Religiosity: driving force behaviour or negligible?

RELIGIOSITY:

DRIVING FORCE

BEHAVIOUR OR

NEGLIGIBLE?

MSc Economics & Governance

Faculty of Governance and Global Affairs Leiden University

June, 2019

ABSTRACT

The question whether religion causes morality is one that dates back to the Ancient Greeks. Religious prescriptions provide moral rules that religious individuals should obey. In this research, we study whether religious individuals differ from non-religious individuals in terms of moral and economic attitudes. The findings show that religious people report more hours of voluntary work and that they have a lower preference for income redistribution. Based on the results we cannot state that religious individuals are less accepting of unethical economic behaviour. In what concerns the provision of informal care, religiosity does not play a role. Considering differences between denominations, Muslims report more volunteering than

non-denomination is compared to the Christian non-denomination, no significant differences are detectable with exception of volunteering.

ACKNOWLEDGEMENT

I would first like to thank the respondents who participated in this research. Without them, this research would not have been possible.

My special and heartily thanks goes to my supervisor, Dr. Kim Fairley. Throughout the whole process I could rely on her encouragement and support, in academic and personal sense. Her critical feedback challenged me, but it also made me more critical of my own work.

Nobody has been more important to me in the pursuit of this project than the members of my family and my best friend. It is truly a blessing to have you by my side, and for that I am very grateful.

I have saved the best acknowledgement for last. I would like to thank my dearest mother for the incredible support in the past years and the past months. I consider myself lucky to be gifted with an inspiring soul that believes in me. You are the reason that I want to do more and better every day. I’d like to dedicate my final work to you, as a product of your unconditional love and support.

LIST OF TABLES

Table 1: Summary statistics: religion ... 19

Table 2: Statistical summary: Moral judgement and behaviour ... 21

Table 3: Summary statistics: control variables ... 22

Table 4: Correlations: Religion ... 26

Table 5: Correlations: Moral Judgement ... 27

Table 6: Religiosity and ethics: adjusted predictions ... 29

Table 7: Denomination and ethics: adjusted predictions ... 31

TABLE OF CONTENT

Acknowledgement ... 2

List of tables ... 3

I. INTRODUCTION ... 6

II. THEORETICAL FRAMEWORK ... 8

2.1 Concepts ... 8

2.1.1 Religiosity ... 8

2.1.2. Morality ... 9

2.2 Religiosity and Ethical behaviour ... 10

2.3. Religiosity and Economic behaviour ... 11

2.4. Religious denominations, ethics, and economic attitudes ... 12

2.5 Hypothesis ... 15

III. DATA AND METHODOLOGY ... 17

3.1 Research design ... 17

3.1.1 Data collection ... 17

3.1.2 Operationalization concepts ... 18

3.1.3 Control variables ... 22

3.2 Statistical analysis ... 23

3.4 Validity and Reliability ... 24

IV. EMPIRICAL FINDINGS ... 25

4.1 Correlation analysis ... 25

4.2 Regression analysis: Dimensions Religiosity ... 27

Moral judgements ... 27

Preference for income redistribution ... 28

Hours spent on voluntary work ... 28

Hours spent on informal care ... 30

4.3 Regression analysis: Muslims and Christians ... 30

4.4 Comparison results ... 33

V. DISCUSSION ... 34

5.1.1 Religiosity and Moral judgements – Hypothesis 1 ... 34

5.1.2 Religiosity and Income redistribution – Hypothesis 2 ... 35

5.1.3 Religiosity and Volunteering- Hypothesis 3 ... 35

5.1.4 Religiosity and Informal care- Hypothesis 4 ... 36

5.2 Theoretical implications ... 36

5.3 Practical implications ... 39

5.4 Limitations and future research ... 39

VI. CONCLUSION ... 41

REFERENCES ... 43

APPENDIX ... 50

A. Survey questions ... 50

Regression results for Table 6 ... 53

I. INTRODUCTION

“As religion and faith are being driven out of the public square, the Judeo-Christian ethical foundations that have sustained our country since its beginning, are being lost and are being replaced with a humanistic amorality, a self-centred, pragmatic indifference that will ensure

that our moral compasses will fail to point us in the right direction in the future” (Caroll, 2003)

This citation is made by Archie Carroll, a renowned writer on the subject of business ethics. is illustrates one the many considerations of religion. For years, people of faith and theologians have maintained that spirituality is the deepest seat and driving force of human motivation (Smith, 2005). Within this vision, change in human behaviour is possible whenever the spiritual idea behind it is understood. Morality is seen as something that comes from religion. This view also implicates that morality is a religious characteristic that atheists do not possess. The question whether morality requires religion is one that dates back to classical antiquity. Prominent figures have answered this question through time in their own way. Socrates have rejected this notion, Dostoevsky confirmed it, Karl Marx rejected it, and many more have given their idea on this matter (Joyce, 2002; McKay & Whitehouse, 2014). Even now in modern days, religion has become one of the most sensitive and controversial topics (Brahm, 2005). On the one hand, it calls for values that are widely supported, such as solidarity and equality. On the other hand, it is mainly associated with misery that is caused in the name of religion. Nevertheless, the majority of the world still considers themselves religious or has a strong affinity with religion (Hackett & McClendon, 2017). Zooming in on The Netherlands, a similar image becomes visible: a country where religion has played a big role throughout history (Ipenburg & Hoekstra, 2000). Until 2017, the majority of the Dutch population belonged to a religious community. In 2017, for the first time, 49% of the population indicated that they do not belong to a particular religious community (Statistics Netherlands. 2018). The same decline is visible when it comes to religious activity, such as church attendance. According to this research, the decline in religiosity applies to all religious communities with the exception of Muslims. In fact, this group have practiced their faith more in recent years (Huijnk, 2018).

The current trend of secularization cannot take place without evoking effects in society, since religion impacts every aspect of it (Guiso et al, 2003). Even the economy is affected by religion, as religion influences behaviour which, in turn, determines the developments in the economy. Besides, all monotheistic religions have prescribed rules regarding economic activities of individuals (Wilson, 1997).

In a predominantly religious world, and until recently country, it becomes questionable what implications the decline in religiosity will have for society and the social cohesion of the economy. More specifically, this research seeks to address the following question: “To what extent does religiosity affect morality and economic behaviour?”. The aim of this study is to determine the extent to which being religious affects individuals behaviour by means of five dimensions of religiosity. Moreover, the effects of the Muslim and Christian denomination will be assessed and compared.

After Christianity, Islam has the most adherents mondial as well as national in The Netherlands (Hackett & McClendon, 2017; Statistics Netherlands, 2018). The majority of studies conducted in the field of religion and its effects, focus on Christianity. However, far too little attention has been paid to Islam. There are almost no empirical researches that focus on Islam and how it relates to ethics. Apart from Chapra (2009), there is a general lack of research in this field. Therefore, this study contributes to science by investigating a group where relatively less is known and written about. It also provides an important opportunity to advance the understanding of a religion that is mainly associated with conflicts, but also has a growing number of adherents (Lipka & Hackett, 2017). By including also Christianity, the effects of the two biggest religions can be examined and compared to each other.

In addition, this study has a societal relevance since it is an actual topic. The world has been confronted with a lot of events that are related to religion lately (Pontifex, 2018). Testing the effect of religion on behaviour allows us to understand whether religion can account for these events or that the effect of religion on individuals behaviour is negligible.

The following chapter begins by laying out the theoretical dimensions of the research, and looks at how religion is related with ethics and economics. Theories are examined with regard to religiosity in general as well as Islam and Christianity and how these affect individuals’ behaviour. This chapter ends with hypotheses that function as the fundament in this study. The third chapter is concerned with the methodology used for this study to explore religiosity and its effect on its adherents. The fourth section presents the findings of the research, focusing on the effect of the dimensions of religiosity on moral and economic attitudes and the Muslim and Christian denomination on the same attitudes. Chapter 5 discusses the findings and assesses the accuracy of the hypotheses. The final chapter concludes the thesis by reflecting upon the findings and critique of the findings.

II.

THEORETICAL FRAMEWORK

There have been numerous studies investigating the relationship between religion and economic behaviour or religion and morality (Wilson, 1997; Guiso, Sapienza & Zingales, 2003; Jackson & Fleischer, 2007; Parboteeah, Hoegl & Cullen, 2007; Noussair, Trautmann, Van de Kuilen & Vellekoop, 2013; Shariff, Piazza & Kramer, 2014). Whether religiosity affects individuals is confirmed by almost all scholars, but still encounters criticism (Shariff & Norenzayan, 2007; Glaesar & Sacerdote, 2008; Chapra, 2009; Benjamin, Choi & Fisher, 2010). Important critics were Marx and Feuerbach, who both questioned the direction of causality regarding religion and behaviour. In Criticism of Hegel’s Law, Marx (1844) stated the famous quote of religion being peoples opium. Within this perspective, religion is merely an expression of reality and societal problems. It is not an independent factor that can influence behaviour, because man made religion and not the other way around. Even though this view is not exceptional, we find most scholars do assign an exogenous role to religion in its effect on society and individuals. Sartre (1985) and Nietzsche (1889) for example shared the view that a Godless world would lead to immorality and chaos. According to this view, religion is an important determinant of the way individuals interact. Since then, this relationship is one that is widely discussed in academic discourses. In the following sections, the main points will be discussed of the debate whether religiosity affects morality and economic behaviour respectively and in particular Islam and Christianity.

2.1 Concepts

2.1.1 Religiosity

In the academic discourse on this topic, there are a set of conceptual limitations that are not taken into account by researchers. One important aspect that is generally overlooked is the conceptualization of religiousness and morality. Many investigations fail to fractionate the two concepts into theoretically grounded units (McKay & Whitehouse, 2014: 448). The concept of religion does not reflect a real natural structure but may comprise an arbitrary gathering of disparate features. One may treat religion as a unidimensional construct; it has provided solid findings that are often constant across cultural contexts and religions (Saroglou & Cohen, in press). However, there is also evidence that religiosity can be conceived as multidimensional (Hill, 2005). The use of a multidimensional approach is rooted in theories of religion that have been repeatedly revised and supported by different scholars since the early 1900s (Glock, 1962; Hill, 2005). This method is favoured because it provides nuanced information on how religion

affects individual lives. Many authors have attempted to identify religion by pointing out its fundamental elements. Even though there is some variability across the number and subcomponents of these elements, there is also a notable consistency. Saroglu (2011) has put forward a psychological model of the Big Four Religious Dimensions that integrates previous formulations of amongst other Glock (1962), Hervieu- Léger (1999) and Hinde (1999). In brief, Saroglu distinguishes between:

1. Believing: beliefs relative to what many people consider as being transcendence. 2. Bonding: self-transcendent experiences that bond the individual with the transcendent

reality. This occurs often within a ritualized framework.

3. Behaving: specific norms and moral arguments that define what is right from a religious perspective. This morality can differ from societal morality.

4. Belonging: affiliating and identifying with a certain community.

Meanwhile, this model is complemented by a fifth dimension like is done by Pearce et al. (2016) and Kirchmaier et al. (2018) and sometimes even by a sixth dimension (Cornwall et al, 1986). These models use the same base, but differ in the concretisation of the dimensions. The five-dimensioned models distinguish between church membership, belief in God, church attendance, prayer and belief in theological concepts. Models based on four dimensions are less specific and combine multiple aspects of religiosity within one dimension. The classic four-dimension model is criticized because there is no distinction between belief in theological concepts, that vary across religious traditions, and belief in God. Also, they are criticized because of the combination of private practices, (e.g. prayer) and religious salience, such as church attendance, in one dimension. These aspects are different enough to keep separate in a model that is likely to be a better fit in an overall model of religiosity.

2.1.2. Morality

One salient feature of Saroglus’ model is the inclusion of morality as a key dimension. He emphasizes that morality refers to religious morality and not morality as an external. Although overlap is possible, the two moralities are independent of one another. In contrast, McKay and Whitehouse differ in their interpretation and state that “this stipulation implies that any inquiry into the effects of “religion” as a whole on “morality” as a whole may be a circular, and therefore futile, enterprise” (2014: 449). This study will be based on the concepts of social morality, referring to a system of internalized social moral rules/standards and values that are

2.2 Religiosity and Ethical behaviour

Different views exist regarding religiosity and ethical behaviour that are not necessarily consistent with each other. Hood et al. (1996: 371) therefore describe their view on this topic as “something of a rollercoaster ride”. The relationship between religion and morality dates back to the Ancient Greeks, where Socrates famously asked whether holiness is loved by the gods because it is holy or whether holiness is holy because it is loved by the gods (Joyce, 2002). Although he preferred the first option, many others after him argued that morality is dictated by God. This view enables conservatives to ascribe weakening moral standards to the increase of secularism (McKay, Whitehouse & Dolores, 2015). However, despite declining religiosity, a recent survey of the Pew Research Centre shows that a majority of respondents agreed that believing in God is essential to morality (PRS, 2014). In the United States, for instance, 53% of the respondents agreed that belief was fundamental to be a good person. The crucial question is how different believers are from non-believers in their morality. On the one hand, it is argued that there would be no reason for believers to be more moral. After all, it is religion that preaches tolerance but feeds intolerance and cultivates contrasting emotion as love and wrath (Van Stokkom, 2012). Besides, religion is also the source of extremism and fundamentalism. On the other hand, there is the obvious effect of religion on behaviour through ethical standards that are imposed on believers. In the Christian religion, for instance, the Ten Commandments provide ethical rules about what is considered right and wrong. Additionally, Islam has clear rules about how one must act to live a proper pious life that is codified in the Quran and Hadith literature, reported sayings of the prophet Muhammed that are received as a source of moral guidance and religious rules (Ali, 2013). Similarly, Buddhism also prescribes that a good life is an ethical life where one lives in coexistence with other humans and nature and is based on compassion (Dalai Lama, 1974). Ali et al. (2000) go even further and suggest that the major monotheistic religions all contain moral premises that provide similar moral guidance. Strong beliefs in religion thus suggest that people have a fundament for a moral life that is built on religion (Vitell & Paolillo, 2003). Strong beliefs imply that the religious principles are taken as guides to discourage yourself and others engaging in unethical behaviours. Furthermore, an important factor that plays a part in the discouragement is the fear of being caught by an omniscient God and being punished (Parboteeah, Hoegl & Cullen, 2008). This is part of the religious prosociality explanation, referring to voluntary behaviour with the aim to intentionally benefit others (e.g. sharing, volunteering) at personal cost. The rationale behind this is that belief in a morally judgmental Supernatural power complements the social scrutiny that theists

Religious priming research reveals that a monitoring God works similarly as social monitoring, thereby promoting prosocial behaviour (Sedikes & Gebauer, 2010). Religious individuals feel pressured by the presence of a punitive God and within the group which ensures that their behaviour corresponds to the demanded values. Thus, in that sense religiosity discourages selfish behaviour that might disrupt the unity of the group they belong to (Shariff & Norenzayan, 2007). According to this theory, individuals’ morality is not entirely intrinsic but a consequence of pressure and reputation.

The relationship between religion and morality have been criticized a lot, for one because “nice” behaviour is not inherently linked to believing (McKay, Whitehouse & Dolores, 2015). Galen (2012) calls this the Religion- Morality Stereotype when religious individuals are seen as more moral and trustworthy. Empirically, there is indeed little agreement on whether adherence to a religion is correlated with more or less ethical behaviour (Kirchmaier, Prüfer & Trautmann, 2018). Empirical assessments are complicated due to the fact that morality differs across religions (Shariff et al., 2014) and religious and non-religious individuals. Religious affiliation may affect opportunity sets, which influence revealed behaviour, but not necessarily attitudes.

2.3. Religiosity and Economic behaviour

Besides the possible effect of religiosity on morality, the effect of religiosity on economic behaviour has been studied thoroughly. Research has uncovered some persistent relationships between religion and economic behaviour (Kirchmaier, Prüfer & Trautmann, 2018). A useful starting point is Mangeloja’s (2005) research that delves into the secularisation hypothesis, i.e. that economic development by increasing education, economic welfare, and urbanization have led to decreased importance of beliefs and religious activities and a minimised role of organised religion (Iyer, 2005; Jackson & Fleischer, 2007). Although this hypothesis has been challenged, it continues to influence the ways in which many think about religion in modern society. Guiso et al. (2003) use the World Values Survey to identify the relationship between the intensity of religious beliefs and economic attitudes. The research shows that on average, religious beliefs are associated with ‘good’ economic attitudes. The identified effects differ considerably across religions, but in general, Christianity has the biggest positive association with economic attitudes favourable to economic growth compared with other religions. These findings support Weber’s original analysis which found that increased religious intensity was a significant factor

beliefs have an influence on individuals’ behaviour which, in turn, affects economic growth through economic choices that individuals make. There is also a hypothesis stating that religious moral codes affect economic behaviour since all (monotheist) religions have prescribed norms concerning economic activity, leading to more sentient individuals. Depending on the focus of the research, different economic attitudes of individuals are considered in researches. Most common concepts are charity work, redistribution, altruism, charity work because there are also topics that can be linked to religious standards.

As concerns voluntary work and charity work, Statistics Netherlands has published a study that shows that frequent church or mosque visitors are more active as a volunteer compared to individuals that never or rarely visit these places. This applies to all denominations, but Christianity and Islam were the main points of focus (Statistics Netherlands, 2015). This research is not an exception: most studies agree on the positive effect of religiosity on voluntary work. People with higher levels of religiosity have been repeatedly shown to be more likely to donate time and money to non-profit organizations than individuals without or no religiosity. The type of organization does not play a role, it could be religious as well as secular (Wilkins-Laflamme, 2018).The time that is donated could be in the form of voluntary work, but informal care is also mentioned as a type of philanthropic activity where time is donated. Religious communities often encourage these activities within their community and provide resources to make the activities possible. From a religious perspective, individuals feel obliged to participate in these activities, because it is perceived as an act of selflessness. As regards to charity and redistribution, it is assumed that irrespective of the religion, there may be a cleavage between the religious and non-religious (Scheve & Stasavage, 2006). Basten and Betz (2013) find evidence that (Reformed) Protestantism has a significant effect on the support for redistribution. Similar effects are also visible if we look to adherents of Islam.

2.4. Religious denominations, ethics, and economic attitudes

If we look at the relationship between religiosity and moral judgement of economic attitudes it is the most obvious to look at Christianity and Islam since these two religions provide rich literature on approaches to economic systems that are rooted within their own religious teaching and moral codes (Jackson & Fleischer, 2007). However, within these two religions, different approaches are noticeable when it comes to economic behaviour with regard to ethics.

Partly, it stems from the difference in their effect on the daily life of adherents: the demands of Islam in everyday life exceed those of Christianity. The demands of Islam makes believers

It follows that these demands inevitably make it difficult for Muslims to isolate the spiritual from the economic domain in their lives, as both are constantly alternating. As Rice (1999: 346) states: “In Islam, it is ethics that dominates economics and not the other way around”. Several studies have reported that Islam differs from other monotheistic religions in the sense that it describes very concrete how individuals attitude should be towards different economic concepts (Keller, 1994; Rice, 1999; Wilson, 1997).

Chapra (2009) states that Islam is strongly intertwined with morality. In fact, she argues that Islam is the uniting force in the Muslim world that has the charisma to attract masses in spite of their diversity and stimulate them in acting righteously. The power behind this force lies in the fundament of Islam, the Arabic word for peace. The Islamic world view gives primary importance to socioeconomic justice, moral values, solidarity, and brotherhood. However, within Islam there is an assumption that most people tend to have a behaviour between two extremes, altruism and selfishness. Hence, a constant effort needs to be made for moral uplift without using force for moral uplift: “There shall be no compulsion in religion” (Quran, 2:256). This refers to individuals that need to work on themselves without coercion and pressure. Eventually, like in Christianity, this pays off in the Hereafter, the ultimate destination of the followers where they will be accountable before God. Their wellbeing will depend then on the way they have lived in this world and whether they have fulfilled their obligations. The assumption thereby is that it would be irrational for a person to sacrifice eternal wellbeing for the sake of relatively short term joy (Chapra, 2009: 8).

An unique concept that differentiates between Islam and other religions that is described by Jackson and Fleischer (2007), is the concept of Homo Islamicus. Just as economic theory is described on the basis of the concept of the Homo Economicus, referring to rational individuals that seek to maximize their utility, the concept of Homo Islamicus is central to the Islamic religion. The Quran prescribes what actions and type of behaviour are considered good and evil. The Homo Islamicus points to the ideal behaviour where one works hard, promotes generosity avoids waste and ostentation, dissuades harmful externalities and also trades using fair prices. Chapra and Ahmed (1992) put it as a transformation of the selfish and greedy Homo Economicus into an epitome of moral behaviour. Economic attitudes of the Homo Islamicus are steeped in moral considerations. Therefore, conducts that involve gambling, speculating and destructive competition are banned.

Another economic notion that has been given significance is that of zakat, which is redistribution. Islam is the only religion that has spelled out concrete rules concerning income redistribution (Wilson, 1997). This system is similar to what contemporary is known as a progressive tax: a system where (high) taxes are imposed on wealthy individuals to assist those that are poor. The rates against which individuals should give zakat are defined in the seventh century, which makes it controversial since it could be seen as outdated. Several analysts, under which Kuran (1995), have stated that redistribution, as described in Islam, may be a powerful force in the alleviation of poverty within a developing economy. Moreover, it could be more effective than the welfare and redistribution systems that are used in modern western states. The moral character of Islamic attitudes is also visible in the ban on interest, riba. The rationale behind this ban is that the practice of interest lead to social friction, as borrowers were pushed into slavery. It has socially damaging characteristics and is also seen as unfair towards borrowers (Fazlur, 1964).

Much of the reviewed literature regarding Islam resembles the Christian attitudes to economic theory. The Bible is the ultimate source which determines Christian moral ideas as well as Christian economic thinking. The New and Old Testament contains insights on economic matters. Based on these sources a variety of concrete guidelines are established about how people should behave and what the implications would be for their economic behaviour (Jackson & Fleischer, 2007). For example, the concept of altruism is emphasized and is of great importance within the Christian tradition. Altruism indicates how generous and caring one is with regard to others. It relies on the belief that individuals are not complete rational agents that ought to maximize their own utility, but that individuals also act for the benefit of others without expecting something in return (Laury & Taylor, 2008). Christianity speaks highly of the concept and puts it down as the moral thing to do for a believer: “Do nothing from selfishness or empty conceit, but with humility of mind regard one another as more important than yourselves; do not merely look out for your own personal interests, but also for the interests of others” (Philippians 2:3). On the other hand, it is also notable that redistribution is mentioned less and more abstract. Wealth distribution of material goods and what amount the rich should give to the poor do not have explicit criteria.

In general, for many Christians in modern society religious belief is primarily a private matter. They may belong to a church and attend gatherings weekly, albeit monthly. Although their beliefs may shape social attitudes and personal values they also may be reluctant to declare these values in their business dealings (Wilson, 1995). Religiosity may affect personal decisions

2.5 Hypothesis

Hypothesis 1: Religiosity leads to stricter moral values

Based on the concepts and theoretical framework, hypotheses can be drawn that will operate as the basis of this research. The notion that religion is a precondition for morality is deeply ingrained in society (McKay & Whitehouse, 2014). Believers are provided with religious frameworks with norms and standards that need to be followed. The belief that there is an all-seeing God reinforces the senses to do good. Adherents of religion are frequently convinced that morality is a condition you need to meet before you can claim religiosity. The hypotheses will be tested by looking at the five dimensions of religiosity as explained in 2.1.1 Religiosity. Thus, the following hypothesis emerge:

Hypothesis 2: Religiosity creates a higher preference for income redistribution

An important part of all religions, and in particular Islam and Christianity, is the principle of charity and sharing (Wilkins-Laflamme, 2018). Islam knows a direct form of income redistribution that is one of the five fundamentals an individual must comply with. The Christian tradition does not have an explicit rule about income redistribution in terms of taxing the rich to provide for the poor, but there is considerable evidence in the Bible for providing for the poor and the act of charity (Wilson, 1997). The religious procedure that should be followed in order to provide for the poor is set up according to its own traditions. Redistribution is seen as a necessity to provide for the weak and points out the importance of care for each other. Therefore, believing should affect the attitude of one towards income redistribution:

Religiosity leads to stricter moral values H1.1 Church members hold stricter moral values

H1.2 Individuals that attend more often in church hold stricter moral values H1.3 Individuals that pray more often hold stricter moral values

H1.4 Individuals with a stronger belief in God hold stricter moral values

H1.5 Individuals with a stronger belief in theological concepts hold stricter moral values

Religiosity creates a higher preference for income redistribution

H2.1 Church members have a higher preference for income redistribution

H2.2 Individuals that attend more often in church have a higher preference for income redistribution. H2.3 Individuals that pray more have a higher preference for income redistribution

Hypothesis 3: Religiosity causes an increase in the time spent on volunteering

Further to this, care over one another is also expressed in the performance of voluntary work. Most religions encourage altruistic behaviour and values. It is therefore not surprising that religion has been closely linked to volunteering throughout history (Wuthnow, 1990). Existing studies have shown that there is a strong relationship between religiosity and participating in philanthropic activities in general. The nature of these activities can be different, ranging from charitable giving to volunteering in the mosque/church, educational organizations, and political platforms. Especially church attendance seems to affect the time that is spent on voluntary work. The following hypotheses can be derived:

Hypothesis 4: Religiosity causes an increase in the time spent on informal care

Finally, the effect of religiosity on hours spent on informal care will be tested. Informal care is the most private type of voluntary work. Every religion has stressed the importance of contributing to society and taking care of those who need it. Religious values like altruism and responsibility for the community imply that religious individuals should play their role in society and take care of the weakest amidst them. Moreover, what makes informal care interesting is that its private character reveals the intentions of individuals: it could been done out of moral reasoning but also for public acknowledgement (Wilkins-Laflamme, 2018).. Therefore, the next hypotheses are considered:

Religiosity causes an increase in the time spent on volunteering

H3.1 Church members spend more hours on voluntary work

H3.2 Individuals that attend church more often spend more hours on voluntary work H3.3 Individuals that pray more often spend more hours on voluntary work

H3.4 Individuals with a stronger belief in God spend more hours on voluntary work

H3.5 Individuals with a stronger belief in theological concepts spend more hours on voluntary work

Religiosity causes an increase in the time spent on informal care

H4.1 Church members spend more hours on informal care

H4.2 Individuals that attend church more often spend more hours on informal care H4.3 Individuals that pray more often spend more hours on informal care

H4.4 Individuals with a stronger belief in God spend more hours on informal care

III. DATA AND METHODOLOGY

This chapter describes the methodology that is used for this research study. The purpose of this study was to determine to what extent religiosity affects moral judgement and economic attitudes. First, a description of the data will be provided along with a description of the sample and procedures used for the study. Second, the data analysis section will define the used variables and the operationalization and conceptualization of the theoretical concepts that are central in this study. Finally, the empirical strategy will be specified and how this affects the validity and reliability of the research.

3.1 Research design

3.1.1 Data collection

In view of all studies that have been mentioned, there seems to be evidence that religiosity affects society, by means of religious precepts with regard to all aspects of life (Guiso et al, 2003). To evaluate the causality between the concepts, a detailed analysis of the effects of religiosity on the ethical judgement of economic attitudes of individuals will be performed. 263 individuals, with a Dutch nationality or/and living in The Netherlands, have completed an online survey regarding morality, religiosity and economic attitudes. The target population for this study were Dutch individuals that are Muslim, Christian and non-religious. The respondents are recruited online and partly in the mosque, given that these people form the focus of this study. The downside of spreading a survey online within the network of the researcher is that it attracts respondents with a similar background. To combat this risk, the survey is shared in as many social media groups and platforms that differ in their target group. Table 1 and Table 2 report how the participant base is set up.

This study is based on micro-level data that is gathered through an anonymous questionnaire (see Appendix, Survey questions). There are two reasons for choosing a questionnaire to collect data. Firstly, the research aims to compare the findings with those of Kirchmaier et al. (2018) who conducted similar research into the effect of religiosity on moral values and economic attitudes of individuals and their parents. The focal point in their study is Christianity, and in particular the Catholic and Protestant denominations. Accordingly, this is where this current study differs from Kirchmaier et al. as the target group in this research are Muslims rather than Christianity. Christianity is used to compare it with the findings of the Islamic subgroup. The transgenerational analysis is also left out in this paper.

To ensure consistency and uniformity, the questionnaire of Kirchmaier et al. is used to measure the ethical judgements of religious individuals. The second reason for using a questionnaire is that it makes it possible to look for statistically significant trends. Moreover, responses can be gathered relatively quick from a large number and efficiently (De Vaus, 1991; Fink, 1995). This allows us to test the effect of religiosity on moral judgement as well as to find out whether there are differences across the denominations.

The survey is assembled in Qualtrics, an online survey software that allows to create surveys. This particular survey website is chosen because it is user-friendly, easy accessible for the respondents, and the data could be transferred to an Excel and Stata file. The final survey consisted of 26 questions: 23 multiple choice and 3 open-ended questions. It was possible to code the answers in Qualtrics beforehand, so that has been done in accordance with Kirchmaier et al. (2016). Some questions that were included in the original study are left out, because they were irrelevant for the focus of this research. Kirchmaier, Prufer, and Trautmann have categorized three types of questions. The first category concerns the religiosity of individuals, the second category their attitudes and judgements of economic activities, and the third provides information about individual characteristics of the respondents, such as their education level, age, and gender.

3.1.2 Operationalization concepts

The questionnaire on religion that the participants have completed encompass data on the religious beliefs and activities of the respondents at the time of the survey. Religiosity is measured by means of five dimensions, that function as the explanatory variables in our analysis. The five chosen dimensions also correspond with the model of Saroglu (2011), with the only difference that ‘Belief” is broken down into two dimensions: belief in God and belief in theological concepts. The example of Kirchmaier, Prüfer, and Trautmann (2016) is followed, where they state that belief in God does not imply belief in theological concepts.

Table 1 provides the summary statistics of the responses of the respondents in the five dimensions. The first column (1) reports the means for each variable over all observations. In column (2) the means are reported for members of a religious community and (3) and (4) report the means for the subsample Muslims and Christians, respectively. The first of the five dimensions of religion on which we focus in this work is Belief in God.

Table 1

Summary statistics: religion

It indicates the strength of belief in God and is reported by a six-point scale that starts with 1: “I do not believe in God” to 6: “I believe without any doubt that God exists”. This dimension

Observations All ₍₁₎ Church members (2) Muslims (3) Christians (4) Religious status Church membership 263 72% Muslim 136 52% 37% Christian 50 19% 14% Other Faiths 3 1% 1% Church attendance Daily 263 6% 8% 9% 2%

More than once a week 263 13% 17% 14% 26%

Once a week 263 8% 11% 7% 24%

At least once a month 263 13% 17% 20% 8%

Only on special days 263 24% 24% 26% 20%

Rarely 263 21% 18% 19% 18%

Never 263 16% 5% 5% 2%

Private prayer

Daily 263 56% 73% 81% 52%

More than once a week 263 6% 7% 4% 18%

Once a week 263 3% 4% 1% 12%

At least once a month 263 3% 3% 1% 10%

Only on special days 263 4% 4% 4% 2%

Rarely 263 9% 4% 5% 2%

Never 263 18% 3% 2% 4%

Belief in God

Degree of belief in God (0-5) 263 3.70 4.75 4.95 4.24

Strong belief in God 263 63% 88% 96% 64%

Belief in theological concepts

Believe in life after death 263 67% 85% 89% 76%

Believe in the existence of heaven 263 70% 93% 99% 80%

Believe in the existence of hell 263 67% 93% 99% 68%

Believe in the existence devil 263 64% 87% 91% 68%

Believe that Adam and Eve existed 263 68% 91% 98% 76%

Believe in Bible as the word of God 263 68% 93% 100% 70%

Believe that prayer makes sense 263 78% 97% 99% 88%

Belief in theological concepts (0-7) 263 4.83 6.39 6.75 4.95

Following on from this dimension, the second dimension we include is Belief In Theological concepts. It is the acceptance of a standard set of religious beliefs, including (1) belief in life after death, (2) existence of heaven, (3) existence of hell, (4) the devil, (5) Adam and Eve, (6) Bible as the word of God and (7) belief in prayer. The strength of the belief in these subdimensions is measured by the answers on the seven questions, reported on a four-point scale ranging from 1: “Yes, I believe” to 4: “I don’t know”. We also define the dummy variable of strong and weak belief in theological concepts based on the median split of the answers to the strength of belief in theological concepts.

Church membership is the third dimension of religion. Respondents had to answer whether they consider themselves as a member of a church or religion community. In the dataset, the different denominations are taken into account.

Following this, is the dimension Church attendance that is present in every model regarding religiosity. This dimension captures the intensity of belief by looking at the frequency of attendance at religious services is looked at. This is reported on a six-point scale ranging from 1: “Every day” through 5: “Only on special religious days” to 7: “Never”. The answers are aggregated to obtain a 3-points scale measure with 0: “No attendance”, 1: “Less than once a week” and 2: “More than once a week”.

The final dimension, Prayer is a private practice that is less visible and may thus have different underlying determinants. Similar as with church attendance, a a three-point scale is defined, ranging from 0: “No attendance”, 1: “Less than once a week” and 2: “At least once a week”.

Moral judgement and behaviour

The dependent variables in this study are the ethical judgements and the economic attitudes that are shown in italics in Table 2. Ethical judgement is measured on the basis of five elements related to economic practices. The variable moral judgement, is an aggregate indicator of the five moral judgements. Respondents were presented several unethical economic behaviours of which they needed to assess whether it is justified or not on a scale of 0: “Always justified” to 10: “Never justified”. The five questions concern (1) Claiming state benefits which you are not entitled to; (2) Cheating on tax; (3) Lying in your own interest; (4) Accepting a bribe; (5) Not paying public transport fare. Subsequently, the average of the five questions is taken and renormalized to the Moral Judgement indicator, ranging from low (0) to high ethics (10).

Table 2

Statistical summary: Moral judgement and behaviour

The second measure involves the attitude of the respondents towards income redistribution. Every religion has prescribed the importance of generosity and sharing, for this reason, a relationship between religion and attitude towards redistribution is plausible. The respondents have indicated their views on a scale ranging from 1 to 5, where 1 means that differences in income should increase and 5 that differences in income should decrease.

In contrast to the previous two measures, the next measures are based on actual behaviour rather than stated preferences. The third measure covers the average time per week that is spent on voluntary work. Voluntary work includes informal care as well as work that is performed in a sports club or other social associations. In the analysis of volunteering a dummy variable is included for individuals that perform some voluntary work. In line with this, the fourth measure is the average hours performed on informal care per week. The distinction is made because the character of informal care differs strongly from that of voluntary work in general. Volunteering in organizations might have a strong social visibility element, certainly when it is performed within a close community. Informal care to the contrary is less visible and therefore may be considered less socially rewarding. Hence, testing informal care apart from voluntary work is a stronger test of charity love.

Observations All (1) Church members (2) Muslims (3) Christians (4) Moral judgement

Social benefit fraud 262 1.53 1.58 1.55 1.64

Cheating on tax 263 2.04 2.13 1.91 3.21

Lying out of self interest 263 2.74 2.48 2.40 2.71

Accepting bribes 263 1.52 1.57 1.49 2.21

Not paying transport fares 263 3.06 2.71 2.49 4.50

Moral judgement 263 2.18 2.10 1.97 1.70

Preferences for redistribution

Preference lower income inequality 263 1.74 1.85 1.85 1.79

Public charity

Hours spent on voluntary work per week 60 5.98 6.50 7.00 5.71 Private charity

3.1.3 Control variables

To isolate the effect of religiosity from other confounding effects, we control for several individual characteristics (Control A) and socioeconomic background (Control B). Table 3 provides descriptive statistics of all control variables included. The statistics are presented in percentages of the whole, except for age that is reported in averages. It is essential to control for these characteristics, as previous studies have reported that it is not just religiosity that acts on judgements of individuals. Therefore, we control for age and gender. In an analysis of charitable contributions, Hrung (2004) found that participation in religious organisations increases with age. In the same vein, research has also shown that morality changes with age. As one grows older, one becomes more aware of doing what is morally right (Wright, 1985). With regard to socioeconomic background variables, we control for marital status, income, educational an occupational status and number of children (Control B). These variables are possibly related to an individual’s economic behaviour and their moral judgement of this. As for income, education level, and marital status: previous studies have reported that they are related to religiosity as well.

Table 3

Summary statistics: control variables

Observations All (1) Church members (2) Muslims (3) Christians (4) Control A Age 263 26.15 25.57 25.65 26.34 Male 263 30% 26% 24% 28% Control B

Living with a partner 263 22% 24% 25% 20%

Having children 263 23% 24% 24% 22% Married 263 16% 22% 24% 39% Divorced 263 3% 3% 3% 7% Higher education 263 55% 50% 49% 58% Self-employed 263 10% 11% 10% 12% Annual income 263 20168 21187 20600 14423

3.2 Statistical analysis

The purpose of this research is to study whether religious people hold stricter views regarding a set of economic attitudes and their moral judgement of this compared to non-religious people. In other words, we want to determine whether there is a causal relationship between religiosity and moral attitudes by comparing how much variability exists in the means of the moral judgements of the groups and across denominations. Different methods are used to investigate this relationship, but on behalf of uniformity in this research, the same method that is used by Kirchmaier et alia will be used for comparable findings.

Firstly, a correlational analysis is used to determine the mutual correlations between the dependent variables and the dimensions of religiosity. The Spearman rank correlations are used to measure and calculate the strength and the direction between the dependent variables. It uses the rank of the variables to measure the association, instead of the variables themselves (Walker & Maddan, 2012). The second part of our analysis consists of Tobit regressions. It is common in the field of ethics to use similar statistical analysis. That is because of the nature of the data: ethical concepts can be investigated best by questioning people and capturing their answers using a Likert scale. The Likert scale is a type of survey scale that ranges from two extreme attitudes regarding a certain topic. The use of the scales is particularly popular in ethical and psychological studies because it is a reliable way to measure behaviours and opinions of individuals. Respondents make a choice between one of the ordered answers that are linked to a number (Likert, 1932). However, an important point is that the observations are censored. The assumption is that respondents that answer a question miss categories to choose from: they are limited to the Likert items that are provided by the researcher. They may choose 5 on the scale when they actually would have chosen a 6 or even a higher score. For this reason, the Tobit model, also known as the censored regression model, is used instead of the classical regression model. Unlike the conventional regression methods, Tobit regressions take into account that the dependent variable is censored, preventing biased parameters (Henningsen, 2010). The dependent variables that are employed in this study are censored: moral judgement (0-10), income redistribution (0-5), hours spent on voluntary work (140) and hours spent of informal care (140). For each of these dependent variables a Tobit regression is run. Two types of regressions are run: one with Control set A and one with Control set B of Table 3. Along with the regressions, the average predictive margins are calculated for each variable to provide a general overview and to make the results more tangible.

The second type of statistical analysis are regressions to estimate again the effect of religiosity on the ethical measures, but this time distinguishing in religious denominations, looking at Islam and Christianity. In this category, the correlations of Muslims and Christians with the four dependent variables are examined and compared Similarly, two models are included with Control set A and B.

All the analyses and the preparations are conducted in STATA. The dataset did not need many adjustments. Due to the forced response option in Qualtrics, it is not possible to finish the questionnaire without answering all questions. Therefore, an additional check for missing values was not necessary. There is also no issue of outliers, because of the multiple choice format.

3.3 Validity and Reliability

Threats to the validity and reliability of the survey instrument are possible, but preventive measures were taken to reduce that probability. The purpose of establishing validity and reliability in research is essential to ensure that data is replicable, and the results are accurate (Mohajan, 2017). Validity and reliability in survey research are often assumed, but different factors need to be taken into account to actually maintain it. Validity in survey research and quantitative research is concerned with the accuracy of the measurements and is often associated with the representativeness of the sample. However, it has also to do with the survey design and to what extent the questions that are asked measure what is supposed to be measured. The questionnaire of Kirchmaier et al. (2016) that is used as a basis is adapted a little to make it fit the focus of this research and to make it more appropriate for the target group. The changes were mainly cosmetic so it is doubtful that they would have affected the validity of the survey. Threats to the reliability, however, are higher. Reliability has to do with the repeatability of the research and the consistency of results over time. The reliability of measurement is called into question when it is prone to random error. Ideally, a large random sample increases the likelihood that the responses will reflect the population accurately and thus increases the reliability. But there lies a threat in the use of web surveys. The principles of probability sampling cannot be applied (Betlehem, 2008). Samples rely on self-selection of respondents rather than probability sampling. This can impact the survey results. On the other hand, a completely random selection would not have been possible anyway given the fact that we mainly focus on Muslims, Christians and non-religious individuals. Targeted sampling, therefore, is inevitable.

IV. EMPIRICAL FINDINGS

This chapter focuses on the results derived from the questionnaire data. Firstly, the correlations between the different aspects of religion and the different behaviours will be shown in Table 4 and Table 5 respectively. Subsequently, Table 6 presents an overview of the marginal effects for each explanatory variable and the significance of the variables based on the results of the Tobit regressions. The numbers express the associations between the dependent and independent variable. The detailed and extensive results of the regressions are presented in the Appendix. Table 7 shows the results obtained on the effect of the Muslim and Christian denomination. The results of these tests partly duplicate earlier efforts that include religious dimensions and ethics.

4.1 Correlation analysis

Table 4 provides an overview of the Spearman Rho correlations among the different measures of religiosity. The grey coloured cells show all statistical significant correlations at the 0.01 level. With exception of the correlations with Christianity, all variables are significant correlated with each other. Looking at the correlation coefficients, the rule is that the closer the Spearman correlation is to -1 or +1, the stronger the monotonic relationship between two variables (Schober, Boer, & Schwarte, 2018). If this rule is applied to the data, it becomes visible that the majority of the correlation coefficients are higher than 0.60, which indicate strong correlations. There are also correlation coefficients higher than 0.80, meaning that the strength of the correlation is very strong. This is not very striking, as the belief in religious concepts is does not stand alone. Individuals have a system of beliefs where the concepts are included in. The coefficients of membership show all a strong and positive correlation with religious aspects. The correlations with the degree of belief in God similarly show strong to very strong positive relationship with the remaining measures. Taken together, these results show a positive and predominantly strong correlations among the different measures.

Table 4

Correlations: Religion

Note: Spearman’s rho for pairwise correlation with p < 0.01

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Member 2. Muslim 0.65 3. Christian 0.30 -0.50 4. Church attendance (0-2) 0.51 0.31 0.19 5. Prayer (0-2) 0.71 0.52 0.19 0.60

6. Degree of belief in God (0-5) 0.78 0.69 0.05 0.52 0.79

7. Strong belief in God 0.73 0.69 -0.00 0.47 0.74 0.97

8. Belief in theological concepts (0-7) 0.74 0.66 0.03 0.48 0.71 0.85 0.81

9. Strong belief in theological concepts 0.61 0.58 -0.01 0.38 0.59 0.70 0.69 0.94

10. Believe in life after death 0.55 0.47 0.08 0.35 0.54 0.66 0.61 0.79 0.71

11. Believe in existence of heaven 0.71 0.61 0.08 0.42 0.65 0.81 0.76 0.83 0.64 0.72

12. Believe in existence of hell 0.71 0.66 -0.01 0.49 0.69 0.84 0.82 0.87 0.71 0.66 0.85

13. Believe in existence devil 0.64 0.57 0.03 0.36 0.58 0.70 0.68 0.84 0.78 0.52 0.69 0.80

14. Believe that Adam and Eve existed 0.69 0.62 0.06 0.48 0.67 0.84 0.79 0.84 0.68 0.60 0.81 0.83 0.70

15. Believe in Bible as the word of God 0.72 0.67 0.01 0.50 0.70 0.81 0.77 0.84 0.71 0.58 0.78 0.81 0.67 0.80

The results of the correlational analysis of the four conducts are shown in Table 5. It is apparent from this table that there are two statistical significant relationships between the four dependent variables. There is a positive association between income redistribution and moral judgement, which indicates that as the value of morality increases, the value of preference for redistribution will also increase. However, the correlation coefficient of 0.19 implies that this relationship is very weak. The second significant correlation is positive as well and is between the variables informal care and volunteering. The correlation coefficient of 0.76 points to a strong correlation. Thus, more hours of informal care are associated with more hours of voluntary work. Informal care is defined as a type of voluntary work with a private nature (Kirchmaier et al, 2018; Wilkins-Laflamme, 2018). Therefore, it is not surprising that these two variables are correlated.

Table 5

Correlations: Moral Judgement

1 2 3

1. Moral judgement

2. Preference for redistribution 0.19

3. Hours spent on voluntary work per week -0.00 0.07

4. Hours spent on informal care per week 0.15 0.17 0.76 Note: Spearman’s rho for pairwise correlation with p < 0.05

4.2 Regression analysis: Dimensions Religiosity

Moral judgements

The Tobit regressions are performed separately to test the effect of each dimension on the explained variable. It is apparent from Table 6 that moral judgement is statistically significantly correlated at the 0.05 level with church membership, church attendance, belief in God and belief in theological concepts. The negative coefficient implies that an increase in the strength of belief makes individuals less accepting of unethical economic behaviours. Interestingly, the significant relation with moral judgements holds only up when control set A is used. Whenever the extended control set B is included, no significant effect on moral values is found due to the significant control variables age and marriage. Age is positively and significantly correlated with all dimensions of religiosity regarding moral judgements at the 0.01 level when control set A is used and the 0.05 level when control set B is added. This result implicates that the moral

Preference for income redistribution

In the questionnaire respondents have expressed their support for income redistribution. Table 6 shows the results obtained from the analysis of the answers. It can be seen from the data in the table that redistribution is positively and significantly correlated with two of five religiosity dimensions, being church membership and regular church attendance at the 0.05 level. The significance for both variables remains when control set B is used, although the strength of the relationship weakens for church membership as it becomes significant at the 0.10 level now. Preference for redistribution is measured on an ascending scale, where higher values imply a higher tolerance for income inequality. Based on the positive coefficients of both dimensions, there can be stated that, on average, acceptance of income inequality is higher by 0.503 units for religious individuals, in terms of church attendance and membership, than for non-religious individuals, all other variables held constant. An interesting result to emerge from the data is that control variable income is highly positively and significantly related to redistribution at the 0.001 level. Redistribution is the only dependent variable that is significantly correlated with individuals’ income level. From economic perspective this correlation is normal, given individuals tend to vote for redistribution if their disposable income increases with redistribution (Neher, 2011).

Hours spent on voluntary work

The results from the analysis of volunteering are presented in column (3) of Table 6. The evidence of religiosity affecting economic attitudes becomes stronger when looked at volunteering. There is a positive correlation between volunteering and the five dimensions of religiosity. The modest, but statistically significant, coefficients suggest that an increase in religious beliefs cause an increase in the hours spent on voluntary work. The significance of this variable is consistent across all five dimensions of religiosity when control set A is used. Church attendance and belief in theological concepts are significant at a higher alpha, namely at the 0.10 level. When the list of control variables is expanded, only prayer and belief in God remain significant, at the 0.05 level. Consistency among the dimensions is also visible in the coefficients: all of the coefficients lay around 2,5 hours of voluntary work per week.

Looking at the control variables, an interesting image arises with regard to higher education. Voluntary work is the only dependent variable that is significantly related to the level of education. The coefficient on this variable is negative, which implies that an increase in education level causes a decrease in the time that is spent on voluntary work per week.

Table 6

Religiosity and ethics: adjusted predictions

Note: Column N shows the number of observations for each category. Columns A and B present the adjusted prediction for each measure, including control set A and B respectively. The significance level of each variable shall be reported as follows: # p< 0.10* p < 0.05, ** p < 0.01, *** p < 0.001. The regression types of all four columns are

Dimension of religiosity Moral judgement (0-10)

Preference for redistribution

(0-5)

Voluntary work

hours per week Informal care hours per week

(1) (2) (3) (4)

N A B A B A B A B

Church membership

Not a church member 74 2.482 2.197 1.442 1.368 4.956 4.310 7.555 12.661

Church member 153 2.089 2.153*A 1.836 1.763*A#B 6.787 6.551*A 8.130 8.071

Church attendance

Never 43 1.772 1.825 1.527 1.386 4.499 3.977 2.181 8.327

Less than once a week 159 2.336 2.400*A,B 1.811 1.844*B 6.111 6.198 10.255 13.054*A

More than once a week 61 1.701 1.827 1.641 1.311 6.893 6.014#A 8.799 3.292

Prayer

Never 52 2.217 2.021 1.724 1.523 4.530 4.430 2.646 5.565

Less than once a week 42 2.598 2.553 1.612 1.758 5.685 7.167*B 6.502 6.672

More than once a week 170 1.934 2.101 1.752 1.643 6.723 5.971*A 10.356 10.103*A

Belief in God

Belief in God < median 94 2.441 2.203 1.630 1.611 4.967 3.854 4.504 3.549

Belief in God > median 169 1.907 2.143*A 1.775 1.653 7.101 7.471*A,B 10.278 12.145*A,B

Belief in theological concepts

Belief in concepts < median 124 2.328 2.262 1.683 1.544 5.240 4.941 10.060 10.331

Hours spent on informal care

In the prolongation of this, column (4) shows the results of the weekly hours that are spent on informal care. In contrast to the previous results, the correlation between the dimensions and the dependent variable is weaker with little consistency across the dimensions. Informal care is the only variable that does not show a significant relationship with church membership. That is, there is no real difference between those who belong to a religious community and those who do not. A positive and significant relation is apparent if looked at church attendance, prayer, and belief in God at the 0.05 level. The coefficients of these three dimensions do lay close to each other, all of them around 13 hours. As regards the control sets, a similar pattern arises as for volunteering in general. The three dimensions are significant control set A is used. Once control set B is added, the coefficients, with the sole exception of belief in God, lose their significance. The majority of the control variables in the correlation between church attendance and informal care are significant. Age is positively and significantly related at the 0.01 level, as well as being married and divorced. Having children and self-employment has a negative and significant effect on the hours spent on informal care per week.

4.3 Regression analysis: Muslims and Christians

Table 7 shows the results obtained on the effect of denominations, especially regarding the differences between the substantial subgroups in the sample, Muslims and Christians. The structure of the table is similar to the structure that is described in Table 6. Two types of regressions are conducted for each subgroup. In the first panel, Denomination I, Muslims and Christians are showed against non-religious individuals. Subsequently, Muslims are directly compared to Christians. In the third panel, Muslims are compared to the non-religious.

It can be seen from the results presented in panel Denomination I in Table 7 that the results shown in Table 6 are partly replicated. Even though not all coefficients are statistically significant, the patterns emerge for most denominations. The results of moral judgements are reaffirmed for each subgroup. As for the preference for income redistribution, a similar effect is apparent in Table 7 for Christians. The exact opposite also occurs when we look at time spent on voluntary hours: the same significance level is visible, but this time only for Muslims. When it comes to the time that is spent on informal care, the results are consistent in not showing a significant effect for both religious and non-religious individuals.

Table 7

Denomination and ethics: adjusted predictions

Note: Column N shows the number of observations for each category. Columns A and B present the adjusted prediction for each measure, including control set A and B respectively. The significance level of each variable shall be reported as follows: # p< 0.10* p < 0.05, ** p < 0.01, *** p < 0.001. The regression types of all four columns are Tobit regressions. Controls A: age and male. Controls B: number of children, partner, divorced, married, income, education and self- employment.

Denomination _{Moral judgement}

(0-10)

Preference for redistribution (0-5)

Voluntary work hours per week

Informal care hours per week

(1) (2) (3) (4)

N A B A B A B A B

Denomination I

Not a church member 74 2.458 2.203 1.464 1.379 4.714 3.772 8.407 10.846

Muslim 134 1.964 2.022*A 1.827 1.756 7.026 7.608#A*B 7.585 9.256

Christian 50 1.766 2.622*A 1.979 1.954*A 6.649 2.502 12.028 3.680

Denomination II

Christian 50 1.812 2.681 1.985 1.929 6.777 6.749 9.983 4.457

Muslim 134 2.007 2.007 1.829 1.810 6.886 6.932 ***B 7.352 7.526

Denomination III

Not a church member 74 2.483 2.483 1.482 1.426 5.036 3.880 7.720 10.478

The second panel shows the refined results of the two subgroups. As shown in the table, there are no significant differences evident between Muslims and Christians regarding moral judgements, preferences for redistribution or the time that is weekly spent on informal care. The only exception is the time that is weekly spent on voluntary work. For Muslims, we observe the salient detail of a negative association with voluntary work at the 0.001 level relative to Christians. A one unit increase in the intensity of Muslim beliefs causes a decline in the hours spent on volunteering per week of 0.154. A similar correlation has not presented itself in earlier findings: neither insignificant nor significant.

Panel Denomination III presents the results of the comparison between Muslims and non-members. Being an adherent of Islam is negatively and significantly correlated with Moral judgement, meaning that Muslims are less accepting of unethical behaviours than non-believers. However, this claim only holds when control set A is employed. After adding the extensive controls list, the coefficient weakens and becomes insignificant. In line with our previous findings, no significant correlation is detected between religiosity and preference for redistribution. Also in line with our previous findings is the correlation between religiosity and voluntary work. As can be seen from the table, there is a significant correlation at the 0.10 level and 0.05 level when control set A and control set B are employed, respectively. The results of the correlational analysis regarding informal care show a different pattern for the denominations compared to religiosity and its dimensions. None of the subgroups is significantly correlated with informal care, while this was the case for the dimensions of religiosity.

Thus, the religion that you adhere to does not affect moral values, the attitude towards income redistribution and the amount of time spent on informal care. In addition, it appears that individuals’ judgement is dependent on more variables. Preferences for redistribution is in all cases and for all denominations significantly related to the level of income. An increase in income leads to a decrease in support for income redistribution. The marital status of religious individuals also affects their behaviour: married Christians hold to stricter moral values and volunteer significantly less. This has also to do with the fact that having children is negatively and significantly correlated with voluntary work. For Muslims, the same negative relationship is observable. Voluntary work is also found to be dependent on the educational level of individuals. The higher the level of education, the fewer voluntary work is done. This is noticeable through all subgroups and through all dimensions of religiosity.