Dimensions of religiosity and attitude towards deviant behaviour

Dimensions of religiosity and attitude towards deviant behaviour A cross-national study in Europe Mina Ehahoui (538942) June the 27 th, 2012 Master thesis Sociology First corrector: Mrs. dr. J.A. Moor Second corrector: Mrs. dr. C.M.C. Verbakel

Index Abstract... 4 Introduction... 5 Social relevance... 6 Scientific relevance... 7 Religious dimensions... 7 New research... 8 Cross-national... 8 Research questions... 9 Tittle & Welch s theoretical framework... 11 First line of thought... 11 Second line of thought... 11 Third line of thought... 12 Fourth line of thought... 12 Fifth line of thought... 12 Overview... 13 Hypotheses... 13 Public Practice... 14 Private Practice... 15 Belief... 15 Respect for authority... 16 Country s religious climate... 16 Data... 17 Operationalization... 18 Dependent variable... 18 Religious dimensions... 18 2

Respect for authority... 19 Control variables - Individual... 19 Country characteristics... 21 Missing values... 21 Results... 22 Method... 22 Bivariate analyses... 23 Public practice... 23 Private practice... 24 Belief... 24 Religious climate... 25 Respect for authority... 25 Multivariate analyses... 26 Multicollinearity... 26 Individual and country level effects... 27 Cross-Level Interaction Effects... 31 Conclusion and Discussion... 33 Conclusion... 33 Discussion... 36 Appendix... 37 References... 43 3

Abstract The aim of this study is to answer the question to what extent religion on both individual and contextual level have an influence on attitude towards deviant behaviour. Based on the theoretical framework by Tittle and Welch (1983), it was expected that the attitude towards deviant behaviour would be more condemning for church members, people with strong beliefs, people who pray and those who often go to church. Respect for authority would provide for an explanation for these outcomes. It was also expected that attitude towards deviant behaviour would be more permissive in the secular countries. Some expectations find support and other do not. Church members for example, do have a more condemning attitude towards deviant behaviour than non church members. This result however differs for different denominations. Permissiveness towards deviant behaviour is not influenced by church attendance at all. Taken into account that the influence is not always in the expected direction, all the religious dimensions, except for church attendance, have an influence on the attitude towards deviant behaviour. 4

Introduction Deviant behaviour is as the word indicates behaviour that is uncommon and indecent. It is of importance to note that deviant behaviour is not only criminal behaviour. An act is considered to be criminal if an infringement is committed upon a law. Next to this, acts against the social standards and norms are also considered to be deviant. Such behaviour can be seen as a source of instability, disorder and a lack of cohesion. For these and other reasons, deviant behaviour is undesirable. However, not everyone shares this stand of view. In this research both deviant behaviour that is seen as a crime and deviant behaviour that is not seen as a crime, are taken into account when looking at attitude towards deviant behaviour. Different deviant behaviours are judged differently by the state laws of the European countries. Claiming state benefits, cheating on tax, joyriding and accepting a bribe are deviant behaviours that are prohibited by the law in all the European countries. However, taking soft drugs is a deviant behaviour that is not prohibited in the Netherlands. In most European countries, adultery is not illegal. Nevertheless, in some European countries (Albania for example), committing adultery might lead to getting a penalty. There is a difference in how certain deviant acts are judged by the state laws in the European countries. It might therefore also be the case that the inhabitants of the European countries differ in their judging of deviant behaviour and that even within a country there might be variety in attitude towards deviant behaviour. When exploring the explanations for deviance and attitude towards this, religion and religiosity seem to appear as an important explanation due to the condemning attitude towards deviant behaviour by most religions. Because scripture such as the Bible, the Torah and the Quran are very concrete and clear about deviance and the punishment that follows deviant behaviour, the assumption is easily made that religion leads to a condemning attitude towards deviant behaviour. Even though this is sometimes seen as a common thought, it is hazardous to make such a statement. Therefore almost a century ago, scientists debated this assumption and started to test whether religion had an inhibitory effect on (the acceptance of) deviant behaviour (Lombroso, 1911; Kvaraceus, 1944). Since then, a lot of research had been conducted in order to continue testing this assumption (Putney & Middleton, 1962; Londen et al., 1964; Wright & Cox, 1967; Tittle & Welch, 1983; Bierbrauer, 1992, Francis, 1992, 1997; Stylianou, 2004; Francis et al., 2005). If religion does have an inhibitory effect on deviant behaviour and attitude towards it, it is important to see to what extent this is. 5

Social relevance It is important to look at attitudes towards deviant behaviour because the majority of people abide to the norms that are set by society. They do not behave in a deviant manner, and for that reason there is hardly any variation within the population. This is in contrast to attitudes towards deviant behaviour. Even though one might act in accordance with the norms, this person s attitude might differ from this. This means that more variation in attitude towards deviant behaviour might be present. If this is the case, and if there are differences found in attitude towards deviant behaviour within the population due to religion, it is important to test to what extent this would be. When the differences in attitudes are large, one might consider a problem in the level of cohesion. If whole groups differ in their attitude towards deviant behaviour, society as a whole can be affected. It might be that people with strong differences in attitude will collide. Attitudes towards deviant behaviour might indirectly affect a country s political climate. For example, if people in a country are very condemning in their attitude towards deviant behaviour, the support for parties who want severe punishments for deviant behaviour will be stronger. It is relevant to see whether even in modern times religion, not only on an individual level but also on contextual level, has an influence on attitude towards deviant behaviour. Religion is often seen as a tool for awareness of different norms. Is this really the case or is the role religion plays in this awareness very small? Norm awareness might also be influenced by other characteristics and not by religion only. With an eye on secularisation, i.e. the disappearance of the role of religion in society and its institutions, people fear for a decline in morality. This means that when the religious climate in a country turns out to be an important determinant for attitude towards deviant behaviour, it is an indication that morality loses grounds within societies. The assumption here is that in religious societies, a strong emphasis is put on morality while this emphasis is less strong in secular societies where religion hardly plays a role in everyday life. Thus, a decline in religion is often linked to a decline in morality. This is mainly stated in politics where religious political parties put an emphasis on the danger of the absence of religion in both lives of individuals and society as a whole. These parties consider holding on to the norms and values and the task for realising this as an important element that holds society together. The question however is whether this fear for a decline in morality due to religion (both individual and contextual) can be justified or whether it is unfounded. 6

Scientific relevance Religious dimensions Over the years, several researchers made an attempt to examine the relationship between religion and deviant behaviour. (Baier and Wright, 2001). However, less research is done about the relationship between different religious dimensions and attitude towards deviant behaviour. In these studies, when examining the effects of different aspects of religion, there was no consideration for the meaning behind and differences between these religious aspects and therefore no clear distinction was made. Personal religiosity, e.g. belief in god, seems to be an important predictor for attitude towards deviant behaviour while practice of religion is not (Putney & Middleton, 1962; Londen et al., 1964, Francis et al., 2005). In all studies a relationship is found between at least one different aspect of religiosity and attitude towards (a particular or a set of) deviant behaviour. Therefore, in continuance of these findings, it is important to distinct dimensions of religiosity when looking at attitude towards deviant behaviour. In this study, a distinction will be made between religious dimensions. The religious aspects that are used in the previous literature are basically dimensions practice and belief, which are distinguished by Glock and Stark in several of their studies (Glock 1962; Glock and Stark 1965, 1966, 1968). By belief, Glock and Stark mean that a person holds to a theological outlook. An example is belief in the traditional aspects of the monotheistic religions e.g. in god, heaven and hell (also called dogmatism). The second dimension by Glock and Stark that is to be used is practice. This dimension is about the acts performed by religious people in order to fulfil their religious commitment. This dimension can be distinguished into public practice and private practice. Reitsma (2006) explains that the distinction made by Glock and Stark between public practice and private practice, refers to church membership and church attendance for public practice and prayer or meditation for private practice. An important annotation must be made concerning church membership and church attendance. Both are seen as public practice but they eventually differ from each other. The same goes if one speaks about these dimensions together. Even though the discussed dimensions are alike (in the meaning that they all say something about the religiousness of a person), it is important to make a distinction because each dimension might have a different influence on attitude towards deviant behaviour. 7

New research Most of the previous studies about religion and attitude towards deviant behaviour are old and not recent (Putney & Middleton, 1962; Londen et al., 1964; Wright & Cox, 1967; Tittle & Welch, 1983; Bierbrauer, 1992). New research is needed because the analyses that are used in these studies are simplistic (mainly correlations). Nowadays, more complicated and extensive statistical methods exist, e.g. multilevel techniques, in order to test effects on both individual and contextual level. Furthermore, the use of these simplistic analyses (such as correlations) made it difficult or impossible to include important characteristics as control variables. Their results are inaccurate and therefore less reliable. Since older studies are also outdated (attitudes change over time due to for example secularisation), new research is needed where recent data must be used. However, there are also more recent studies that examine the relationship between religion or religiosity and attitude towards deviant behaviour. In spite of this, these recent studies are more often based attitude towards (different kind of) substance use, instead of a set of diverse behaviours (Francis, 1992, 1997; Stylianou, 2004; Francis et al., 2005). Even though these studies are more recent, the focus on only one this one particular deviant act, hardly says anything about attitude towards deviant behaviour in general. Cross-national Hardly any research is done on a cross-national level where important country characteristics like secularisation (or rather the religious climate in a country) are taken into account. Even though Tittle & Welch (1983) did take religious climate into account, it was not crossnational. The study was in the United States of America and they looked at the religious climate of the states that were used. For this study, a closer look will be taken at the variation in attitudes towards deviant behaviour in the European countries. Since there is variation in attitude towards deviant behaviour among Europeans and the European countries, it is essential to seek for explanations for the differences in these attitudes. A high number of European countries will be included (44 countries) which means a great source of information. In figure 1 it is shown that countries with strong permissiveness towards deviant behaviour, such as Sweden, Iceland, Denmark France and Finland are secular countries. On the other hand, countries where attitude towards deviant behaviour is more condemning are the religious countries such as Malta, Northern-Ireland, Ireland, Kosovo, Turkey and Greece (Appendix A). This gives an indication that religious climate plays a role in forming an attitude towards deviant behaviour. The reason why it is important to take religious climate 8

into account is due to its influence on the values of individuals. One who lives in a religious country, but is not religious at all, might be affected by the religious values that are common and shared in this country. In this case this person s own religion is not the determinant for the values it holds, but the religious climate is. Therefore, next to the individual characteristics, the country characteristic religious climate will be included to see whether this characteristic influence attitude towards deviant behaviour. Research questions Based on previous literature, social relevance and scientific relevance the main research question is to what extent does religion on both individual and contextual level have an influence on attitude towards deviant behaviour. Because of the importance of making a distinction between religious dimensions, the sub question is To what extent do different dimensions of religiosity have an influence on attitude towards deviant behaviour? The two main dimensions to be used are belief and practice, which on its turn can be divided into public and private practice. Finally, it is interesting to see whether the differences between the European countries are due to the religious climate of a country. The next sub question therefore is To what extent does religious climate of a country provide for an explanation for the difference in permissiveness between the European countries? 9

Source: EVS 2008, N=44 10

Tittle & Welch s theoretical framework Tittle & Welch (1983) discussed that up until then, studies regarding religion and attitude towards deviant behaviour lacked in providing a substantive theoretical framework. They refer to this as a knowledge gap and consider it to be a problem in this field. Based on five lines of thought, Tittle & Welch mapped out a theoretical concept of why religiosity would have an inhibitory effect on attitude toward deviance. These same lines of thought will be discussed and used in order to derive hypotheses about religious dimensions and religious climate of a country. First line of thought The first line of thought is based on functionalism. Tittle & Welch use Parson s view that social order consists when values are collectively held. Because these values are assumed to be based on moral commitment, it is expected that people will not behave deviant. This means that the main idea is that people, due do their shared values, will not behave in a deviant manner nor will their attitude towards deviant behaviour be permissive. Now, religion is seen as a tool that enables people to link sanction set by the divine to moral rules, which are the same as the norms shared in society. This means that because of religion, people are more easily obedient to the shared norms since they are used to obey the divine. In addition, religious people who are actively involved in religious activities and institutions are presumed to have a stronger emphasis on norm obedience and moral commitment because they are repeatedly exposed to information about sanctions by the divine and moral rules. According to this thought, active religious participation such as church attendance is to have an inhibitory effect on both deviance and attitude towards deviant behaviour. Second line of thought This line of thought that is discussed by Tittle & Welch is based on the social control theory in general and Hirshi s social bonding theory in particular. According to Hirschi s social bonding theory, someone who has a weak bond with their own community and to conventional others and who does not have social attachments, is more likely to go against the norms and therefore accept deviant behaviour (Hirschi, 1969). Those who have a strong bond with their own community and conventional others will think the other way around. This is because individuals in the social network; family, friends, neighbourhood members and other 11

people with whom there is a tie, cam affect someone s life in different ways. According to Hirschi, it is important that people are attached, committed and involved to their community and conventional others in order to act and think according to the norms shared in society as a whole and norms shared by the community in particular. This theory is an application of Durkheim s socialisation theory (Durkheim, 1897) where the extent to which people are integrated within a group influences the norm obedience of this group. Tittle & Welch explain that since the majority of the religious denominations are conventional and participation in religious institutions or activities lead to having ties with conventional others, people who are member of a church and who are religiously participating will be more likely to support the norms shared in society, hence they will be less permissive towards deviant behaviour. Third line of thought The differential association perspective by Sutherland & Cressey (1966) is the third line of thought that is used. According to this theory values are learned by others, which means that interaction is a key element. The environment of a person is crucial in this case. Tittle & Welch mention that almost all religious denominations promote norm obedience and that being surrounded by religious people leads to a condemning attitude towards deviance. An example is when someone lives in a country with a strong religious climate; the interaction of this person with religious people is hardly inevitable. Hence, this person s values will be learned through these interactions. Despite of their own religious background, they will eventually share the same values as religious people. Fourth line of thought The next line of thought is based on several, mainly social psychological, perspectives. These perspectives basically explain that belief, i.e. belief in traditional aspects of religion, and worship of god, will lead to respect for authority. This will be impersonated to respect and conformity to rules, whether divine or legitimate. The assumption is that one who is fostered with traditional belief and worship will have developed a strong value about respect for authority and conformity to rules. Deviant behaviour will therefore be more likely to be convicted by the people with strong religious beliefs and people who worship god. Fifth line of thought Finally, the deterrent doctrine by Gibbs (1975) which has a utilitarian point of view is used. According to Gibbs, norm obedience is based on the costs that will be made and the potential 12

reward that can be received. If the reward for norm violation is higher than the costs (sanction), people will behave deviant and their attitude towards deviant behaviour will become more permissive. Now, according to Tittle & Welch, being a part of a religious community, which is being a member of this group, will increase the costs of norm violence because the probability of being sanctioned by a group member (due to for example control) are higher. In addition, a person who is a member of a religious group and violates the norms, by behaving in a deviant manner, loses aspects such as respect and support from the other members. Attitude towards deviant behaviour will be negative because it is an unattractive behaviour both in morality and costs. Overview The discussed lines of thought are all aiming in same direction. That is religious aspects lead to a condemning attitude towards deviant behaviour. Except for the fourth line of thought, where psychological perspectives are used as a mechanism, the mechanisms of all other lines are to be sought in the influence of the environment of people. Whether it is due to integration, learning process, exposure to religious information or expected costs. All explain that when you are a member of a church or when you are active within this church or religious organisation, norm obedience, directly or indirectly, becomes more important. The theoretical framework will therefore function as a tool to explain why different religious dimensions would have an inhibitory effect on attitude towards deviant behaviour. Since the underlying mechanisms are the same, the different lines will not be tested individually in order to seek support for the different theories. Hypotheses The theoretical framework by Title and Welch cover the religious dimensions by Glock and Stark (Glock 1962; Glock and Stark 1965, 1966, 1968) that are discussed in the introduction. The influence of the dimension belief and private practice on attitude towards deviant behaviour can be explained by the fourth line of thought. For the dimension public practice, hypothesis will be derived from the first, second and fifth line of thought. The second line of thought can be used to explain the influence of the religious climate in a country on attitude 13

towards deviant behaviour. Different hypotheses will therefore be derived from different lines of thought mentioned in the theoretical framework. Public Practice Both Hirshi s social bonding theory (1969), i.e. second line of thought, and Gibss deterrent doctrine (1975), i.e. fifth line of thought, explain that church membership might lead to a less permissive attitude towards deviant behaviour. This is because having ties with conventional others leads to stronger norm obedience. The assumption here is that due to membership, a tie consists between the member and conventional others. In addition, a person gains certain benefits just by being a member of a church. Committing a violation of the norms while you are a member of a church, knows certain costs which makes deviant behaviour an unappealing matter. The benefits you receive as a church member can suddenly be taken away when your values are not the same as the values of the church you are a member of. This means that it can be expected that H1. Attitude towards deviant behaviour will be less permissive for people who are a member of a church than for people who are not a member of a church. For church attendance, the same lines of thought can be used to explain why people who frequently go to church will have a more condemning attitude. In addition, support can also be found within the first line of thought. First Hirshi s bonding theory; because of religious participation and therefore ties to conventional others religious people have a stronger bond with the so called larger community. They have a stronger bond with their community because they are to be more attached, involved and committed. Religious practices can be seen as participation. Examples are church attendance and other involvement in church activities such as voluntary services. Also Gibbs deterrent doctrine can be applied. Committing a violation of the norms as a frequent church attendee, knows certain costs which makes deviant behaviour an unappealing matter. These costs may even be higher than for people who are only a member of a church and do not attend or whose church attendance is very low. This is because there is more control and therefore a higher chance of losing benefits when not abiding to the norms. According to Parsons idea, active religious participation such as church attendance is to have an inhibitory effect on both deviance and attitude towards deviant behaviour because it enables people to link sanctions by the divine to moral commitment. They are more often exposed to information about sanctioning than those who are not actively participating in the religious community. This leads to the hypothesis H2a. Attitude towards deviant behaviour will become less permissive when frequency in 14

church attendance increases. It is also expected that H2b. this effect will be stronger than the effect of church membership on attitude towards deviant behaviour. Even though both church membership and church attendance are seen as one dimension, that is public practice (Reitsma, 2006), differences based on attachment to their religious affiliation might be considered. A member whose church attendance is frequently is not the same as a member who never goes to church. The bond with the religious affiliation is probably stronger for someone who is often surrounded by religion and religious people which might affect norm obedience. Private Practice For private practice, a hypothesis can be derived based on the psychological perspective that worship of god, i.e. devotion, fosters respect for authority. Because it fosters respect for authority, people who worship god through meditation or prayer will have a more condemning attitude towards deviant behaviour. The assumption here is that people who take time to pray have a strong bond with god and have therefore a stronger conviction that obedience towards god s rules is an important element in life. Since they are used to obey god s rules, obedience in general is considered to be the standard. It is therefore expected that H3. Attitude towards deviant behaviour will be more condemning for people who pray compared to people who do not pray. Belief Besides religious practice, there is a form of religiosity, i.e. belief, which is based on how someone experiences religion. This has nothing to do with how people act but rather how people think and feel towards the contents and purposes of religion in that way that they adapt to and belief in theological concepts of their religious affiliation. This does not mean that religious practice is not important for these people. In contrary, it might even be that because of religious practice, belief is affected in a matter that people feel more connected to their religious affiliation and therefore believe in the theological aspects of their affiliation. Examples of personal religiosity are belief in the traditional dogmas, e.g. god, heaven, hell and life in hereafter and the importance to have a personal god in one s life. According to the social psychological perspective, also belief fosters respect for authority and norm obedience. If someone has a strong personal belief, it is likely that respect for authority is very strong. Hence it is expected that H4. attitude towards deviant behaviour will be less permissive when strength of belief increases. 15

Respect for authority The social psychological perspective expects that belief and private practice would have an inhibitory effect on attitude towards deviant behaviour because of a heightened respect for authority. It might be that the impact of the dimensions belief and private practice can be explained by respect for authority. Furthermore, this might also be the case for both aspects of public practice since religion in general might be a tool that creates more emphasis for authority. This means that it is presumed that people who are a member of the church, people whose belief is strong and people who practice religion, will have more respect for authority. It can therefore be expected that H5. The relationship between religious dimensions and attitude towards deviant behaviour can be explained by respect for authority. Country s religious climate The third line of thought can be used to explain the differences in attitude towards deviant behaviour for the different European countries. Since values are learned by others and interaction with these other people forms a key element, the environment influences the values. As explained, if someone lives in a country with a strong religious climate there is more interaction of this person with religious people then in case it concerned a secular country. The values will be learned through these interactions and despite of own religious background, non religious people will eventually share the same values as religious people. It is expected that H6. People who live in the more religious European countries will have a more negative attitude towards deviant behaviour than people who live in more secularized countries. It might also be that religious climate of a country might interact with the relationship between individual religiosity and attitude towards deviant behaviour. The attachments to society (based on individual religiosity) are less important because of the effect of society s religious climate. In a more religious country, both religious and non religious people are being influenced by the values of others. Even though someone is non religious and lives in a religious country, that person s attitude towards deviant behaviour will not be that different from religious people because the values of his or her fellow citizens, influences this persons attitude. Therefore, H7. the relationship individual religiosity and attitude towards deviant behaviour will be less strong in the more religious countries. 16

Figure 2. Conceptual model Data The data that are used to test the hypotheses are from the European Value Study (EVS), 2008. The EVS is a large scaled, cross-national and longitudinal research program about values in Europe. Every nine years (1981, 1990, 1999 and 2008), standardized questionnaires are used in face to face interviews that were conducted in different European countries. For this study, only data is used from the year 2008. The integrated dataset contains data of 47 countries (Albania, Armenia, Austria, Azerbaijan, Bosnia Herzegovina, Belgium, Bulgaria, Belarus, Switzerland, Cyprus, Czech Republic, Croatia, Germany, Denmark, Estonia, Spain, Finland, France, Great Britain, Georgia, Greece, Hungary, Ireland, Iceland, Italy, Kosovo, Lithuania, Luxembourg, Latvia, Moldova, Montenegro, Malta, Northern Cyprus, Northern Ireland, Netherlands, Norway, Macedonia, Poland, Portugal, Romania, Serbia, Russian Federation, Sweden, Slovenia, Slovak Republic, Turkey and Ukraine). Three countries, Croatia, Macedonia and Azerbaijan had to be excluded from the analyses. Azerbaijan due to methodological reasons (data is not representative for the country) and Croatia and Macedonia are excluded because the country characteristic GDP was not available for these countries. A representative random sampling is used which lead to representative respondents of the people within these European countries. The N on the country level is therefore 44. 17

The numbers for Gross Domestic Product (GDP) are purchased from a different data source then the EVS. This external data source that is used for GDP is Eurostat (2008) which is the statistical office of the European Union. Their data contains different structural indicators of countries in the European Union, the euro-zone, and the Member States and their partners. For countries with missing information, numbers for GDP (same scale) were taken from the CIA World Factbook. Operationalization Dependent variable The dependent variable, attitude towards deviant behaviour, is based on a 7-item scale. The main question is do you justify and then per item followed by the deviant behaviour - claiming state benefits, cheating on tax, joyriding, taking soft drugs, lying in your own interest, adultery and accepting a bribe. Respondents were able to choose on a scale between 1 never and 10 always. After a factor analysis (Oblimin rotation) that showed that this scale consisted of one component, a reliability test was performed. The reliability when all countries are included is good; there was a cronbach s alpha of 0.792. The reliability when performing the test for each country separate was also acceptable. Religious dimensions Four variables that measure religious dimensions will be used. Public practice is measured with two variables; church membership and church attendance. First, for church membership the question do you belong to a religious denomination is used. This is because claiming to belong to a religious denomination indicates that this person is a member of the church or a religious denomination. The respondents could answer either yes or no. Those who answered yes, which makes them a church member, are used as the reference category. Second, church attendance will be measured by the frequency in church attendance per year. The question that is used is how often do you attend religious services, answers are: 1 more than once a week, 2 once a week, 3 once a month, 4 only on specific holy days, 5 once a year, 6 less often, 7 never, practically never. These categories are changed into frequency of church attendance per year by assigning the number of how many times per year a person visits the church. Because categories more than once a week and less often than once a year do not state the frequency per year; for more than once a week, the frequency is 104 visits per year and for 18

less often than once a year, the frequency is 0.5 visits per year. Private practice will be measured by the variable prayer: do you take moments of prayer or meditation?, answer 1 yes or no. The respondents that answered this question with a confirmation are used as a reference category. For the dimension belief a scale of five items will be used that consists of questions about belief in the traditional dogmas. The questions are do you believe in God, do you believe in life after death, do you believe in hell, do you believe in heaven and do you believe in sin. It is counted how many times the respondents answered with yes. Also for this scale a factor analysis (Oblimin rotation) was performed to test whether the different items can be used to create one scale. This happened to be the case and a reliability test was performed when all countries included. The result was very good; there was a Cronbach s alpha of 0.835. Also the reliability of this scale for each different country was high. The scale then, runs from 0 does not believe at all to 5 believes in all traditional aspects. Respect for authority Respect for authority will be included as an explanatory variable. Respondents were asked whether more respect for authority is a good or a bad thing. The answers were good, bad and don t know. These three categories remain, and a fourth category with the missing values is added in order to take them in the analyses (there are a lot of missing values for this variable). Respondents who answered this question with good will be used as the reference category. This variable (respect for authority) is included because it is expected that this will explain (a part of) the relationship between religious dimensions and attitude towards deviant behaviour. Control variables - Individual Control variables are included because they are seen as elements that can influence both religious dimensions and attitude towards deviant behaviour. First, I would like to control for socio-economic status (SES) based on income and education. SES will be controlled for because it is assumed that people with a low SES might understand why one would commit deviant behaviour (Antonaccio & Tittle, 2008). For example, someone who has a low income understands that someone would steal or lie because he or she is trying to survive in a poor situation. This means that in case that people with a low SES are the same as the people who consider themselves as religious, the effect of religion on attitude towards deviant behaviour becomes positive. SES is therefore a variable that might disturb the relationship between religion and attitude towards deviant behaviour. The variable for income is used where the monthly household income (x1000), corrected for Purchasing Power Parity (PPP) in Euros is 19

available. Because there are a lot of missing values for this variable, and the choice is made not to delete them, the mean income is aggregated and assigned to these missing values. Thus, each country has a mean on the variable income. Each country will have a different mean that will be assigned to the missing values for the matching country. In addition, a dummy variable is made where the missing values will be the reference category. This variable will be included when analyses are performed. Second, education is a variable where the educational level of a respondent is corrected in order to compare education between different countries. The categories are 0 'Pre- primary education or none education, 1 Primary education or first stage of basic education, 2 Lower secondary or second stage of basic education, 3 (Upper) secondary education, 4 Post- secondary non- tertiary education, 5 First stage of tertiary education, and finally 6 Second stage of tertiary education. These categories are sorted into the three categories: low educated consists of pre-primary education or none education, middle educated of primary education or first stage of basic education, lower secondary or second stage of basic education, (upper) secondary education and post-secondary non-tertiary education and finally high educated of first stage of tertiary education and second stage of tertiary education. In the analyses, high educated is taken as a reference category The other two control variables are gender and age (for an interpretative reason, when included in multilevel analyses, age is centred at the minimum). In previous research about attitude towards deviant behaviour (Putney & Middleton, 1962), a difference between men and woman was found. Women were more likely to condemn deviant behaviour then men. If women, in contrast to men, happen to see themselves more often as religious (Miller & Stark, 2002), the effect of religion on deviant behaviour might be spurious. This might also be the case for age. It is assumed that older people compared to younger people are less positive towards deviant behaviour (Tittle, Ward & Grasmick, 2003). They are also more likely to be religious compared to young people (Markides, 1983; Ainlay & Randall Smith, 1984). Since age affects both religion and attitude towards deviant behaviour, the effect of religion on attitude towards deviant behaviour might disappear or become less strong. Finally, denomination is included as a control variable. This is in order to make sure that the variation in attitude towards deviant behaviour explained by dimension is due to the differences in denomination instead of religious dimensions (Londen et al, 1964). The denominations that are used are Catholic, Protestant, Orthodox, Muslim, other religion and no denomination. Those who do not belong to a religious denomination are included as the reference category. Since the original variable that is used to control for denomination is the 20

same as the variable to distinguish members from non members, the variables will not be included at same time. When controlling for denomination, the dichotomous variable membership will be excluded. The reference category is respondents with no denomination. The religious denomination (member) will therefore be compared to the group no religious denomination (non member). Country characteristics The religious climate in a country was measured by the mean of the frequency of church attendance per year (for an interpretative reason, when included in multilevel analyses, religious climate is centred at the mean). The countries with the strongest religious climate are Malta and Poland and the countries with the lowest religious climate (the most secular country) are Sweden, Denmark, Finland and Estonia. In addition to the individual control variables, I also want to control for welfare on a country level. Basically the same mechanism as religious climate can be applied. A person who lives in a secure environment, that is for example in countries with a high GDP, is more likely to share the same values such as condemning deviant behaviour (Antonaccio & Tittle, 2008) as their surroundings where the majority of people are not poor. GDP might influence both individual religiosity and attitude towards deviant behaviour. There is however no such variable in the EVS. Therefore, the numbers for Gross Domestic Product (GDP/10.000) are to be found in a different data source, Eurostat (2008). This is the statistical office of the European Union. Their data contains different structural indicators of countries in the European Union, the euro-zone, and the Member States and their partners. GDP for Croatia and Macedonia was not available. Missing values Apart from income and respect for authority, the remaining variables did not have a large number of missing values. Therefore a listwise deletion is used in order to take them out of the analyses. The total N before this deletion was 63256 and after deletion 58738 (thus 92.86% was valid). In table 1, the descriptive statistics of the whole sample are presented. The descriptive statistics for each country are to be found in the appendix. 21

Table 1. Descriptive Statistics N min max mean % std, dev, Attitude deviance 58738 1 10 4.915 2.279 Church member 58738 0 1 75.3 Catholic 58738 0 1 29 Protestant 58738 0 1 12.3 Orthodox 58738 0 1 22.3 Muslim 58738 0 1 9.3 Other religion 58738 0 1 2.5 Belief 58738 0 5 2.611 1.905 Church attendance 58738 0 104 15.643 26.188 Prayer 58738 0 1 65.5 Respect authority good 58738 0 1 61 Respect authority bad 58738 0 1 13.8 Respect authority no opinion 58738 0 1 25.2 Low educated 58738 0 1 13.1 Middle educated 58738 0 1 63.4 High educated 58738 0 1 23.5 Age 58738 15 108 47 17.739 Male 58738 0 1 44.9 Income 58738 0 15 1.320 1.232 Country - religious climate 58738 4.21 59.9 15.477 10.918 Country - GDP 58738 0.23 8 2.415 1.534 Source: EVS 2008, Eurostat 2008, CIA World Factbook 2008 Results Method First, for the bivariate analyses, f-tests will be carried out and presented. For the multivariate analyses, a multi-level multiple regression analyses will be performed in which both individual and country characteristics are included. Before that, a test for multicollinearity is executed to see whether this causes a problem when doing the multivariate analyses. When performing the multi-level multiple regression analysis, a null (empty) model will be created first in order to determine the variance on the individual and country level in permissiveness towards deviant behaviour. In model 1 only dimensions of individual religiosity are included and in model 2 the individual control variables are added. This is done so that changes in the effects of religious dimensions on attitude towards deviant behaviour due to the control variables can be noticed. In model 3, respect for authority is included in order to see to what extent this variable explains the relationship between religious dimensions and attitude towards deviant behaviour. Finally in model 4, the country characteristic religious climate and 22

the country control variable GDP are added. Four cross level interactions will be tested; the difference in the effects of individual religiosity (membership, belief, church attendance and prayer) on attitude towards deviant behaviour when religious climate changes. Apart from church membership, the model to which the interaction variables are included is similar to model 4. As explained before, two different variables are used for church membership. In the first four models, church membership will be included as a nominal variable. That is, people who are a member of a certain religious denomination will be compared to people who are not a member of a denomination (reference category). In order to control for different religious denominations, a distinction will be made between these different religious groups. In contrast to this, for the models with the cross level interactions, church membership will be included as a dichotomous variable. Because of interpretative reasons, no distinction will be made between the religious denominations in the cross level interaction models. Bivariate analyses Public practice The mean in attitude towards deviant behaviour for church membership is higher for non members than for members (figure 3). The difference between these two groups is significant. This is what was expected; attitude towards deviant behaviour is more permissive for those who are not a member of the church. However, since it is only a bivariate analysis and no controls are included, it is too early to state that this is actually the case. N=58738, Sig.=0.000, F=37.279 23

An f-test was performed to see whether there is an increase in permissiveness towards deviant behaviour when the frequency in church attendance increases. The f-test showed an F of 44.954 with a Sig. of 0.000. This means that there is a negative relationship between frequency in church attendance and attitude towards deviant behaviour. Private practice N=58738, Sig.=0.770, F=0.086 There does not seem to be a significant difference between respondents who pray and respondents who do not pray (figure 4). The mean in attitude towards deviant behaviour are for both groups basically the same. This gives the expectation that private practice is not important when explaining attitude towards deviant behaviour. The multivariate analyses will have to show whether this expectation holds. Belief In figure 5, the differences in attitude towards deviant behaviour for different levels of belief are presented. The overall picture is when strength of belief increases, permissiveness towards deviant behaviour decreases. However, up until 3, the decrease is very little and probably not even significant. Only after this, a strong decrease is visible. The change in permissiveness towards deviance when traditional belief increases is significant. 24

N=58738, Sig.=0.000, F=48.399 Religious climate An f-test was performed to see whether there is an increase in permissiveness towards deviant behaviour when religious climate of a country becomes more secular. The f-test showed an F of 138.195 with a Sig. of 0.000. This means that there is a relationship between religious climate and attitude towards deviant behaviour. As noted before, the countries with the most permissive attitude towards deviant behaviour are often the secular countries and the countries with the most condemning attitude towards deviant behaviour are the more religious countries. This however, does not necessarily mean that when religious climate in a country becomes more secular, people s attitude towards deviant behaviour will be more permissive. The question whether this is the case will be answered after performing the multivariate analyses. Respect for authority The mean for attitude towards deviant behaviour is highest for those who see having more respect for authority as a bad thing and lowest for those who see it as a good thing (figure 6). This is not suprising since people who foster respect for authority are more likely to conform to the rules set by these authorities. 25

N=58738, Sig.=0.000, F=117.415 Multivariate analyses Multicollinearity Table 2. Test for Multicollinearity Tolerance VIF Belief 0.553 1.807 Catholic 0.465 2.152 Protestant 0.670 1.491 Orthodox 0.501 1.997 Muslim 0.556 1.799 Other religion 0.866 1.154 Church attendance 0.689 1.452 Prayer 0.676 1.480 Low educated 0.593 1.686 Middle educated 0.693 1.443 Age 0.838 1.193 Male 0.961 1.040 Income 0.651 1.535 Respect authority - bad 0.906 1.104 Respect authority - no opin. 0.898 1.113 Country - religious climate 0.647 1.545 Country - GDP 0.549 1.820 Condition number 13.426 Source: EVS 2008, Eurostat 2008, CIA World Factbook 2008 Dependent: attitude deviant behaviour 26

When multiple variables are overlapping and perhaps measuring a part of the same concept, it becomes unclear what part of an effect is caused by what variable. This might lead to a disturbance in the effects. For that reason, a test for multicollinearity is performed to see wether this causes a problem in the analyses. In table 2 the VIF, tolerance and the condition number are mentioned. Multicollinearity does not seem to be a problem. The tolerance does not come below 0.10 and the VIF does not go above 10. Also the condition number of 13.426 does not imply serious problems concerning multicollinearity. Individual and country level effects As notified before, only religious dimensions are included in model 1 (table 3). When not controlling for denomination, churh members have a more condeming attitude towards deviant behaviour than non church members. When controlled for, there are differences between the denomination when comparing members to non members. People who are a church member and belong to the Catholic or Protestant denomination are less permissive than non church members. The effects for both denominations are equally in strength. For the denomination Islam and other religions, being a church member means having a more permissive attitude towards deviant behaviour than non church members. The effects are not very strong considering that the effect is strongest for the Muslim denomination (0.428). The effect of the group other religion is very smaller (0.140) and only significant at a p-value of 0.05. Orthodox Christians do not differ in permissiveness from non church members in this model. Church members do differ in their attitude from non church members; however whether they are permissive or condemning in their attitude depends on the denominations to which the church member belongs to. The reason why the overall picture (when not controlling for denomination) shows that church members are more condemning in their attitude towards deviant behaviour than non church members is due to the fact that main denominations that are present in Europe are the Catholic, Orthodox and Protestant denominations (table 1). The denominations that show a more condemning attitude towards deviant behaviour than the other denominations are the Catholic and Protestant denominations. When strength of belief increases, attitude towards deviant behaviour becomes more condemning. This is also the case for church attendance; attitude towards deviant behaviour becomes less permissive when the frequency in church attendance increases. However both effects are rather small and the effect of church attendance is only 27