Theory without data is myth: data without theory is madness.
Since my emancipatory introduction to secular humanism through Free Inquiry about eight years ago, I have immersed myself in the literature, having read authors from A (Bob Avakian’s Away with All Gods!) through Z (Phil Zuckerman’s Atheism and Secularity and Society Without God). My readings have also included works by Richard Dawkins, Daniel C. Dennett, David Eller, Tom Flynn, Sam Harris, Christopher Hitchens, Robert G. Ingersoll, Susan Jacoby, Paul Kurtz, Michael Martin, David Mills, Thomas Paine, Elizabeth Cady Stanton, Peter Singer, and Victor Stenger.
Recently, I focused on applying the perspectives and methods of my discipline, sociology, to matters pertaining to religion and society and to secular humanism.* Although I have the utmost respect, admiration, and appreciation for the work of eminent scholars such as Dawkins, Harris, and Hitchens and their contributions to the literature of secular humanism are important and enlightening, from the perspective of social science many of their claims are not substantiated by findings deriving from research applying appropriate social scientific methodology and, thus, are prone to serious errors, the most common and flagrant being that of overgeneralization. According to these authors, theism in any and all forms is irredeemably malignant and dysfunctional for individuals and societies and must ultimately be eradicated for the betterment of humankind. In The God Delusion, Dawkins states, “As long as we accept the principle that religious faith must be respected simply because it is religious faith, it is hard to withhold respect from the faith of Osama bin Laden and the suicide bombers. The alternative, one so transparent that it should need no urging, is to abandon the principle of automatic respect for religious faith. This is one reason I do everything in my power to warn people against faith itself, not just against so-called ‘extremist’ faith. The teachings of ‘moderate’ religion, though not extremist in themselves, are an open invitation to extremism.”
Really? First, it is not hard to withhold respect from Bin Laden’s version of faith, and to equate his interpretation of Islam with all religious faith is preposterous, an overgeneralization unworthy of the scholar who has authored The Greatest Show on Earth: The Evidence for Evolution. Where is the empirical evidence to support the claim proffered in his last sentence? Furthermore, could not the same be said regarding “moderate atheism,” as compared with the extremes of, say, Stalin’s version of communism/socialism?
The apocalyptic tenor of the critique of religious beliefs and practices of all forms is articulated even more melodramatically by Harris in The End of Faith: “One of the central themes of this book . . . is that religious moderates are themselves the bearers of a terrible dogma: they imagine that the path to peace will be paved once each of us has learned to respect the unjustified beliefs of others. I hope to show that the very ideal of religious tolerance—born of the notion that every human being should be free to believe whatever he wants about God—is one of the principal forces driving us to the abyss.”
Should we rewrite the First Amendment to the U.S. Constitution to accommodate Harris’s own particular form of intolerance? Might there possibly be other factors responsible for societal dysfunction, such as the gross inequities and intolerable social inequalities that are the consequences and concomitants of global capitalism?
And from the late Hitchens’s God Is Not Great: How Religion Poisons Everything: “As I write these words, and as you read them, people of faith are in their different ways planning your and my destruction, and the destruction of all the hard-won human attainments. . . . Religion poisons everything” (emphasis in original). How can we characterize this if not as exaggeration? Overgeneralization? Paranoia? Perhaps the foreword to a Stephen King novel?
On my view, in these examples Dawkins, Harris, and Hitchens have offered us mere rants, reminiscent of political scientist/sociologist Chip Berlet’s discussion of the extreme religious Right. According to Berlet, the following are among the styles utilized by the extreme Right in articulating its ideologies and mobilizing support.
Dualism . . . a form of binary thinking that divides the world into good versus evil, with no middle ground tolerated. There is no acknowledgement of complexity, nuance, or ambiguity in debates, and hostility is expressed toward those who suggest coexistence, toleration, pragmatism, compromise, or mediation. . . .
An Apocalyptic Style. . . . A handful of people have been given a warning so they can make appropriate preparations. Apocalyptic . . . social movements often combine demonization, scapegoating, and conspiracism with a sense that time is running out, so quick action is needed. . . .
Conspiracism . . . is a particular narrative form of scapegoating that frames demonized enemies as part of a vast insidious plot against the common good, while it valorizes the scapegoater as a hero for sounding the alarm. . . .†
Purpose of Study
The purpose of the current study is to illustrate by one select example the potential to apply the sociological perspective and a very basic research methodology in order to increase our understanding of the complex interrelationships between religion and society, particularly the “dysfunctional” effects of the former on the latter. I hope to demonstrate that the errors of overgeneralization evident in the “broad brush” critiques of prominent atheists such as Dawkins, Harris, and Hitchens may be corrected by more careful and detailed examination of the relationships between religiosity and societal dysfunction.
Methodology and Findings
The Human Development Index. Although numerous indicators of societal health or dysfunction have been used to characterize a society’s health, one widely accepted measure is the Human Development Index, or HDI. The HDI was developed as an alternative to economic-based measures of societal development such as the Gross National Product (GNP) or Gross Domestic Product (GDP), so as “to shift the focus of development economics from national accounting to people-centered policies.” The HDI has been included in the United Nations Human Development Report since 1990. As of 2010, the HDI composite measure incorporated three components:
- A long and healthy life: life expectancy at birth
- Access to knowledge: mean years of schooling and expected years of schooling
- A decent standard of living: Gross National Income (GNI) per capita
An initiative of the Social Science Research Council (SSRC), the “American Human Development Project” designed The Modified American Human Development Index, with scores for the fifty U.S. states and Washington, D.C., ranging from 6.30 for Connecticut to 3.85 for West Virginia (see www.ssrc.org and www.measureofamerica.org).
Religiosity. Reliable state-by-state data on religiosity are available from the Pew Forum on Religion and Public Life’s U.S. Religious Landscape Survey, which involved a representative sample of more than 35,000 adults. Among the dimensions of religiosity it examined were denominational affiliation and religious beliefs and practices. Most pertinent to the objectives of the current study were the following general findings, fro
m which I have developed composite measures of high, moderate, and low religiosity for the fifty U.S. states and the District of Columbia. The percentages of the 35,000+ total respondents selecting identified alternatives are presented in parentheses:
- Belief regarding the existence of God or universal spirit. Absolutely certain that God exists (71%); Fairly certain (17%); Not too certain/not at all certain/unsure how certain (4%); Does not believe in God (5%); and Don’t know/refused to answer (3%).
- Belief regarding interpretation of Scripture [Bible or Holy Book]. Word of God, literally true, word for word (33%); Word of God, but not literally true word for word/unsure if literally true (30%); Book written by man, not the word of God (28%); Don’t know/refused to answer/other (9%).
- Belief regarding importance of religion in one’s life. Very important (56%); Somewhat important (26%); Not too important/not at all important (16%); Don’t know/refused (1%).
- Practice of frequency of attendance at religious services. At least once a week (39%); Once or twice monthly/few times a year (33%); Seldom or never (27%); Don’t know/refused (1%).
- Practice of frequency of prayer. At least once a day (58%); Once a week/a few times a week (17%); A few times a month (6%); Seldom or never (18%); Don’t know/refused (2%).
Disregarding the “Don’t know/refused to answer” selections, the first response alternative for each of the above survey question topics were jointly considered as “high religiosity” responses: Absolutely certain that God exists; Word of God, literally true, word for word; Very important; At least once a week; At least once a day. “Low religiosity” was defined by the last response alternative(s) to each question: Not too certain/not at all certain/unsure how certain or Does not believe in God; Book written by man, not the word of God; Not too important/not at all important; Seldom or never; and Seldom or never. “Moderate religiosity” included the middle or in-between responses: Fairly certain; Word of God, but not literally true word for word/unsure if literally true; Somewhat important; Once or twice monthly/few times a year; Once a week/a few times a week or A few times a month.
Hypotheses 1–3. If we accept the generalizations proffered by Dawkins, Harris, and Hitchens regarding the dysfunctionality or malignant effect of religion on society, then it is reasonable to hypothesize the following relationships between religiosity and the American Human Development Index:
- There is a negative association (correlation) between high religiosity and the American HDI; namely, the higher the percentages of state populations self-reporting high religiosity of religious beliefs and practices the lower the scores on the American HDI.
- There is a negative association (correlation) between moderate religiosity and the American HDI; the higher the percentages of state populations self-reporting moderate religiosity of religious beliefs and practices the lower the scores on the American HDI, but this relationship is expected to be somewhat less strong than that predicted in hypothesis 1.
- There is a positive association (correlation) between low religiosity and the American HDI; namely, the higher the percentages of state populations self-reporting low religiosity of religious beliefs and practices the higher the score on the American HDI.
My analyses of the data produced evidence supporting hypotheses 1 and 3 but not hypothesis 2, as shown in figure 1.
The findings of this analysis do not support hypothesis 2. In fact, whereas a negative correlation was predicted, a positive one is observed. The higher the moderate religiosity score, the higher the score on the American HDI. Having found support for hypotheses 1 and 3 but not for hypothesis 2, I next examined another measure of religiosity, namely, denominational affiliation.
In addition to querying survey respondents about their religious beliefs and practices, the Pew Survey asked about denominational affiliation. Of the sample of 35,000+, 84 percent identified with one of the following major religious traditions: Evangelical Protestant (26%); Mainline Protestant (18%); Catholic (24%); and Unaffiliated (16%).
No “tradition” other than the four above exceeded 10 percent. The next largest group self-identified with Historically Black Churches at 7 percent, with the Jewish tradition at only 1.7 percent; while Buddhists comprised but .7 percent and Muslims and Hindus were .6 percent and .4 percent, respectively. Among the Unaffiliated were atheists at 1.6 percent, agnostics at 2.4 percent, and “nothing in particular” at 12.1 percent.
Hypotheses 4–7. Guided again by the writings of Dawkins, Harris, and Hitchens, the following hypotheses are suggested:
- The higher the percentage of state populations self-identifying as Evangelical Protestant, the lower the score on the American HDI; a negative correlation is predicted.
- The higher the percentages of state populations self-identifying as Mainline Protestant, the lower the score on the American HDI. Again a negative correlation is predicted.
- The higher the percentages of state populations self-identifying as Catholic, the lower the score on the American HDI, with a negative correlation again predicted.
- The higher the percentages of state populations self-identifying as Unaffiliated, the higher the score on the American HDI. A positive correlation is predicted.
My analyses yielded findings supportive of hypotheses 4 and 7 but not hypotheses 5 and 6, as shown in the following figures.
Summary and Conclusion
In summary, reviewing empirical evidence bearing upon the truth or falsity of seven hypotheses deriving from the writings of atheists Dawkins, Harris, and Hitchens, we have found evidence in support of four but in refutation of three others. Although it certainly may be asserted that “Four out of seven ain’t bad,” my analyses support my original contention that Dawkins, Harris, and Hitchens have committed the egregious methodological error of overgeneralization. While they are apparently correct in characterizing as malignant or dysfunctional the effects of high religiosity and self-identification with the evangelical Protestant tradition on the American HDI, and also in hypothesizing positive or functional effects of low religiosity and self-identification as unaffiliated on the HDI, their errors of overgeneralization are evidenced by the following three findings that refute hypotheses 2, 5, and 6:
- Moderate levels of religiosity are associated with high scores on the American HDI, so that the effect is positive or functional, not malignant or dysfunctional.
- Higher percentages of state populations self-identifying as Mainline Protestants is unrelated to scores on the HDI. Its effect is neither functional nor dysfunctional but rather benign.
- Higher percentages of state populations self-identifying as Catholic is positively associated with scores of the HDI, so its effect is positive or functional, neither malignant nor dysfunctional.
It is hoped and anticipated that further social-scientific research will help to illuminate some of the underlying reasons for these findings. My own related analyses, in which I introduce the social inequality variables of income, educational attainment, and race, reveal tha
t the former two account for or “explain” 93 percent of the variation in HDI scores, while introducing the religiosity variable increases the amount of explained variation by only an additional 2 percent.
The direct, confrontational, no-holds-barred assault upon organized religion by contemporary atheists is misguided—a terrible waste of time, precious talent, intelligence, and energy and has little chance to help bring about the realization of a life for humankind guided by humanist ideals of the Enlightenment: reason, science, compassion and the several other normative standards articulated in “The Affirmations of Humanism: A Statement of Principles” and in The Humanist Manifesto.
While we should continue to pursue a policy of zero tolerance for pedophile priests and terrorists inspired by religious extremism, we should simultaneously advocate for more aggressive measures to decrease currently intolerable levels of income, wealth, and educational inequality. If we expect substantial proportions of our population to live lives guided by reason, science, and compassion, our national budgetary priorities must change in a manner that more closely approximates those of the more egalitarian and less dysfunctional social-democratic European nation-states.
I am grateful for the substantive reviews and editorial suggestions provided by confidantes Nancy D., Peter D., John D., and Joan H., and I am indebted, especially, to William H. Miller, professor emeritus at Iowa State University, for access to his OpenStat multivariate statistical package. Finally, I thank Tom Flynn for his infinite patience and especially for his assistance, through insistence that my findings be described in a fashion more suitable to an intelligent and informed general readership and less in the format of a specialized quantitatively oriented research journal, distinguished by detailed and admittedly laborious statistical analyses, a style to which I am naturally predisposed. A substantially more detailed version of this paper is available from the author; send requests in care of this magazine.
* See, for example, Delamontagne, R. Georges. 2010. “High Reli- giosity and Societal Dysfunction in the United States during the First Decade of the Twenty-First Century.” Evolutionary Psychology 8(4): 617–57 (www.epjournal.net/filestore/EP08617657.pdf).
† “Mapping the Political Right: Gender and Race Oppression in Right- Wing Movements” in Abby L. Ferber, ed., Home Grown Hate: Gender and Organized Racism (New York: Routledge, 2004).