Values Related to the Religious Adherence

The religions and their value systems play a crucial role in the history of human civilization. In the past and in the recent time, the value-based religious differences substantially contribute to the societal conflicts. Thus, the research of the values related to the religious orientation is an important task of psychology and other social sciences. This study is aimed to obtain a more complete insight into the differences in the value orientations between the adherents of the seven major religions in the world: Buddhist, Hindu, Jewish, Muslim, Christian Orthodox, Christian Protestant and Christian Catholic. The results clearly demonstrated, (1), the essential association of the religious or non-religious beliefs with the values, value priorities and value orientations and, (2), the substantial differences between religious or non-religious groups in the value systems. These differences are very probably related to the globally observed distinctions between secularism and fundamentalism and underlying ideological and educational doctrines.


Introduction
The values can be defined as the general categories of beliefs, which serve as the guiding principles in the life of the individuals and the societies (Hofstede, 1980(Hofstede, , 2001;;Musek, 2000Musek, , 2011;;Schwartz & Bilsky, 1987, 1990).The values are connected to the religion in many ways.In the majority of the cultures and societies, the religion and values are closely associated with the ethics, moral principles and virtues.Furthermore, in the eyes of many people, the religious beliefs are even the ultimate basis and rationale for the values, ethical standards and morality.Consequently, the differences and controversies in the value systems of different religions can lead to the conflicts between the adherents of the particular religions and may represent a constant source of the societal tensions and antagonisms.In the modern world, many local and even global confrontations are obviously connected with the disagreement in the value systems developed within the respective religious orientations.
The religious beliefs and the values systems are doubtlessly related to the personal and group decisions connected with the respective behavior and influence therefore the political and social occurrences in a given society.As frequently shown in the history and also corroborated in the scientific research, the practical impact of the religious beliefs and values is real and often misused by political leaders and parties (Habermas, 2004;Pratt, Cheetham, Pratt, & Thomas, 2013).There are many reasons to believe that the controversies between the religion-related values lie in the same core of the past and present controversies between fundamentalistic and secularistic ideologies in the world (Fulton, Gorsuch, & Maynard, 1999;Nandy, 1988;Newman, 1982;Plantinga, 1995;Pratt et al., 2013;Roberts & Sandberg, 2015;Turner, 2015;Williams, 2013).
The better and deeper understanding of the religion-related differences in the value systems is therefore necessary in order to manage the interreligious adversities and conflicts more effectively.Despite the forthcoming recognition that different religions have much in common in their basic value systems, the adversities among the numerous adherents and groups of different religious denominations are still severe and perpetuating.Religious values have been often the target of the empirical research (for a good review see Gorsuch, 1988), yet this research has been rarely especially focused on the value differences.Thus, contrary to the expectations, the empirical research of the religious differences in the value systems is relatively scarce.Comparing four religious groups (Jews, Catholics, Protestants and Greek Orthodox), Schwartz and Huismans (1995) confirmed the hypothesis that the religiosity positively correlated with the "values that enhance transcendence, preserve the social order, and protect individuals against uncertainty", and negatively with the "values that emphasize self-indulgence and favor intellectual or emotional openness to change" (Schwartz & Huismans, 1995, p. 88).The meta-analysis of 21 studies (Saroglou, Delpierre, & Dernelle, 2004) was also focused primarily on religiosity-values associations and less on the differences in the value ratings between the religions.
In this study, the primary objective is to obtain a more complete insight into the differences in the value orientations between the adherents of the seven major religions in the world: Buddhist, Hindu, Jewish, Muslim, Christian Orthodox, Christian Protestant and Christian Catholic.

Method
In the present study, we intended to examine the relations between the value orientations and the adherence to major religious denominations including Buddhistic, Hinduistic, Muslim, Jewish, Christian Catholic, Christian Protestant and Christian Orthodox.The data being analyzed were taken from the sixth wave of the World Values Survey (WVS6, World Values Survey, 2015).Participants 86292 participants have been included into the whole WVS6 project.The data were collected in the period 2010 to 2014.Exactly 60 national samples were drawn from the respective populations equal or older than 18 years.In this study, the full data records for the adherents of seven religions have been retained for the respective analyses: Buddhist (3851), Hindu (1712), Jews (406), Muslim (18079), Orthodox (8324), Protestant (5764), and Catholic (14817).The adherents of all other (very numerous) religions were classified together into the category Else (17304).Additionally, the participants who did not adhere to any religion were categorized into a separate group None (16015).Thus, the records of 86292 participants of both sexes (41101 males, 45083 females, 88 missing) and different ages (M=42.09,SD=16.57) were included into the final analyses in the study (see below the list of all variables in the research model).

Material
The WVS6 represents the sixth, the last already accomplished data collection on 60 national samples throughout the world.It is a part of the World Values Survey (WVS) project leading by the global network of social scientists with the headquarter of WVS association and secretariat in Stockholm.The project is dedicated to the study of values and their influence on social and political life.It started in 1981, and now, it is entering the seventh wave, the phase, which is planned to end in 2018.The entire WVS6 is a huge survey questionnaire, which contains 430 items measuring respective demographic, sociological and psychological variables.The following variables from this instrument were entered into the research model in this study: 1.  (Schwartz Value Survey;Schwartz & Bilsky, 1987, 1990), measured by 11 items:  "It is important to this person to think up new ideas and be creative; to do things one's own way."(code V70): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable self-direction of the Schwartz model (coded as selfdir in results section). "It is important to this person to be rich; to have a lot of money and expensive things" (code V71): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable power of the Schwartz model (coded as selfdir in results section). "Living in secure surroundings is important to this person; to avoid anything that might be dangerous" (code V72): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable security of the Schwartz model (coded as secur in results section). "It is important to this person to have a good time; to "spoil oneself" (code V73): rating scale from 1 ("very much like me") to 6 ("not at all like me").
Item is measuring the variable hedonism of the Schwartz model (coded as hedon in results section). "It is important to this person to do something for the good of society" (code V74): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable benevolence of the Schwartz model (coded as benev in results section). "It is important to help people living nearby; to care for their needs" (code V74B): rating scale from 1 ("very much like me") to 6 ("not at all like me").
Item is also measuring the variable benevolence of the Schwartz model (coded as benev2 in results section).
 "Being very successful is important to this person; to have people recognize one's achievements" (code V75): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable achievement of the Schwartz model (coded as achiev in results section). "Adventure and taking risks are important to this person; to have an exciting life" (code V76): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable stimulation of the Schwartz model (coded as stimul in results section). "It is important to this person to always behave properly; to avoid doing anything people would say is wrong" (code V77): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable conformism of the Schwartz model (coded as conform in results section). "Looking after the environment is important to this person; to care for nature and save life resources" (code V78): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable universalism of the Schwartz model (coded as univer in results section). "Tradition is important to this person; to follow the customs handed down by one's religion or family" (code V79): rating scale from 1 ("very much like me") to 6 ("not at all like me").Item is measuring the variable tradition of the Schwartz model (coded as tradit in results section).5. General preferences for 6 values, measured by 6 items:  Important in life: Family (code V4): rating scale from 1 ("very important") to 4 ("not at all important").Item is coded as family in results section. Important in life: Friends (code V5): rating scale from 1 ("very important") to 4 ("not at all important").Item is coded as friends in results section. Important in life: Leisure time (code V6): rating scale from 1 ("very important") to 4 ("not at all important").Item is coded as leisure in results section. Important in life: Politics (code V7): rating scale from 1 ("very important") to 4 ("not at all important").Item is coded as politics in results section. Important in life: Work (code V8): rating scale from 1 ("very important") to 4 ("not at all important").Item is coded as work in results section. Important in life: Religions (code V9): rating scale from 1 ("very important") to 4 ("not at all important").& Welzel, 2005).Item is coded as wemancip in results section.

Procedure
All participants fulfilled the WVS6 questionnaire items during the period 2010 to 2014.The questionnaire, translated into the various national languages, was administered in the face-to-face interviews by the instructed experts.The details of the processing the WVS6 are available in the online address http://www.worldvaluessurvey.org/WVSContents.jsp.The data contain also the instruments needed for online analyses and are available on the WVS website mentioned above (World Values Survey, 2015).
The WVS6 data are available in several formats (SPSS, Stata, CSV Ascii text, Excel).For our analyses, we utilized the data in SPSS or CSV format with possibility to be analyzed by means of appropriated packages of R program (R Core Team, 2015;Rstudio Team, 2012).Before the statistical analyses, the responses to all items except Y001, Y003, SACSECVAL and RESEMAVAL were reversely coded in order to make the interpretation of the results easier.As a consequence, the higher values of all items denotes higher ratings of the respective values.The values of the variables SACCESVAL and RESEMAVAL were multiplied by the respective weights (variables SECVALWGT and WIEGHTB), as recommended in the instructions of the use of WVS6 data.
For control reasons, all data analyses have been accomplished also by the corresponding algorithms in the IBM SPSS statistics package (IBM SPSS Statistics for Macintosh, 2015).In the following section, only the results obtained by the analyses utilizing the R packages were shown.

Results
In this study, the analyses of data are focused on the differences in the value orientations between 9 religious groups.The value orientations were represented by 21 variables including 10 dimensions of the Schwartz model of values, measured by 11 variables (self-direction /coded as selfdir/, power /power/, security /secur/, hedonism /hedon/, benevolence /benev and benev2/, achievement /achiev/, stimulation /stimul/, conformism /conform/, universalism /univer/, tradition /tradit/), 6 variables measuring the life importance of family (family), friends (friends), leisure time (leisure), politics (politics), work (work), religion (relig) and 4 variables measuring postmaterialistic values (postmat), autonomy (autonindex), secular values (wsecular) and emancipatory values (wemancip).Before the analyses, the values of all variables have been scaled to the continuum from minimum value 0 to maximum value 100 for the sake of better comparability.In the first part of the results section, the relationship between the value dimensions were analyzed in order to define the underlying latent dimensions.In the next parts of this section, the analyses were focused on the differences in the value dimensions between all groups of religious denomination.

Dimensional Analysis of 21 Value Variables
Table 1 displays the means, standard deviations and correlations for all 21 value variables.The correlations between the variables extend from -.53 to .59.With rare exceptions, the correlations are significant due to the great number of respondents (even some correlations low as .01 are significant).Although low to moderate correlations clearly prevail, a lot of correlations are substantial enough to assume that further structural analyses of the correlation matrix of 21 variables would be justified.
Indeed, the respective indices suggest the suitability of factor analysis: Kaiser-Meyer-Olkin coefficient (KMO) amounts 0.83 and Bartlett's test of sphericity is highly significant.The criteria for the number of latent dimensions to be extracted suggest seven factors for factor analysis and five components for component analysis (see Figure 1).
For control reasons, all data analyses have been accomplished also by the corresponding algorithms in the IBM SPSS statistics package (IBM SPSS Statistics for Macintosh, 2015).In the following section, only the results obtained by the analyses utilizing the R packages were shown.

Results
In this study, the analyses of data are focused on the differences in the value orientations between 9 religious groups.The value orientations were represented by 21 variables including 10 dimensions of the Schwartz model of values, measured by 11 variables (self-direction /coded as selfdir/, power /power/, security /secur/, hedonism /hedon/, benevolence /benev and benev2/, achievement /achiev/, stimulation /stimul/, conformism /conform/, universalism /univer/, tradition /tradit/), 6 variables measuring the life importance of family (family), friends (friends), leisure time (leisure), politics (politics), work (work), religion (relig) and 4 variables measuring postmaterialistic values (postmat), autonomy (autonindex), secular values  We decided for the extraction of five components (Principal Components method) and factors (MINRES method) using the fa algorithm in the package psych in R program (Revelle, 2015).Five components account for 51 percent of the total variance in the correlation matrix, while seven factors within the MINRES factor analysis explain about 42 percent of the shared variance.The fit indices also showed that the five-factor solution for the factor analysis is sufficient (RMSEA = 0.46; RMSR = 0.02; TLI = 0.90), the extraction of seven factors will be thus unnecessary.The factors and components were rotated by Promax technique.The loadings of 21 variables on the extracted five factors and components after the rotation are displayed in Table 2 in the first 10 numerical columns.The results of factor and component analyses are very congruent provided slightly different consequence and orientation of the extracted dimensions.The first factor (F1) is practically identical to the first component (C1).The second factor (F2) is congruent with the fifth component (C5), the third factor (F3) with the third component (C3), the fourth factor (F4) with the second component (C2), and the fifth factor (F5) with the fourth component (C4).Note also, that the fourth factor and the second component have the reverse loading orientations.It seems that the results of factor and component analyses are very plausible for the interpretation.For the sake of simplicity, we will consider the results of the component analysis as the basis of psychological interpretation.The first extracted component heavily saturated the variables, which are representative for the self-transcendence in Schwartz model of values or Apollonian values in Musek's model of values (Musek, 2000(Musek, , 2011)): benevolence, universalism, conformism, security and tradition.Thus, the interpretation of this higher-order dimensions of values as low versus high Apollonian value orientation is almost self-evident.Similarly, the second higherorder component (C2) is evidently a bipolar dimension separating the autonomy, secular and emancipatory values from the religious value orientation.This dimension can be interpreted as the religious versus secular value orientation.The third component is associated with the variables power, stimulation, hedonism, achievement and self-direction and can be interpreted as the low versus high Dionysian value orientation according to the Musek's model of values (Musek, 2000(Musek, , 2011)).The fourth higher-order component is closely connected with the variables stressing the importance of friends, leisure time and family in life.This component can be interpreted as low versus high orientation to affiliative values.Finally, the fifth component is closely connected with the postmaterialist and emancipative values and can be interpreted as low versus high emancipative value orientation.

Connections of Value Dimensions with the Religious Groups
How are the value dimensions connected with the categories of religious denomination?To answer this question, which is an important aspect of the problem in this study, we performed multivariate analyses on mixed data, continuous and categorical.For the sake of sparing space, we will focus only on the results of PCAmix and PCArot procedure from the PCAmix package of R program (Chavent et al., 2014).The PCAmix algorithm performs principal component analysis of the analyzed variables including the properly transformed categorical variable (religions9, categories of religion denomination in our case), while the PCArot algorithm executes the orthogonal (varimax) rotation of the extracted dimension.The extracted five latent dimensions with the loadings on both continuous and categorical variables are shown in the last 5 numerical columns in Table 2.The extracted five dimensions accounted for 40 percent of the total variance in the variable matrix.Note that the loadings of the nominal categories are represented by coordinate coefficients that can exceed the value of 1.
The loadings on five dimensions (from dim1 to dim5 in the header of the table) correspond very closely to the factor and component loadings obtained by previous multivariate analyses.Thus, the interpretation of the dimensions should be practically the same as mentioned before.The dimensions of mixed data are highly congruent with the dimensions from the component and factor analyses.The important additional information is the connection of the newly obtained latent dimensions with the religious adherence.We can see, that the Orthodox group is very high on the dimension of Apollonian values (dim1; benevolence, universalism, conformism, security, tradition), while the Hindu, Protestant, and Buddhist group are rather low.On the religious versus secular values dimension (dim2; religious values versus autonomy, secular and emancipatory values), the None denomination group is extremely high, as expected.This dimension strongly separates the non-religious respondents with secularistic orientation from all religious groups.Interestingly, Jewish, Buddhist and Orthodox group stand remarkably higher on the secularism dimension than other religious groups, which represent more strict religious orientation.Dionysian values (dim3; power, stimulation, hedonism, achievement and self-direction) are associated with the Hindu, Jewish and Protestant group and are remarkably low in the Orthodox group.Affiliative values (dim4; importance of friends, leisure time and family in life) are most expressed in the Protestant, Buddhist and Orthodox group, and are very low in the Jewish and Hindu group.The emancipative and postmaterialistic values (dim5) are high in Hindu, Jewish and Catholic group and remarkably low in the Muslim group.

Differences between Religious Groups
Still more accurate insight into the differences in the value orientations between the religious groups can be obtained by means of analysis of variance and discriminant function analysis.Table 3 is showing the results of the multivariate analysis of variance (MANOVA) in the first five numerical columns and the results of the discriminant function analysis in the last six columns.According to the MANOVA, the religious groups differ very significantly in all value variables (certainly, the overall difference is particularly significant).The smaller the Wilks's Lambda and the greater the F value, the more the respective variable contribute to the difference between the religious groups.The Wilks's Lambda and F values indicate the largest differences between the religious groups on the importance of religion (relig).
It is clear from the Figure 2, that the non-religious group (None) is far lowest on rated importance of religion in opposition to all religious groups, especially Muslim, Orthodox and minor religion group (Else).Very pronounced differences between the groups can also be detected along the secular, emancipatory, autonomy and tradition values.Again, the non-religious group is at extreme position, the highest on the first three value variables and the lowest on the tradition.According to the more thorough inspection of the Figure 2, the Muslim and Orthodox group are low on the secular and emancipatory values and high on the tradition.Interestingly, the Orthodox group has the leading position in several value scales: beneath tradition in benevolence, universalism, conformism and rated importance of friends.On the other side, this group is the lowest on hedonism, rated importance of politics and secular values.Buddhist group is high on the rated importance of politics and autonomy index yet low on the self-direction, stimulation and achievement.Hindu group is low on security and rated family importance and the highest on stimulation values.The Jewish group is the highest on power and post materialistic values and very low on the rated importance of friends, leisure time and work.The Muslim group is the highest on security, hedonism, achievement and rated importance of religion, but very low on autonomy index and emancipatory values.The Catholic group has highest scores on self-direction and rated importance of leisure time and work.The Protestant group has neither highest nor the lowest position on any of the value variables.The loadings on five dimensions (from dim1 to dim5 in the header of the table) correspond very closely to the factor and component loadings obtained by previous multivariate analyses.Thus, the interpretation of the dimensions should be practically the same as mentioned before.The dimensions of mixed data are highly congruent with the dimensions from the component and factor analyses.The important additional information is the connection of the newly obtained latent dimensions with the religious adherence.We can see, that the Orthodox group is very high on the dimension of Apollonian values (dim1; benevolence, universalism, conformism, security, tradition), while the Hindu, Protestant, and Buddhist group are rather low.On the religious versus secular values dimension (dim2; religious values versus autonomy, secular and emancipatory values), the None denomination group is extremely high, as expected.This dimension strongly separates the non-religious respondents with secularistic orientation from all religious groups.Interestingly, Jewish, Buddhist and Orthodox group stand remarkably higher on the secularism dimension than other religious groups, which represent more strict religious orientation.Dionysian values (dim3; power, stimulation, hedonism, achievement and self-direction) are associated with the Hindu, Jewish and Protestant group and are remarkably low in the Orthodox group.Affiliative values (dim4; importance of friends, leisure time and family in life) are most expressed in the Protestant, Buddhist and Orthodox group, and are very low in the Jewish and Hindu group.The emancipative and postmaterialistic values (dim5) are high in Hindu, Jewish and Catholic group and remarkably low in the Muslim group.

Differences between Religious Groups
Still more accurate insight into the differences in the value orientations between the religious groups can be obtained by means of analysis of variance and discriminant function analysis.Table 3 is showing the results of the multivariate analysis of variance (MANOVA) in the first five numerical columns and the results of the discriminant function analysis in the last six columns.According to the MANOVA, the religious groups differ very significantly in all value variables (certainly, the overall difference is particularly significant).The smaller the Wilks's Lambda and the greater the F value, the more the respective variable contribute to the difference between the religious groups.The Wilks's Lambda and F values indicate the largest differences between the religious groups on the importance of religion (relig).
It is clear from the Figure 2, that the non-religious group (None) is far lowest on rated importance of religion in opposition to all religious groups, especially Muslim, Orthodox and minor religion group (Else).Very pronounced differences between the groups can also be detected along the secular, emancipatory, autonomy and tradition values.Again, the non-religious group is at extreme position, the highest on the first three value variables and the lowest on the tradition.According to the more thorough inspection of the Figure 2, the Muslim and Orthodox group are More condensed information is provided by the differences between the religious groups on five higher-order dimensions of values (see last 9 rows of Table 3 and the graphical picture in Figure 3).We can see a sharp difference in factor scores of Apollonian values between the Orthodox group (high) and Buddhist, Hindu and non-religious group.Even more evident is the contrast on the secular values between the non-religious (extremely high) and Muslim group (extremely low).On the other higher-order values, the differences are also very significant, yet somewhat less obvious.The Muslim, Hindu and Jewish group are the highest on the Dionysian values, where the Buddhist and the non-religious group have the lowest scores.The Hindu and Jewish group have outstanding lowest scores on the affiliative values (but highest on the emancipatory values) and the Muslim and Orthodox group are very low on the emancipatory values.The next method, useful for investigate the differences between the religious groups, is Discriminant Function Analysis.Usually, the Discriminant Function Analysis serves to predict a categorical variable (religions in our case) by different continuous variables.21 value variables are highly significant predictors of religious adherence (by them, 45.69% of individuals can be correctly classified into the categories of religious adherence), as well as the five higher-order value dimensions (41.27% correctly classified individuals).Discriminant Function Analysis yielded eight significant discriminant functions, yet the first three explained about 90 percent of the discriminant variance (the first function itself explained about 77 percent).Thus, we should retain only first three functions for further analyses.The last six columns in the Table 3 display the standardized coefficients and the respective loadings on the first three discriminant functions.The predictor variables that have highest standardized coefficients and highest loadings on discriminant functions contribute most to the differentiation between the groups of religious adherence.
Discriminant Function Analysis yielded several other interesting results.Canonical correlation for the first discriminant function has a value of .69suggesting that about 47 percent of the variance in the religious adherence is accounted for by our model.Considering all discriminant functions, about 72 percent of the variance in the religious adherence is accounted for by the model.Thus, the results of discriminant function analysis indicate a very significant predictive power of the independents (value variables) for the adherence to the religious groups.The value of the overall Wilks's test of 0.41 is highly significant demonstrating thus the substantial relationship between the value variables and the discriminant groups (note that the overall Wilks's test is important for discriminant analysis and for MANOVA).The Press Q Statistics 37689.3exceeded very strong the critical value of 20.090 suggesting thus that the analyzed model has fair predictive power.
Provided the relative importance of the first discriminant function, the rated importance of religion is far strongest predictor of religious adherence.Considering the group centroids (mean values of discriminant scores) for given religious category (Table 4), we can see that this predictor is mostly connected with the Muslim and Orthodox group.The traditional values are the next important predictor, connected to both mentioned groups in the same direction.The Jewish and the Buddhist group took the opposite position connected with the predictors with highest negative loadings on the first discriminant function: secular values, emancipatory values and autonomy index.All other variables have smaller although almost always significant predictive power for the classification into the religious categories.The relations between the group centroids and the value variables are displayed in the Figure 4.Even more inspective insight into the relationships between five higher-order dimensions and the respective group centroids is shown by the Figure 5. Figure 4 is showing the positions of the group centroids for all religious categories in the space of first two discriminant canonical functions (Can1 and Can2) as well as the directions and values for the 21 predictor variables.Very clearly, we can see the separate position of the non-religious group, extremely low (as expected) on the rated importance of religion (relig) and quite high on the secular values.Among the religious groups, the Muslim, Orthodox and minor religions (Else) groups are the highest on the rated importance of religion and the lowest on secular values taking thus the position that is most distant to the non-religious group.Beside this, the Muslim group is closely connected with the tradition and achievement values, while the Orthodox and Else group are close to the benevolence, conformism, universalism and self-direction values.The Buddhist group is connected to the secular values, autonomy index and rated importance of leisure time and has the position closest to non-religious group among all religious categories.Rather close to the Buddhist group is the Jewish group, which is distinctively oriented toward the emancipative values.Finally, the Hindu group is characterized by post materialistic values, stimulation and emancipative values.The remaining Christian groups, Catholic and Protestant have a rather balanced position between the non-religious, Buddhist, Jewish and Hindu groups on the one side and Orthodox, Else and Muslim angroup on the other side.Note.Included are the vector directions and values for the 21 predictive variables.
Five higher-order value dimensions retain a great amount of the predictive power of the model represented by the 21 predictors (Wilks's test = 0.564, Press Q Statistics = 28669.55,both highly significant).In the model with five higher-order value dimensions as predictors, all discriminant functions were significant, the first three accounting for about 97.5 percent of the variance (the first for 84 percent alone).
Canonical correlation for the first discriminant function was 0.610.Thus, about 37 percent of the variance in the religious adherence is accounted for by the model considering the first discriminant function and 48 percent considering all discriminant functions.
Figure 5 displays the religious group centroids together with the directions and values of five higher-order value dimensions.Again, the non-religious group has most distant position to other groups far exceeding them on Secular higher order dimension.The Muslim, Else and Orthodox group stand on the opposite pole away from the secular position (and therefore on the opposite, religion side of the dimension).Among them, the Orthodox group is connected to the Apollonian and Affiliative second-order dimension, the Muslim group to the Apollonian and Dionysian dimension, and the Else group mostly to the Dionysian dimension.The Jewish and Hindu group are associated with the emancipatory and Dionysian secondorder dimension and both groups are low on Apollonian and Affiliative dimension.The Buddhist group is oriented toward Secular, Emancipatory and Affiliative dimension.Among all religious denomination groups, this group is the closest to the non-religious group.The Protestant group is placed near to the Else, Catholic and Buddhist group with the affinity to the Dionysian and Emancipatory dimension.Finally, the Catholic group is well-balanced with almost central position along the secondary value dimensions.A further test of the predictive power of the values was made by means of the multinomial regression analysis.Multinomial regression analysis applies the logic of logistic regression analysis for the case of the multiple nominal dependent variable (religious adherence in this study).The results of the multinomial regression analysis complied well with the results of discriminant analysis.Again, the dependent variable (the religion adherence) was regressed by means of two predictive models, the 21-predictors model with 21 values as predictors and the 5-predictors model with five higher-order value dimensions.The overall impact of the predictive variables on the religious adherence is highly significant: the overall Likelihood Ratio Test yielded chi square 23771.932with p-value less than 0.001 for the model with 21 values as predictors and 16479.914with p-value less than 0.001 for five higher-order predictors.All single Likelihood Ratio Tests for 21-predictor model and for the 5predictor model were significant.The strongest predictor in 21-predictor model was the rated importance of religion, while the strongest predictor in 5-predictor model was the secular higher-order dimension.So-called pseudo R squares amounted .53 for 21 value variables and .41 for five higher-order value dimensions (Cox-Snell pseudo R squared), .55 and .42(Nagelkerke or Cragg-Uhler pseudo R squared) and .21and .14(McFadden pseudo R squared).All pseudo R squared coefficients are rough estimates of the proportion of the total variability in dependent categories accounted for by the model.The percent of correctly classified cases was 46 for 21predictors model and 41 for 5-predictors model, almost exactly the same as in the discriminant function analysis.

Discussion
As hypothesized, the religious groups substantially differ in their value orientations.In Table 5, the main characteristics of different groups' value orientations are summarized according to the obtained results.The results clearly demonstrated that the strongest difference exists between the non-religious group (None) and the genuine religious groups.Nevertheless, there are also strong differences between the religious groups themselves.The value-profile of some religious groups, for example Muslim, Else and Orthodox, is almost the inverse of the profile of non-religious group.However, some other religious groups have profiles that resemble the non-religious value-profile (Jew, Buddhist) and still other are somewhere in between (Hindu, Protestant, Catholic).

Overall Intergroup Distances in Value Orientations
A further reasonable question is, how close or distant are different religious groups considering their value orientations.The inspection of distances between the religious group means (group centroids) is an accurate way in order to assess the closeness or distances between the groups representing the religious adherence (see Figures 4 and 5).Yet, we must keep in mind that the first discriminant function very dominantly represents the differences between the groups.Thus, the distances between the groups are well represented by the distances along the first discriminant function.It is obvious that the non-religious group (None) has far the most distant position in relation to other groups.This is logical and expected for the group of nonreligious people, whereas all other groups represent the samples with factual religious adherence.Among all religious groups in the strict sense, the Muslim group is the most distant from the non-religious group and also the most distant within the strict religious groups.The Muslim group has also strongest religious and lowest secular orientation among all groups.Similar structures of the relationships between different religious groups can be obtained by using the algorithms for calculating the Euclidean distances among the groups.Figures 6 and 7 are showing metric scaling results (Multidimensional Scaling: Cox & Cox, 2001;Gower, 1966) and the results of a robust hierarchical cluster algorithm (hclust in R package cluster, Maechler, Rousseeuw, Struyf, Hubert, & Hornik, 2015), both based on the Euclidean distances.Finally, the differences between the groups of religious adherence can be integrated and projected into one single dimension.Using the ICLUST algorithm (Revelle, 2015), a special combination of clustering and factoring approach may be performed, revealing the hierarchical cluster structure (see Figure 8) and structure loadings on a single dimension, which subsumes or integrates the differences between the groups (Figure 9).Thus, we can really obtain a good insight into the overall differences between the groups.
As we already know, the strongest predictors of the religious intergroup differences are rated importance of religion among 21 value variables and secular dimension among higher-order value dimensions.Considering also other predictors, the common distance dimension very clearly distinguished between the high secular, emancipatory, autonomy and low religious and Apollonian orientation (represented most pregnantly by the non-religious group) versus high pro-religious and Apollonian and low secular, emancipatory and autonomy orientation (represented mostly by Muslim, Else and Orthodox groups).Figure 8.The Hierarchical Cluster Structure of Religious Groups Ssing the ICLUST Algorithm (Revelle, 2015) ICLUST of the religious group data Note.The algorithm is based on the special combination of clustering and factoring approach.
The pronounced religious and anti-secular position of the adherents of minor religions (Else group) is not surprising.A great number of minor religious groups is characterized by the strong religious commitment including the demands for a decisive role of the religion in public life.The social seclusion of minor religious communities is more the rule than exception.These communities tend to form the life conditions that maximally possible conform the religious rules.Nevertheless, there are some minor religious groups that are secularly oriented and that even evade classical definitions of a religion, for example the nature worshippers, Wicca or Scientology.Another, more traditional subgroup of the Else category is represented by the local religions like Voodoo.

Causal Factors of Differences in Value Orientations
The results clearly demonstrate that the religious or non-religious beliefs are essentially associated with the values, value priorities and value orientations.The education in a given society is certainly under influence of the cultural and other societal premises of that society including the religious conditions.All known religions strongly emphasize different values within and beyond the strict domain of pure religious values.The faith, the belief in God and religious doctrine represent only the core of a religious value framework, which is most often accompanied by further values, e. g. love, hope, honesty, justice, decent life, concern for others, marital fidelity, courage, modesty and others.The value system of a religious person is therefore substantially connected with the value priorities of the respective religious group.On the other side, the non-religious people are not lacking a value system at all.Yet, it is different and sometimes opposite to the values of the religious communities.A non-religious person is normally sharing a secular ideology, which can be more or less confronted with the pro-religious orientation.It is plausible therefore, that the value orientations of the non-religious people will be at difference with the value orientations of the adherents of different religions the more stronger is the degree of the religiosity and anti-secular attitude of the religious group.
Several factors might be in the play as causal factors of the value differences between the religious groups.Specific cultural milieu and education probably represent major common denominators of the influences shaping the value orientations of the adherents of different religious groups.For example, religious education varies from the extremely exclusive or fundamentalist indoctrination stressing the respective religion as the only true religion, to the opposite, more tolerant and respectable look at other religions.In similar manner, a non-religious education may vary from the extreme antireligious secularism to the non-religious orientation, which is friendly and tolerant toward the religious people.
According to the results of this study, it seems very probable that the adherents of some religions are closer to the religious exclusivism or even fundamentalism and more strongly reject the secular ideology.This is true especially for three religious groups: Muslim, Christian Orthodox and minor religions group (Else).The adherents of other religions have more tolerant and respecting attitudes toward the other religions (Catholic, Protestant, Hindu), or are even close to the secular group (Jew, Buddhist).

Societal Impacts of Religion-Based Differences in Value Orientations
The debate over the value differences between the religions is far from being of only academic or experts' interest.The great majority (80 percent) of world population is religious and a wide percent of the rest has a definite secular value orientation.The values are by definitions the beliefs, which serve as life guiding principles (Musek, 2000(Musek, , 2011;;Schwartz & Bilsky, 1987, 1990).Coupled with the proper behavioral intentions, the values may deeply impact our daily and life decisions as well as our life routines.Our every-day experiences massively testimony about the importance of the religion-based differences in the values and value orientations.In more extreme cases, the antireligious and pro-religious value systems can serve as the ideological basis for serious forms of intolerance, prejudices, discrimination and violence in the society, not to mention the religion-based terrorism, which is notorious in the contemporary world.Both extreme antireligious secularism and extreme religious exclusivism or fundamentalism accompanied different totalitarianisms very often in the history.Interestingly, the extreme antireligious secularism has been until recently a part of the ideological justification for the repression of the religious people that represented a majority of population in the countries suffering under totalitarian regimes.

Figure 1 .
Figure 1.Extraction Criteria for the Factor Analysis of 21 Value Variables Parallel Analysis Scree Plots Note.M and SD are used to represent mean and standard deviation, respectively.* p<.05; ** p<.01.

Figure 2 .
Figure 2. The Differences between Religious Groups along the Dimensions of Value Orientations

Figure 3 .
Figure 3.The Differences between Religious Groups along Main Factors of Value Orientations

Figure 4 .
Figure 4.The Positions of the Group Centroids of all Religious Groups in the Space of First Two Discriminant Canonical Functions (Can1 and Can2)

Figure 5 .
Figure 5.The Positions of the Group Centroids of all Religious Groups in the Space of First Two Discriminant Canonical Functions (Can1 and Can2)

Figure 9 .
Figure 9.The Loadings of Nine Religious Groups along the Common Distance Dimension

Table 1 .
Means, Standard Deviations, and Correlations

Table 2 .
Factor and Component Loadings of 21 Variables

Table 3 .
Main Results of the Multiple Analysis of Variance (MANOVA) and Discriminant Function Analysis

Table 4 .
Mean Discriminant Scores (Group Centroids) for Religious Groups

Table 5 .
Summarized Value Orientations of Nine Religious Adherence Groups