SUBJECTIVE SOURCES OF OPPOSITION TO ABORTION

A number of studies (Adamczyk Reference Adamczyk2013; Fisher Reference Fisher2011; Jelen and Wilcox Reference Jelen and Wilcox2003; Reference Jelen and Wilcox2005) have shown that religiosity is associated with disapproval of abortion. There are important differences and similarities across religious traditions in both the extent of opposition to legal abortion and to the rationales underlying such support or opposition. Both the “behaving” and “belonging” aspects of religion (Green Reference Green2010; Ben-Nun Bloom and Arikan Reference Ben-Nun Bloom and Arikan2012) appear to have independent effects on approval or disapproval of the practice of abortion. The abortion issue poses a number of ethical questions. Indeed, Staggenborg (Reference Staggenborg1994) has characterized the issue as a “condensational symbol,” which invokes several very basic attitudes toward human life, sexuality, and gender roles.

The most common antiabortion rationale represents the label favored by most opponents of legal abortion in the United States: Pro-life. To the extent that most religious traditions can be characterized as having “a” position on abortion, the most frequently invoked rationale is respect for the human life represented by the embryo, and later, the fetus. For example, the contemporary Roman Catholic position holds that abortion involves the taking of a human life, since “ensoulment” begins at the moment of conception (Cook et al. Reference Cook, Adell, Jelen and Wilcox1992; Connery Reference Connery1977; Noonan Reference Noonan and Noonan1970). Further, this position is based on the Catholic doctrine of natural law. Conversely, the opposition to legal abortion among some evangelical Protestants is based on a biblical heuristic.

Although Protestant leaders have not neglected the importance of concern for human life in the abortion controversy (see especially Schaeffer Reference Schaeffer2005), the humanity of the fetus has been but one consideration in the antiabortion rhetoric of some Protestants (at least in the United States). Concerns for sexual morality and appropriate gender roles also permeate the abortion discourse of many Protestants (particularly those with evangelical orientations). Some prominent evangelicals have suggested that the demand for legal abortion is an inevitable and unfortunate consequence of the “sexual revolution” which was thought to characterize the 1960s and 1970s (Falwell Reference Falwell1981; see Williams Reference Williams2010, for a review). Reciprocally, legal abortion was often characterized as a source, as well as a consequence, of sexual promiscuity. Jelen (Reference Jelen1984) showed that, in the period immediately following the Roe v. Wade decision in the United States, doctrinally conservative Protestants were more likely to oppose legal abortion for reasons of sexual morality, rather than basing their antiabortion positions on respect for human life. However, as time passed, and the abortion issue entered the realm of political controversy, the antiabortion rationales of US Catholics and evangelicals largely converged (Jelen Reference Jelen1988; see also Dillon Reference Dillon1995).

Identifying the subjective sources of Muslim attitudes is somewhat more complex. Some Muslim authorities have suggested that ensoulment does not occur until much later in pregnancy, but that the potential ensoulment of fetuses at earlier stages of development and the role of children and women in organic conceptions of the family must also be considered (see, for example, Abdul Hussain Reference Abdul Hussain2005; Aramesch Reference Aramesh2007; Sekaleshfar Reference Sekaleshfar2008; Alamri Reference Alamri2011; and Hedayat, Shooshtarizadeh, and Raza Reference Hedayat, Shooshtarizadeh and Raza2006). That is, among many Muslim theologians, the ontological status of the fetus is a vital component in the construction of a rationale for opposing abortion, but is by no means the only one. Concerns with the structure of family life and the role of women in the family, as well as the importance of chastity, are all important considerations in evaluating the morality of abortion, and providing rationales for the legal restriction of the practice.

A similarly complex set of concerns permeates the Hindu perspective on abortion. Indeed, the extent of decentralization in the theology and practice of Hinduism makes a characterization of “a” Hindu position problematic. Nevertheless, analysts of Hinduism have suggested that abortion is inappropriate for both moral and social reasons. The fetus does, in most accounts, possess attributes (unrealized) of a life distinct from the mother (Jain Reference Jain and Maguire2003). However, abortion is also regarded by some as a disruption of family continuity and generational replacement (Damian Reference Damien2010; Lipner Reference Lipner, Coward, Lipner and Young1989).

By contrast, both Eastern Orthodox and Buddhist traditions place primary emphasis on the humanity of the fetus in formulating reasons for opposing legal abortion. Indeed, some analysts have argued that, in Buddhist theology, there is no moral distinction between an unborn fetus and a living person (Florida Reference Florida1991; Damian Reference Damien2010). However, a few analysts have suggested that abortion is a violation of appropriate ethical and social roles for women, as the practice indicates selfishness on the woman seeking to terminate a pregnancy (Harvey Reference Harvey2000; Florida Reference Florida1991). The Eastern Orthodoxy tradition is similarly decentralized, owing to the organization of Eastern Orthodoxy into national churches. However, there appears to be a strong emphasis on the “human life” aspect of abortion (see Garvey Reference Garvey1996). Indeed, the Orthodox Church in the United States submitted an amicus curiae brief in the case of Webster v. Planned Parenthood (1988) in which the humanity of the fetus was the sole moral issue raised (Orthodoxy Christian Information Center 1988).

Thus, for spokespersons of a number of religious traditions, abortion is indeed “a matter of life and death” (Baker et al. Reference Baker, Epstein and Forth1981). However, some religious leaders have emphasized other considerations, such as the maintenance of traditional standards of sexual morality or the importance of specialized gender roles for family and social life.

Religious traditions vary not only in doctrine, but in organizational structure and scope as well. The obvious limiting case here is Roman Catholicism, in which an international, hierarchical organization permits the existence of an “official” position and rationale on the abortion issue. Analogously, although Islam does not exhibit the same organizational structure of Catholicism, both faiths have strong missionary traditions (Jelen and Tamadonfar Reference Jelen and Tamadonfar2011), which render cross-national socialization possible. In contrast, Eastern Orthodoxy is organized around national churches, while the three other religious traditions considered here (Protestantism, Hinduism, and Buddhism) are stubbornly decentralized.

Although abortion attitudes, and abortion policies, differ across diverse global regions, there are important centripetal forces which have made abortion policies the subject of international politics as well. For example, the International Conference on Population and Development in Cairo in 1994 presented an excellent case study of the opportunities for, and limitations of, Roman Catholic/Muslim cooperation with respect to issues of reproductive freedom and population control (Bowen Reference Bowen1997; Cowell Reference Cowall1994; Waldman Reference Waldman1994; Al-Sharq al-Awzat 1994). Abortion politics have been a source of contention among member states of the European Union, with national policies having been modified by the acts of the European Union (Lyall Reference Lyall2010; New York Times 2007), and in Latin America (Economist 2013). Such internationally focused attention to abortion politics may increase the effect of transnational religious traditions.

Although there are obvious differences in the extent to which citizens of various nations approve or disapprove of abortion, this study is focused on the rationales by which citizens and adherents of different religious traditions may reach or justify their abortion stances. The specific purpose of this study is to compare the subjective bases of attitudes toward abortion across a variety of religious-defined groups. The effects of three possible bases for support for or opposition to legal abortion are considered. First, the relationship between a general respect for human life and abortion attitudes will be considered.

Second, in this study, I will examine the effects of attitudes toward sexual morality on attitudes toward abortion. For many activist opponents of legal abortion, a prohibition against legal abortion may increase the risk of non-marital sex (Luker Reference Luker1985; Cook et al. Reference Cook, Adell, Jelen and Wilcox1992; Falwell Reference Falwell1981) and thus may be an effective means of discouraging sex outside of marriage. It is hypothesized that respondents who hold more traditional attitudes toward sexual morality will hold less permissive attitudes toward abortion.

Finally, I hypothesize that respondents who hold egalitarian attitudes toward the appropriate social and economic roles of women will be more likely to regard abortion as justified. Reproductive freedom, and the resultant control over one's fertility, is often regarded as important prerequisites to full female equality in politics and in business (Luker Reference Luker1985; Cook et al. Reference Cook, Adell, Jelen and Wilcox1992). One might anticipate that respondents who hold more traditional expectations about gender roles will be less likely to approve of abortion. As will be described below, a distinction is made between “public” and “private” feminism: In the case of the former, attitudes toward paid labor versus homemaking are measured. With respect to the latter, the independent variable under consideration deals with attitudes toward the role of woman as mother, in relationship to her children.

My specific task in this study is to compare the effects of these three sets of attitudinal variables across six religious traditions. I anticipate that considerations of human life, sexual morality, and gender roles will all exhibit significant relationships with abortion attitudes for all groups. However, based on other scholarly accounts of the theological positions taken by religious elites, I would also hypothesize that considerations of other than human life would be most prominent among Muslims, Hindus, and Protestants, in which considerations of family life and sexuality may be more explicit. Because of the transnational character of the traditions, I also hypothesize that relationships among predictor variables will be stronger for Roman Catholics and Muslims.

DATA AND METHOD

Data for this study were taken from the World Values Study for 1999–2007.Footnote ¹ The main analytic strategy is to compare abortion attitudes, and multivariate models explaining such attitudes, for members of six religiously-defined groups: Roman Catholics, Protestants,Footnote ² Muslims,Footnote ³ Eastern Orthodox, Hindus, and Buddhists.Footnote ⁴ Of course, while all traditions would contain some theological diversity, “Protestantism” is an unusually theologically diverse category, and the results for this group should thus be interpreted with caution. The estimation strategy is to assess the relative importance of the three attitudinal sources of abortion attitudes across religious traditions. Does support for, opposition to, abortion have similar meanings across different religious affiliations, or are there important differences in the importance of distinct sources of abortion attitudes?

The main dependent variable is a 10 point scale, in which respondents are asked to place themselves on a continuum, on which 1 represents a belief that abortion is “never justified,” and 10 indicates a stance in which abortion is “always justified.” This is a 10 point scale, on which respondents are ask to place themselves. No cues as to the meaning of intermediate responses were provided.

In this study, the most important independent variables are the subjective bases on which respondents may choose to support or oppose legal abortion. One such variable is, as noted, a measure of respect for human life. Such a concept poses a difficult and intriguing measurement problem. Following earlier work (Jelen Reference Jelen1984; Reference Jelen1988; Cook et al. Reference Cook, Adell, Jelen and Wilcox1992), respect for human life is operationalized as a 10 point scale (in the same format as the abortion item) measuring respondent attitudes toward euthanasia. A score of 1 indicates that the respondent believes that euthanasia is “never justified,” and 10 suggest that the respondent believes that euthanasia is “always justified.” This measure is both indirect and controversial. It is difficult to imagine a survey item which directly taps a general respect for human life and more specific applications of such respect (such as Cardinal Bernardin's [Reference Bernardin1988] “consistent ethic of life” has some difficulties. Moreover, at least in the United States, previous research has shown that very few respondents (even devout Catholics) take consistently “pro-life” positions on issues such as euthanasia, military spending, capital punishment, and abortion. Given the apparent empirical independence of these attitudes (see especially Jelen Reference Jelen1990), the choice of attitudes toward euthanasia as a surrogate for a more general respect for human life seems defensible. In the cases of abortion and euthanasia, the “person” whose life is ostensibly being taken is presumably innocent, and not deserving of her/his fate. This feature of euthanasia stands in contrast to other “life attitudes,” in which possible “victims” include convicted murderers or enemy soldiers, whose demise may be considered justified. This is not to suggest that respondents who provide high scores on this item do not respect life, but that such respondents may have a more qualified or nuanced view of the circumstances under which human life should be preserved. Moreover, as the findings reported below suggest, the euthanasia item exhibits a substantial level of construct validity.Footnote ⁵

Another subjective predictor of abortion attitudes in this study is the respondent's attitude toward sexual morality. Here, this concept is operationalized as the mean of two “justification” scales, similar to the abortion and euthanasia items, tapping attitudes toward homosexuality and prostitution. Again, this measure is controversial, but these items appear to tap a more general attitude toward acceptance of sex outside of monogamous heterosexual marriage. Further, these items are empirically related quite strongly.Footnote ⁶

Finally, the effects of two items measuring attitudes toward appropriate gender roles are considered. A measure of “public feminism” is a Likert scale which reads “Being a housewife is just as fulfilling as working for pay.” An item which taps attitudes toward “private feminism” (see Cook et al. Reference Cook, Adell, Jelen and Wilcox1992 for a discussion of the distinction between public and private feminism) reads: “A working mother can establish just as warm and secure a relationship with her children as a mother who does not work.”Footnote ⁷ Again, these attitudes are conceptually and empirically distinct, and the effects of these variables are considered separately.Footnote ⁸

The multivariate models which are the focus of this study include two controls for religiosity: attendance at religious servicesFootnote ⁹ and frequency of private prayer.Footnote ¹⁰ The models also include controls for respondent gender, age, and education.Footnote ¹¹ Because the dependent variable is nearly continuous, the models presented below are estimated using Ordinary Least Squares (OLS).

THE SUBJECTIVE BASES OF ABORTION ATTITUDES: COMPARISONS ACROSS RELIGIOUS TRADITIONS

Differences in abortion attitudes across religious traditions are shown in Figure 1. As the data presented in this graph indicate, there are substantial variations in the extent to which members of different religiously-define groups consider abortion justified. Muslims are the least likely to regard abortion as justified, followed by Protestants, Hindus, Roman Catholics, Buddhists, and Eastern Orthodox respondents.Footnote ¹²

Figure 1. (Color online) Mean abortion attitudes by religious tradition.

The main findings of this study are reported in Table 1. What is most noteworthy about this table is that the statistical models which account for variation in abortion attitudes are quite similar for all six groups. In all models, the most important predictor is the index of sexual morality. In all seven OLS models, the effects of this measure are strong, significant, and in the expected direction. Similarly, the effects of the euthanasia variable are also quite robust. However, the effects of attitudes toward sexual morality are about twice as strong as the effects of the euthanasia item for each comparison group. Indeed, the most striking feature of Table 1 is the uniformity of effects of sexual traditionalism and respect for life across very different religious traditions.Footnote ¹³

Table 1. Multivariate models of abortion attitudes by religious tradition

*significant at 0.05.

**significant at 0.01.

***significant at 0.001.

Source: World Values Survey, 1999–2007.

By contrast, the effects of the items tapping aspects of gender role attitudes are much weaker and less consistent. The measure of public feminism (Housewife fulfilling) has significant effects on abortion attitudes among all three Christian groups (Catholic, Protestant, and Eastern Orthodox) and among Hindus. The effects of the “working mom” item (tapping private feminism) are significant only for Eastern Orthodox respondents. Moreover, this somewhat surprising “non-finding” does not appear to be dependent on variations in model specification. The relationships between the gender role items and the measures of sexual morality and euthanasia are quite weak (typically, r < 0.10), and are often of marginal statistical significance. It is perhaps of interest to note that the relationships between both gender items and the abortion scale are strongest for adherents of the Eastern Orthodox tradition.

These findings are somewhat surprising. One possible explanation for these results is that many respondents may not have clearly formed or defined attitudes about gender roles (Converse Reference Converse and Apter1964; Zaller Reference Zaller1992). To provide a partial test of this hypothesis, a measure of attitude constraint for gender roles was computed by taking the individual standard deviation of three items measuring attitudes toward being a housewife, the relative importance of university education for boys and girls, and respondent beliefs about the superiority of men as business managers. The expectation here is that respondents who take consistently egalitarian or consistently traditionalist positions on these issues have more consistent (and perhaps more sophisticated) gender role attitudes. Interaction terms between this constraint measure and the Housewife Fulfilling and Working Mom measures were computed, and included in alternative specifications of the multivariate models in Table 1 (data not shown). The substitution of gender role measures which tap both direction and consistency of attitudes toward the respective roles of men and women make no discernible difference in the results reported here. Similarly, computing interaction terms between each gender role item and education makes no discernable difference in the direction or magnitude of any of the relationships reported here. While these tests are certainly not definitive, the data suggest that the limited relevance of gender role attitudes to attitudes toward abortion may not be based on differences in political sophistication.Footnote ¹⁴

For all groups except Buddhists, frequency of private prayer is significantly related to respondent attitudes toward the acceptability of abortion, while attendance at religious services exhibits a statistically significant relationship with the dependent variable among Roman Catholics, Protestants, and Muslims. In all cases, the effects of prayer are stronger than those associated with public religious observance. This finding suggests that religiosity exerts influence on adherents indirectly, rather than via direct socialization from clergy or coreligionists.

Finally, the effects of demographic variables are rather inconsistent. Age significantly predicts abortion attitudes among Protestants, Buddhists, Muslims, and the religiously unaffiliated. Women are significantly more “pro-choice” only among Protestants and the Eastern Orthodox, while higher levels of education are associated with lower disapproval of abortion for all groups except Catholics and Muslims.

The general pattern, in which sexual morality and attitudes toward euthanasia are the strongest predictors of abortion attitudes, with sexual traditionalism having the stronger effect, seems remarkably consistent across national settings in which the political roles and direct effects of religion are quite different. To illustrate, Table 2 compares the effects of these two variables in five different countries: Poland (the Catholic nation with the lowest mean approval on the abortion scale), Switzerland (the most “pro-choice” Catholic country), Columbia (the most conservative Catholic country outside of Europe), Malaysia (the most approving Muslim majority country), and Jordan (the least approving Muslim majority country). In these five countries, the general pattern holds up rather neatly, with both the effects of both variables attaining statistical significance, and the sexual morality variable having consistently stronger effects than the euthanasia item.

Table 2. Effects of attitudes toward euthanasia and sexual morality on attitudes toward abortion, selected countries

**significant at 0.01.

***significant at 0.001.

Entries are unstandardized regression coefficients. Models include controls for two feminism items (Housewife Fulfilling and Working Mom), frequency of private prayer, and frequency of attendance at religious services, gender, age, and education.

Given the consistent effects of attitudes toward human life and sexual morality, and the weak, inconsistent effects of other variables, how can variations across religiously-defined traditions be explained? In general, the answer appears to lie in different marginal distributions of the main independent variables. Table 1 contains the mean values for the independent variables for adherents of each religious tradition considered here. Comparing means across religious traditions is fairly revealing. The clearest finding is one of Muslim exceptionalism. Muslims, who constitute the group most disapproving of abortion, are also least likely to approve of euthanasia, and are less likely than all other groups to regard homosexuality and prostitution as justified.

Second, the effects of the euthanasia item are rather consistent across religious traditions. Although the mean value for Buddhists is a relatively high 4.67, and for Muslims a very low 2.14, the range of mean values on the measure of respect for human life across Christian traditions, and for Hindus, is very limited. By contrast, there exists a substantial variation across religious traditions with respect to the sexual morality index. Thus, although the relationships between the euthanasia and sexual morality items do not vary very much across traditions, variation in the marginal distributions of attitudes toward non-marital sex appear to make a greater contribution to inter-religious differences on the abortion scale, especially among different Christian traditions.

Third, it is also noteworthy that Roman Catholics do not exhibit distinctive patterns. The effects of the euthanasia and sexual morality scales on abortion are quite similar for Catholics as for other religious groups. In terms of the marginal distributions of the main independent variables, Catholics approach the grand median on the euthanasia measure (despite the Church's emphasis on the sanctity of human life), and are the most permissive group on the sexual morality scale. These results should put to rest any notion that the Catholic Church (at the level of the laity) is in any way distinctive on either the abortion issue itself, or on the most important attitudinal correlates of abortion attitudes.Footnote ¹⁵

More generally, the fact that differences across religious traditions can be moved to the marginal distributions of independent variables suggests that these models are capturing comparable phenomena across traditions (Converse and Dupeux Reference Converse, Dupeux, Campbell, Converse, Miller and Stokes1966; Converse Reference Converse1969). The overall impression is one of impressive similarity. Attitudes toward sex dominate the multivariate models explaining variation in abortion attitudes, followed by attitudes toward the sanctity of human life. By contrast, the effects of attitudes toward the appropriate economic and familial roles of women are much weaker and less consistent within and across religious traditions.