PREDICTING CIVIL RELIGION AT A CROSS-CULTURAL LEVEL

The concept of civil religion has caught major attention among scholars studying the junction of religion and politics (J.-J. Rousseau, E. Durkheim, R. Bellah). The notion focuses on the phenomenon of cultural contents sacralizing and ritualizing the ruling political institutions of a society, extending support to the integration of the political and social system at a cultural level. The notion of civil religion has recently been operationalized crossculturally, but light has not been shed upon its predictors. In this paper authoritarianism is tested as a predictor of civil religion cross-culturally. Four student samples of Bosnian, Serbian, Slovenian and US students were analyzed. Very strong, significant associations between authoritarianism, as operationalized by a modified Lane scale, and civil religion were found in all cases. Moreover, upon introducing femininity, anxiety and gender into the analysis, a strong, dominant and significant impact on the part of authoritarianism was still found when civil religion was observed crossculturally. When the same predictors were applied to explaining general religiosity, authoritarianism fell short of being a significant predictor in most of the environments observed. Such results suggest an especially close link between civil religion and authoritarianism.


INTRODUCTION
Recently, a cross-culturally valid composite measure of civil religion has been presented (Flere & Lavrič, 2007).The authors indicated, that regarding civil religion, 'it is evident that it entails some linking of the nation, of the state, of the ethnic group and of sovereign political authority to some principle that is higher than the empirical, the observable or the human, tending towards finding a metaphysical, transcendent, religious or theistic explanation and legitimation of secular phenomena '. (p. 596) In arriving at a working definition of civil religion the authors follow Coleman's precise definition of civil religion, as phrased by Gehrig, as 'the religious symbol system which relates the citizen's role and the … society's place in space, time and history to the conditions of ultimate existence and meaning ' (1981, p. 52).Within this framework, the idea of the 'broken covenant' between the chosen people and their God is usually included (Bellah 1967;Gehrig, 1981, p. 53).
Flere and Lavrič introduced a cross-culturally valid scale for the study of this phenomenon, the study of which was previously limited mostly to the United States (Wimberely et al., 1976;Wimberely & Christenson, 1979), Furseth's (1994) being the main attempt to define the phenomenon in a European national context, encountered a lukewarm form of the phenomenon.The authors studied civil religion in the Bosnian, Serbian, Slovenian and US environment, on student samples (Flere & Lavrič, 2007, p. 595).
Following Bellah (1967), central to the concept of civil religion is the understanding of a special position of one's people/nation/state vis-à-vis God or some similar construct, denoting transcendence.This is usually elaborated by way of an allegedly existing covenant, a special contract between God and one's people.The covenant imposes obligations on the people, but God's protection on the other.The concept allows for the people not quite living up to the obligations set in the covenant, creating a tension and allowing for special political dynamics.
The Flere and Lavrič scale (2007) contained the following items: "Our country has a special covenant with God", "It is not really possible to be a good patriot without being a true follower of my religion", which the authors hold to presuppose the superiority and uniqueness of one's religion in the meaning of patriotism, where religion and civil loyalty blend into one, "It might be said that our country does not comply with all of its religious duties", "Each individual of my nationality should belong to my religion", aiming to access the sacred link between nation and religion, and »God has intervened in the history of my nation by testing its true faith" suggesting God's special care for and attention to the particular nation, a statement too general to strictly imply a covenant.The authors attained very good metric characteristics for the scale by way of exploratory and confirmatory factor analysis and by way of Cronbach reliability, at cross cultural inspection.
As to averages found by these authors, they were highest in the American sample.Serbs come second, which could be expected, owing to the nature of Serbian Orthodoxy, which stresses ethno-nationalism (Vrcan 1995).That the Serbian averages distantly trail those of the Americans was considered by the authors as possibly to be unexpected (owing to the nature of their religion) and may indicate "fatigue" after the Balkan wars of the 1990s.Somewhat more surprising is that Bosnians from the Muslim part of the country are very close to the averages in the Serbian sample.Bosnians only recently established themselves firmly as a nation, "the Bosnyaks," during the Bosnian war of the 1990s, when their difference from the Serbs and Croats was confirmed as being not merely religious, but also political and "ethnic."The Slovenians indicate invariant lows.Civil religion in this case does exist, but it is a minor phenomenon in size.One might ask whether such a lukewarm situation would be encountered in the European West as well (in line with Furseth 1994).Flere and Lavrič (2007) also underscore that 'it is essential to make a clear distinction between civil religion and religion in the standard meaning of the term, following Bellah (1967) and Wimberley et al. (1976).Bellah treated civil religion as a 'religious dimension', implying that it is somehow 'part of a wider concept of general religiosity, but still particularized ' (1967, p. 11).
However, the 2007 study falls short of explaining the predictors of civil religion and of illuminating the social and psychological factors producing its presence and strength.In this paper, we present some results following the same framework of analysis and applying the same instrument.
The relationship between denominational religion and loyalty, on the one hand, and civil religion, on the other, may be an obstacle difficult to bypass in crosscultural analysis, as situations may be truly different.The American case of civil religion is classic, where civil religion overlies denominational loyalty within a tolerant pluralism, allowing for a peculiar and distinct civil religion to develop.This may be radically different from the situation in countries that are basically monodenominational, the denomination being also historically entrenched and hegemonic (e.g., Slovenia and Serbia, in our study).Further types of civil religion are acknowledged (Hammond 1980).
Since civil religion has to do with the social integration of nation, state, ethnic group and sovereign authority, it seems logical to study whether authoritarianism might be a good predictor of civil religion.
Civil religion as social phenomenon and as notion is closely associated with social integration and may be considered an instrument for its provision, enhancement and imposition (Gehrig, 1981, 56).In fact, although civil religion is sometimes noted for its 'democratic nature', Gehring is right in contending it is 'is a socially constructed myth, which is more or less believable, depending on societal conditions ' (1981, p. 53).The very idea of 'broken covenant' leaves ample space for variance between belief and practice, whereas Bellah allows for despotism to be the future of the American political system (1998).As Flere and Lavrič have found (2007), civil religion appears as a cross-cultural feature of political culture, possibly not universal in the contemporary world, but nevertheless functioning and growing in very diverse environments.It is thus a pertinent question what it is predicted by.
One is tempted to consider authoritarianism.A common element in authoritarianism, regardless of whether it is considered as deep structure as do the original authors (Adorno et al., 1950), or an attitude, as does Altemeyer (1988), civil religion could be associated with the conservative and right wing nature of the two.
The assumption present in both approaches integration of the political system, of the ethnic groups and/or of the state, by a sacred past enters into the picture.To the contrary, attempts to demonstrate the existence of left wing authoritarianism (Eysenck, 1954;Shils, 1954) have been fruitless (Stone & Smith, 1993).Thus one might well look for an association between authoritarianism and civil religion.
Authoritarianism has been proven to be not only a psychological phenomenon, but also a mechanism having to do with group and societal integration.The findings on the 'rigid cognitive style' and 'intolerance of ambiguity' (Frenkel-Brunswik, 1954), 'narrow mindedness' and discrimination (Haddock & Zanna, 1994;Hunsberger, 1996), typical of authoritarianism, making it 'psychologically unhealthy' (Roiser & Willig, 2002, 74) may well still hold, but it is possibly imminent to be found to a certain degree in any population, because of the associated need for social conformity, as demonstrated by Feldman (2003).Newer conceptualizations of authoritarianism underscore 'perceived threat' (Feldman, 2003) and 'insecurity' (Oesterreich 2005), constructs which may also be of relevance in the formation of a concept of civil religion.
We shall limit our analysis to three well-known psychological correlates of religiosity: 1. Authoritarianism has been found to correlate positively with religiosity by many researchers.The existence of such a relationship was brought to the forefront by Fromm (1950).Pursuing from his critical theory positions, he advanced the idea of two types of religion and religiousness: an authoritarian type, underscoring man's unconditional submission and degrading him, and a humanist one.Similarly, Adorno and associates (1950) made arguments that facilitated the study of religion within this context.Adorno wrote: 'Properties of religion/ assume an aspect of rigidity and intolerance such as we expect to find in the prejudiced person… Under contemporary circumstances religion loses 'the intrinsic claim to truth'…'and more hostile, destructive and negative features come to the fore' (Adorno 1997, pp. 434-435).
Other efforts to examine the relationship between religiousness and authoritarianism included Allport (1950), who focused on the comprehension of religiousness in relation to prejudice (the latter being considered as closely associated with authoritarianism) and found a positive relationship between the two (Adorno et al., 1950; see also McFarland & Adelson, 1996).Allport solved this seemingly surprising and disturbing finding by differentiating between immature (i.e.extrinsic) and mature (i.e.intrinsic) religiousness (Allport, 1950): the latter was treated as being authentic, while the former was instrumental for extra-religious ends (Allport & Ross, 1967), thus possibly including prejudice, thus coming close to identifying extrinsic religion as authoritarian.
2. Femininity, the explanatory variable suggested in the study of gender differences, which was to tap 'cognitive schemas' typical of females, although present in males to a lesser extent as well (Bem, 1981).We are speaking of the feminine orientation including compassion, understanding, warmth, an inclination to help the suffering, which is usually measured by the Bem scale.Femininity is known to be associated with religiosity (Thompson, 1991;Francis 1997Francis , 1998;;Flere, 2007).
3. Trait anxiety, as measured by the trait form of the State-Trait Anxiety Inventory (STAI - Spielberger et al., 1970).The relationship between trait anxiety and civil religion could pursue from the relationship between trait anxiety and general religiosity.Concerning this relationship, C.G. Jung wrote that of his many hundreds of clients, he believed that each 'fell ill because he had lost that which the living religions of every age have given to their followers, and none of them has been really healed who did not regain his religious outlook' (as cited in Koenig et al., 2001, p. 201).In this manner religiosity is often seen as a coping mechanism for dealing with various stressors (e.g.Pargament, 1997) and a number of studies have indeed related religiosity to decreased levels of anxiety (e.g.Baker & Gorsuch, 1982;Hertsgaard & Light, 1984;Petersen & Roy, 1985;Bergin et al., 1987;Koenig 1993;Koenig, 1995;Ventis, 1995;Maltby, Lewis & Day 1999;Davis et al., 2003).Admittedly, the results of studies relating anxiety and religiosity are not entirely consistent, especially when a cross-cultural perspective is taken into account.Nevertheless, substantial research literature associates various measures of religiosity with positive mental health (for review see: Koenig, 1994;Koenig et al., 2001).It thus seems reasonable to use a measure of psychological well-being as an important psychological predictor of religiosity, whereas civil religion is considered a dimension or closely associated with religiosity.The use of the STAI instrument is especially appropriate in this regard since, according to Andrade and collaborators (2001), it represents a good measure of general psychopathology, rather than an anxiety-specific construct.In their words: 'In fact, none of the 20 STAI-T items measures specific features of anxiety: 15 items correspond to the measure of general negative affect and 5 measure positive affectivity' (p.373).

HYPOTHESES
On the basis of what has been said about authoritarianism we first hypothesized that there would be a strong association between this construct and the civil religion that we wanted to explain.Such a hypothesis may be questionable and too bold, if civil religion is regarded as part of the moral cement of a democratic polity.In fact, statements to such an effect are to be found mainly outside strictly scientific discourse.These include statements describing civil religion as 'a synthesis of democratic and deistic values' (quoted by Gehrig, 1981, p. 53), or talk of a 'democratic faith'…'based on liberty, equality and justice' (Gehrig, 1981, p. 52).Gehrig attributes such parlance to theologians and philosophers, but does mention it in the context of the scientific study of civil religion.Thus we established this hypothesis mainly following the finding of civil religion being strongly associated with traditionalism (Flere & Lavrič, 2007) and religiousness being strongly associated with authoritarianism (Kahoe, 1977;Flere & Klanjšek, in press).
Our second hypothesis is more specific.We hypothesize that authoritarianism is a stronger and more reliable predictor of civil religion as compared to general religiosity.We expect to find different psychological predictors for religiosity as compared to civil religion, the main difference being precisely in a greater association between civil religion and authoritarianism.According to Flere and Lavrič (2007, p. 560), this would ensue from the particular position of civil religion, its association with social integration and, in particular, with the value complex of traditionalism that glorifies the societal past.

METHOD Sample
Within a more extensive study of religiosity, we applied a scale of civil religion to samples of university students in Maribor, Slovenia (Roman Catholic environment, N = 468), Sarajevo, Bosnia and Herzegovina (Muslim environment, N = 439), Niš, Serbia (Christian Orthodox environment, N = 427) and Auburn, Alabama, USA (predominantly Protestant environment, N = 450).
Data collection was carried out in the spring of 2005 at the respective universities, almost exclusively among students of the social sciences and humanities and almost exclusively among first-and second-year students.The portion of male respondents varied from 34% in Bosnia and Herzegovina to 48% in the USA.The mean age was 21.3 (S.D. = 1.5), all national samples being close.Data collection was carried out with the support of teachers and teaching assistants at the specified institutions.
Our set of four environments includes the following variants: one where there is a more ethnically pronounced dominant Eastern Orthodox religion (Serbian Orthodoxy); one universal allowing for ethnicity as a 'natural community' and in fact serving ethnic consolidation well (Roman Catholicism); one that down-plays the relevance of ethnic belonging and nationality in favor of the universal community of believers (umma, Islam) and, finally, the American, predominantly Protestant denominationalist environment, known for a variety of stands on this issue and tending to favor an amalgamation of ethnicities into 'the first new nation' (Lipset, 1964).Besides these essential differences among the environments under study, situational circumstances also vary: Bosnia and Herzegovina and Serbia underwent a major war experience on their territories, an experience still fresh in the minds of the respondents, whereas the same applies neither to Slovenia nor to the USA.Thus, results should vary, with cultural and situational factors intermingling and possibly being difficult to decipher.

Procedure
The instrument applied was a questionnaire containing varied items, concentrated on various measures of religiosity and its possible correlates (e.g.anxiety, gender orientation, delinquency, demographic variables and so on).The filling out of the questionnaire was conducted in groups of 10 to 40 students under the supervision of research team member and took about 40 minutes.It was carried out in spring, 2005.The questionnaires were translated from the original and back, so that possible errors in translation were avoided.

Variables and instruments
Authoritarianism was tapped by the Lane scale, which is regarded as 'the best short form of the F scale currently available' (Christie 1991;556), supplemented by one item, as described by Flere (2007).
Femininity was tapped by the Bem scale, abbreviated as reported by Flere (2007).
Civil religion was tapped by the 5-item scale of civil religion formulated by Flere and Lavrič (2007).
Religiosity was tapped by a summation scale of 5-point response items pertaining to belief in God, in an afterlife, in hell, in heaven and in the soul.
Trait-anxiety was tapped by STAI (Spielberger et al., 1970).Gender as variable was also included.Its relevance has been explained above with regard to femininity.

RESULTS
Zero-order correlations between civil religion and authoritarianism by samples attained the value of .475**for the Bosnian sample, .498**for the Serbian one, .407**for the Slovenian one and an extremely high value of .545**for the US sample.Thus, we were working from unquestionably very high associations in the analysis of the phenomenon.As for religiosity, the respective values were .335**,.165**,.136*,and .097,indicating positive correlations, but very culture dependent and not even significant in one case.The correlations between civil religion and religiosity ranged from .423** for the Bosnian sample to .524** for the Slovenian one.In Table 2, one finds much cross-cultural variety, but the impact of authoritarianism and only authoritarianism remains unquestionably strong in all cases.Femininity has significant, but much lower, impact on civil religion in three out of four samples, while anxiety is a significant predictor only among the Serbs.As demographic variable, gender indicated one marginal case of significance, as to females being more civil religious in the US sample.Basically, this analysis upheld our first hypothesis.
We further tested the same constructs as predictors of religiosity (Table 3).The data in Table 3 support our second hypothesis.Authoritarianism, when controlled for the chosen variables, is not a significant predictor of religiosity in three out of four samples observed.Even in the case of Bosnia where Beta is significant, its value is substantially lower than it is in the case of explaining civil religion.Instead of authoritarianism, it is rather femininity which prevails in explaining religiosity.Again, probably for specific cultural reasons, Bosnia is an exception in this regard, allowing for a somewhat different set of psychological correlates in the case of Islam.Authoritarianism is a strong predictor of religiosity in that environment, even withholding such a control.

CONCLUSION
Our investigation contributed to the illumination of the predictors of civil religion, while simultaneously describing some of its deep structures.It also indicated some in-depth differences between civil religion and religiosity.
We confirmed that (1) authoritarianism is a strong and robust predictor of civil religion at a cross-cultural level and (2) that authoritarianism is a stronger and more reliable predictor of civil religion as compared to predicting general religiosity.Thus, our data suggest civil religion to be deeply bound-up with an authoritarian personal structure.
This confirms a difference, a specificity of civil religion when compared to the general notion of religiosity, possibly a more direct association of civil religion with societal integration.It also lends support to Flere and Lavrič's (2007) contention on civil religion being cross-culturally observable, since predictors are comparable and the major one is the same in all instances.
There are numerous limitations to the validity and generalizability of our findings, pertaining to the insufficient number of independent variables tested, to limited cross-cultural validity, although the variety among the environments is significant.Finally, it has not been proven that the same findings could be found among general population samples, although there are no specific reasons why cross-cultural variation should be impaired by this type of analysis (Cheung & Rensvold, 1998).Probably, within individual societal samples, social stratification would be relevant in general population samples, as authoritarianism should be more prevalent there, i.e. among the lower strata, as has already been established by Lipset (1959).