 ‘Religion as a Cultural
System’, in C. Geertz (ed.) The Interpretation
of Cultures. Selected Essays, London:
Fontana Press.
Gill, A. and Lundsgaarde, E. 2004 ‘State
Welfare Spending and Religiosity: A CrossNational
Analysis’, Rationality and Society
16(4): 399–436.
Goldthorpe, J.H. 2000 On Sociology.
Numbers, Narratives, and the Integration of
Research and Theory, Oxford: Oxford University
Press.
Hedström, P. 2005 ‘Dissecting the Social. On
the Principles of Analytical Sociology’, Cambridge:
Cambridge University Press.
Hoge, D.R., Petrillo, G.H. and Smith, E.I.
1982 ‘Transmission of Religious and Social
Values from Parents to Teenage Children’,
Journal of Marriage and the Family 44: 569–
80.
Houtman, D. and Mascini, P. 2002 ‘Why
Do Churches Become Empty, While New
Age Grows? Secularization and Religious
Change in the Netherlands’, Journal for
the Scientific Study of Religion 41(3): 455–
73.
Hox, J. 2002 Multilevel Analysis. Techniques
and Applications, Mahwah: Lawrence
Erlbaum Associates.
374 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Iannaccone, L.R. 1990 ‘Religious Practice:A
Human Capital Approach’, Journal for the
Scientific Study of Religion 29(3): 297–
314.
Iannaccone, L.R. 1991 ‘The Consequences
of Religious Market Structure. Adam Smith
and the Economics of Religion’, Rationality
and Society 3(2): 156–77.
Kelley, J. and De Graaf, N.D. 1997 ‘National
Context, Parental Socialization, and Religious
Belief: Results from 15 Nations’,
American Sociological Review 62(4): 639–
59.
Lindenberg, S. 1989 ‘Social Production
Functions, Deficits, and Social Revolutions.
Prerevolutionary France and Russia’, Rationality
and Society 1(1): 51–77.
Lindner Gunnoe, M. and Moore, K.A. 2002
‘Predictors of Religiosity Among Youth
Aged 17–22: A Longitudinal Study of the
National Survey of Children’, Journal for the
Scientific Study of Religion 41(4): 613–22.
Luhmann, N. 1982 Funktion der Religion,
Frankfurt am Main: Suhrkamp.
Mayer, J.-F. 1993 Les Nouvelles Voies
Spirituelles. Enquête sur la religiosité parallèle
en Suisse, Lausanne: Editions L’Age
d’Homme.
Mears, D.P. and Ellison, C.G. 2000 ‘Who
Buys New Age Materials? Exploring
Sociodemographic, Religious, Network, and
Contextual Correlates Of New Age Consumption’,
Sociology of Religion 61(3): 289–
313.
McLeod, H. 1998 ‘The Urban/rural
Dichotomy in European and North American
Religious History from the Eighteenth
Century to the Twentieth’, Social Compass
45 (1): 7–19.
Need, A. and De Graaf, N.D. 1996 ‘ “Losing
My Religion”: A Dynamic Analysis of
Leaving the Church in the Netherlands’,
European Sociological Review 12(1): 87–
99.
Norris, P. and Inglehart, R. 2004 Sacred and
Secular: Religion and Politics Worldwide,
Cambridge: Cambridge University Press.
Olson, D.V.A. 1999 ‘Religious Pluralism and
US Church Membership: A Reassessment’,
Sociology of Religion 60(2): 149–73.
Poos, L.R. 1995 ‘Sex, Lies, and the Church
Courts of Pre-Reformation England’,
Journal of Interdisciplinary History XXV (4
(Spring)): 585–607.
Schütz, A. and Luckmann, T. 1974 The
Structures of the Life World, London:
Heinemann.
Sherkat, D.E. 1997 ‘Embedding Religious
Choices. Preferences and Social Constraints
Into Rational Choice Theories of Religious
Behaviour in L.A. Young (ed.) Rational
Choice Theory and Religion. Summary and
Assessment, New York: Routledge.
Sherkat, D.E. and Ellison, C.G. 1999 ‘Recent
Developments and Current Controversies
in the Sociology of Religion’, Annual
Review of Sociology 25: 363–94.
Simon, H.A. 1983 Reason in Human Affairs,
Stanford, California: Stanford University
Press.
Snijders, T.A.B. and Bosker, R.J. 2000 Multilevel
Analysis.An Introduction to Basic and
Advanced Multilevel Modeling, London:
Sage.
Stark, R. and Bainbridge, W.S. 1985 The
Future of Religion, Berkeley: University of
California Press, Ltd.
Stark, R. and Finke, R. 2000 Acts of Faith:
Explaining the Human Side of Religion, Berkeley:
University of California Press.
Stark, R. and Iannaccone, L.R. 1994 ‘A
Supply-Side Reinterpretation of the “Secularization
of Europe” ’, Journal for the
Scientific Study of Religion 33(3): 230–
52.
Stolz, F. 2004 Religion und Rekonstruktion.
Ausgewählte Aufsätze herausgegeben von
Daria Pezzoli-Olgiati, Göttingen: Vandenhoeck
& Ruprecht.
Stolz, J. 2006 ‘Salvation Goods and Religious
Markets: Integrating Rational Choice
and Weberian Perspectives’, Social Compass
53(1): 13–32.
Stolz, J. 2008 ‘Secularization Theory and
Rational Choice. An Integration of Microand
Macro-Theories of Secularization using
the example of Switzerland’ in D. Pollack
and D.V.A. Olson (eds) The Role of
Religion in Modern Societies, New York:
Routledge.
Explaining religiosity 375
© London School of Economics and Political Science 2009British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Stolz,J. 2009a‘Gods and Social Mechanisms.
New Perspectives for an Explanatory Sociology
of Religion’ in M. Cherkaoui and P.
Hamilton (eds) Raymond Boudon.A Life in
Sociology: The Bardwell Press.
Stolz, J. 2009b ‘A Silent Battle. Theorizing
the Effects of Competition between
Churches and Secular Institutions’, Review
of Religious Research.
Swanson, R.N. 1990 ‘Problems of the Priesthood
in Pre-Reformation England’, The
English Historical Review 105 (417 (Oct.)):
845–69.
Theissen, G. 2000 Die Religion der ersten
Christen: eine Theorie des Urchristentums,
Gütersloh: Kaiser.
Voas, D. 2003 ‘Intermarriage and the
Demography of Secularization’, The British
Journal of Sociology 54(1): 83–108.
Voas, D. and Crockett, A. 2005 ‘Religion in
Britain: Neither Believing nor Belonging’,
Sociology 39(1): 11–28.
Wallis, R. and Bruce, S. 1995 ‘Secularization:
The Orthodox Model’ in S. Bruce (ed.) The
Sociology of Religion, Aldershot: Elgar.
Warner, R.S. 1993 ‘Work in Progress toward
a New Paradigm for the Sociological Study
of Religion in the United States’, American
Journal of Sociology 98(5): 1044–93.
Warner, R.S. 1998 ‘Religion and Migration
in the United States’, Social Compass 45(1):
123–34.
Weber, M. 1978 [1920] Economy and
Society: An Outline of Interpretive Sociology,
Berkeley: University of California
Press.
Wilson, B. 1982 Religion in Sociological
Perspective, Oxford: Oxford University
Press.
Woodhead, L. 2007 ‘Why so Many Women
in Holistic Spirituality? A Puzzle Revisited’
in K. Flanagan and P.C. Jupp (eds) A Sociology
of Spirituality, Aldershot: Ashgate.
376 Jörg Stolz
© London School of Economics and Political Science 2009 British Journal of Sociology 60(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
AUTHOR QUERY FORM
Dear Author,
During the preparation of your manuscript for publication, the questions listed
below have arisen. Please attend to these matters and return this form with your
proof.
Many thanks for your assistance.
Query
References
Query Remark
q1 AUTHOR: Stolz 2005 has not been included in
the Reference List, please supply full publication
details.
q2 AUTHOR: Statistisches Jahrbuch der Schweiz
1998 has not been included in the Reference List,
please supply full publication details.
SNP Best-set Typesetter Ltd.
Journal Code: BJOS Proofreader: Elsie
Article No: 1234 Delivery date: 19 March 2009
Page Extent: 32