Unraveling Religious Worldviews: The Relationship between Images of God and Political Ideology in a Cross-Cultural Analysis

Abstract

Not only do few studies address the issue of how religious belief relates to political ideology, but little attempt has also been made to analyze this relationship from a comparative perspective. Using data from the International Social Survey Program, we examine how images of God, as measured by God's perceived level of engagement and authority, relate to political ideology in seven Western industrial and postindustrial societies. We find that variation in images of God has no effect on whether individuals are politically liberal or conservative in five of seven countries. Nonetheless, beliefs about God are strongly related to abortion and sexual morality attitudes in every country, but only sporadically related to ideas about social and economic justice. In the end, we argue that theological beliefs tend to be unrelated to a general measure of political ideology, not because religious beliefs are politically unimportant in these societies, but rather because religious perspectives are rarely fully liberal or conservative in their political orientation. In addition, we find that Americans hold unique views of God in comparison to other countries in our sample and that the American tendency to view God as more active and authoritative affects policy attitudes in ways contrary to the effects of church attendance.

NOTES

Notes

¹ Contemporary explorations into the social significance of religious belief tend to examine how religious, social, moral, and political beliefs comingle (Hunter 1991; Guinness 1993; Wuthnow 1996; Evans 1997; Davis and Robinson 1999; Baker 2005). However, this same literature often uses measures or concepts that mix moral, social, theological, and political worldviews, making it difficult to determine the unique effects of each. For example, Wayne Baker's (2005:9) study of value differences in the American public equates an individual's belief in a transcendent authority with moral absolutism. For Baker, moral absolutists look toward the “transcendental sphere” for authority—basing their decisions about right or wrong, social justice, and public policy on God or the Scriptures. Moral relativists, on the other hand, find their basis of authority in the social sphere—basing decisions upon worldly concerns. While these are useful conceptual categories, they do not allow us to discover whether certain religious beliefs lead to absolutism or relativism, because the connection is assumed to exist from the outset. Potentially, one could be a moral relativist while believing in a powerful and authoritative God or, in contrast, one could be an atheist and a moral absolutist—a Marxist-Leninist, for instance, would fit this category.

² Other national surveys find majority belief in God in the United States. For example, according to the General Social Survey 2000, 8–4 percent of Americans believe in God or in a higher power (percentage combines belief in God without doubt, belief in God with some doubts, belief in God sometimes, and belief in a higher power).

³ George Lakoff's (2002) explanation of moral and political differences in the United States provides us with an additional theoretical framework to understand Greeley's findings. Conceptually, Lakoff (2002:163) argues that politically conservative attitudes evoke a “Strict Father” morality, which stresses self-discipline, responsibility, and self-reliance. As such, political conservatives use Strict Father morality to justify policies aimed at increasing individual responsibility, such as harsher criminal punishments and reductions in social welfare. For Lakoff, political liberals express a “Nurturant Parent” morality, which stresses fairness and equality, and is used to justify policies such as increases in social welfare and equal rights legislation. While Lakoff does not make this connection, his moral categories correspond directly to Greeley's distinctions between different images of God. Therefore, we might expect that individuals who perceive God as less maternal will tend to support Strict Father or politically conservative policies. And those who view God as more loving will favor “Nurturant Parent” or politically liberal policies.

⁴ In his study of American values, Wayne Baker (2005) speaks of a moral continuum between absolutist and relativist worldviews. We find his conceptualization highly useful, measurable, and nonjudgmental, but, unlike Baker, we detach the idea of moral absolutism from beliefs about moral authority. In other words, a moral absolutist views a behavior as “always wrong,” while a moral relativist weighs circumstances or outcomes—regardless of his or her beliefs about transcendent authority. With this slightly revised conception of moral differences, we can directly test how absolutism and relativism relate to theological worldviews, rather than simply assuming that the existence of such a relationship in our measures.

⁵ Our cross-sectional analyses have some clear limitations. It is impossible to determine the causal link between the variables in question. Therefore, we cannot address the topic of how images of God are developed or whether they logically lead individuals to hold certain attitudes or act in specific ways. Above, we have hypothesized reasons why we expect images of God to be related to attitudes and behaviors and these explanations have necessarily incorporated causal assumptions. The following analyses test whether our expectations are supported, but do not establish the causality of these relationships.

⁶ With the exception of one item that was not available in the 1998 ISSP and the Biblical literalism item, our measure of God's level of engagement is quite similar to Davis and Robinson's (1999) measure of “religious orthodoxy.” While we find the same items useful, we disagree with Davis and Robinson (1999) regarding what they represent. All of their items refer directly to beliefs about God. A measure of orthodoxy might also involve beliefs about the importance of religious practices and ritual and the quality of those rituals. One can imagine a religious traditionalist who is orthodox in one sense, strictly adhering to ritual practices, yet less traditional in theological beliefs. For instance, many Roman Catholics believe strongly in the power of religious sacraments, yet also espouse less orthodox beliefs about God's personality than American Protestants (see Greeley 1989:100). Rather than an incomplete measure of orthodoxy, we believe that, together, these items are more properly understood as measuring beliefs about God's engagement in the world.

⁷ We do not claim that these items perfectly capture the distinction between an authoritarian and nonauthoritarian God. After all, people may have different conceptions of what a “King” and “Master” mean. But clearly, a respondent who sees God as a master, king, judge, and father has a different image in mind than someone who views God as spouse, friend, mother, and lover.

⁸ The variables contrasting master/spouse and judge/lover were reverse-coded so that, on all four variables, higher scores related to more judgmental images of God (father, king, judge, master).

⁹ We feel that there is a legitimate reason for why our measure constitutes a conceptually justifiable scale despite its less than ideal alpha score. Basically, we find that the manner in questions were asked naturally produced a lower than ideal alpha score. Specifically, each of the component variables asks respondents to place their personal image of God between two different descriptors on a scale of 1 to 7. For example, an item asks respondents to place their personal image of God anywhere from mother (1) to father (7). Although the variable ranges from 1 to 7, the vast majority of respondents pick one of three categories—either extreme (1, 7), or the middle (4). Very few people claim that God is somewhat of a father (e.g., a 5 or a 6) or somewhat of a mother—although many appear to not be able to decide between the two. A better way to ask these questions would be to allow individuals to determine how well a single descriptor applies to God, rather than force them to locate themselves between two descriptors. The latter method assumes that the two descriptors are in fact opposites. Data from the Baylor Religion Survey 2005 confirms that if you give respondents the choice of one characteristic per question, the more judgmental characteristics group together statistically. And if we dichotomize the responses from the God variables on the ISSP, we find the alpha for Authoritative God is .74.

¹⁰ In analyzing the meaning of “left/right” labels, Conover and Feldman (1981:643) conclude that “liberals seem to favor change and progress at the expense of government involvement; conservatives, on the other hand, wish to preserve traditional arrangements, particularly those threatened by government involvement.”

¹¹ In all of our analyses, we performed stepwise inclusion of variables, and we also added Biblical literalism as a control. These changes to the models did not produce any results that dramatically differed from those presented.

¹² We replicated the analyses in this article by removing any individuals who expressly indicated that they had no belief in God; our findings do not change significantly. Tables can be provided upon request.

¹³ In other words, we are not looking at a moral continuum between libertarian and moral absolutist. Instead, the data show that individuals tend to agree on the morality of behavior, but differ in how strongly they are willing to condemn it in others.

¹⁴ Theological individualism flows out of a belief that God is watching and taking score of individual choices. As Stark and Glock (1969:80) point out: “Underlying traditional Christian thought is an image of man as a free actor, as essentially unfettered by social circumstances, free to choose and thus free to effect his own salvation. This free-will conception of man has been central to the doctrines of sin and salvation. For only if man is totally free does it seem just to hold him responsible for his acts.”

¹⁵ Certain religious traditions such as Calvinism espouse the doctrine of predetermination. But most traditional Christian theologies view the individual as free to make her own choices.

¹⁶ We would expect even greater variation in the Authoritative God scores when including countries that are not predominantly Christian. Particularly, Asian religions are vastly differently in how they describe supernatural concepts in nonmonotheistic terms.

¹⁷ Authoritative God remains the sum of four variables that ask respondents to characterize God between the opposites mother and father (MAPA), master and spouse (MASTERSP), judge and lover (JUDGELUV) and friend and king (FRNDKING) (Alpha = .61). The items were converted to z-scores before summation. Active God remains an index created using two items (Alpha = .80). We standardize and sum items that ask respondents if life is only meaningful because God exists (GODMEANS) and if there is a God who concerns himself with every human being personally (THEISM).

¹⁸ Original names of GSS variables used/collapsed to create controls: gender (SEX); age (AGE); income (INCOME98); college education (EDUC); Catholic (RELIG); Protestant (RELIG); church attendance (ATTEND).

¹⁹ Gun regulations in the United States vary dramatically from state to state, with some states not requiring certain forms of licensing and registration. All other countries in our sample have stiff licensing, registration, and purchasing gun laws.

²⁰ Capital punishment is legal in the United States for common criminal offenses. Every other country in our sample has banned capital punishment, except for Latvia, which reserves capital punishment for exceptional crimes.