Readers Forum

Victor J. Stenger on the Epidemiology of Religion

By Victor J. Stenger

Epidemiological studies, in general, are notoriously difficult to interpret reliably because of so-called "confounding factors." A given study may indicate a correlation between an illness and some factor, but this does not necessarily prove that the factor is the cause (or cure) of the illness. In an amusing example given by psychologist Richard P. Sloan, a study might find that lung cancer is more prevalent among people who carry around matches in their pockets. It would not follow that matches cause lung cancer.

Recently, we have seen an upsurge in epidemiological studies that attempt to discern a connection between religion and health. In a 1998 review published in Archives of Family Medicine, Dale Matthews and five collaborators examined a large number of papers and concluded that "a large proportion of published empirical data suggests that religious commitment may play a beneficial role in preventing mental and physical illness, improving how people cope with mental and physical illness, and facilitating recovery from illness."

One of the co-authors of this study, which was partially fund by the Templeton Foundation, was Harold G. Koenig, director of the Center for the Study of Religion/Spirituality and Health located at Duke University. Koenig is also the author of a popular book, The Healing Power of Faith, and has appeared on national TV promoting the power of religion to heal. He writes that people have had their lives changed by "faith in a transcendent Spirit with a power to heal emotionally, mentally, and physically."

A review of the literature has been provided by R. P. Sloan, E. Bagiella, and T. Powell in the British medical journal Lancet. They concluded that the published work lacks consistency and is not based on sufficiently large samples of data. Linda Gundersen has also reviewed the subject in Annals of Internal Medicine and finds that the conclusions of several of the studies are doubtful.

By far, the greater number of these investigations are of the epidemiological variety where various religious behaviors are examined for their possible association with health.

Churchgoing seems to produce the greatest correlation, but none of these reports have the statistical significance that is normally required for the publication of extraordinary claims in the hard sciences. Furthermore, the studies have been unable to adequately establish a clean causal connection between religion and health.

As Sloan and his co-authors explain, when confounding variables are considered, these can explain most, if not all, of the effects observed. For example, a 1972 study that is often cited as evidence for a positive association between church attendance and health was later found, by its own authors, to be due to failure to take into account people with reduced mobility. People in poorer health were simply less likely to go to church.

A similar question can be raised with the frequently heard claims of a connection between positive thinking and health. Perhaps such a connection exists, but is it positive thinking that causes good health or good health that causes positive thinking?

A much ballyhooed recent report has also been misinterpreted by those who did not read carefully, or chose to ignore, what the investigators really said. The autobiographies of 180 Catholic nuns were studied, and it was found that those who reported positive emotions in childhood lived longer than those who did not, with chance odds of greater than 1,000 to one. This did not mean, as the media reported, that nuns live longer than the rest of us because of supernatural intervention ó just that positive-thinking nuns may live longer than negative-thinking nuns.

The best way to extract a single factor for a difference in observations among two samples is for the samples to be equal in every way except for that one factor. Thus, religious people need to be compared not with the rest of the population, or with themselves, as in the nun study, but with a comparable group with similar healthy habits who only differ in their religious practices. This should be possible. Not all atheists indulge in behaviors that are damaging to their health.

In fact, studies exist which indicate a negative influence of religion on health, although you have to dig to find them since they are rarely reported in the media or in the uncritical reviews of those who have already made up their minds. As Kevin Courcey has discovered, Koenig and his collaborators selected mainly favorable studies in their review and ignored even some of their own work that gave conclusions opposite to the ones they preferred to see. For example, a 1994 Koenig paper notes that "several studies have reported an association between psychiatric disorder and religious affiliation, with the rates of disorder highest among non-mainline Protestant religious groups."

Another Koenig study ignored in his review showed that the likelihood of major depression among highly religious Pentecostals was three times greater than among persons who said they had no religious preference. Other investigators have also shown that the religious are frequently among the lowest in mental health. Even if some positive correlation between religion and health can be found in some areas (mental health does not seem to be one of them), religious people, as a group, are less likely than the general population to engage in risky behaviors such as smoking, excessive alcohol, and promiscuous sex. One might attribute this causally to religious behavior but not necessarily to anything supernatural.

Victor J. Stenger is professor emeritus of physics at the University of Hawaii and visiting fellow in philosophy at the University of Colorado.

This article is used by permission of the Skeptical Briefs ( and is based on material from Dr. Stengerís upcoming book Has Science Found God? due to be published by Prometheus Books. Those interested may visit his website at

Research on Religion and Health: A Second Opinion

By Andrew J. Weaver, Ph.D.

Hundreds of scholars representing many of the major universities in North America, Europe, and Israel have published peer-reviewed scientific findings about the positive effects of religion on health in a wide array of settings (Koenig, McCullough, & Larson, 2001). A recent review of more than 1,200 studies on religion and health published by Oxford University Press reported that at least two-thirds of the studies that were evaluated had shown significant associations between religious activity and better mental health, better physical health, or lower use of health services (Koenig, McCullough, & Larson, 2001).

This massive review, with over 2,000 references which included religious effects on mental health, reports that religious involvement was related to greater social support in 19 of 20 studies, better marital satisfaction and stability in 35 of 38 studies, and enhanced purpose and meaning in life in 15 of 16 studies.

It also found lower rates of depression and better recovery in 60 of 93 studies, lower substance abuse in 124 of 138 studies, lower delinquency in 28 of 36 studies, lower suicide in 57 of 68 studies, and less fear and anxiety in 35 of 69 studies. One ofthe most striking facts about this research is that the religious measures used in these studies are often rather crude (i.e., church or synagogue attendance and other religious activities, such as prayer), yet the studies continue to show remarkably consistent positive effects after controlling for multiple confounding variables. As more refined measures are used in future studies we would expect to find even stronger positive effects.

Following the lead of the psychologist Richard P. Sloan, physicist Victor J. Stenger says these studies are of doubtful value. Dr. Sloan declared with certitude in the British journal Lancet in 1999 that the scientific studies conducted in the last half-century that have documented the positive effects of faith are hopelessly flawed and of no scientific value (Sloan, Bagiella, & Powell, 1999). The Lancet review that Dr. Sloan called "comprehensive" and Dr. Stenger appears to accept as authoritative has serious problems of fact according to many experts.

Several noted scientists asked the editors of Lancet in 1999 to allow them to respond to Dr. Sloanís negative criticisms of the field of religious research. They were denied the opportunity. These respected academics and widely published scientists include the former president of the American Gerontological Society and professor at Duke University, Dr. Linda George; Dr. Herbert Benson at Harvard; Dr. Ellen Idler at Rutgers; Dr. Harold Koenig, Dr. David Larson and Dr. Judith Hayes at Duke; Dr. Marc Musick at the University of Michigan; Dr. Stanislav Kasl at Yale; and Dr. Terrence Collins at Kentucky. They published their analysis of Dr. Sloanís review in the International Journal of Psychiatry and Medicine in 1999 (Koenig et al., 1999).

These scholars representing the disciplines of medicine, epidemiology, public health, and social sciences documented in detail the ways in which Dr. Sloan misrepresented the facts by "erroneous, incorrect and misleading statements." They showed how Dr. Sloan was exceedingly selective in what he declared to be a "comprehensive" review. They noted that "only 24 of the approximately 325 studies of religionís relationship to physical health" were mentioned in the article (Koenig et al., 1999).

In addition, even though mortality studies were the central focus of Dr. Sloanís review, "only 17 of the nearly 100 mortality studies were examined in detail in the article, and nine recent high-quality studies were not even mentioned." They pointed out, contrary to Dr. Sloanís declarations, that "eight of these nine prospective studies found a significant inverse relationship between religious measures and mortality after controlling for multiple covariates." Furthermore, the scientists noted that Dr. Sloan mentioned "none of the nearly 900 studies on mental health" (Koenig, et al., 1999). This is a serious omission, as these experts stated, "because one of the strongest rationales for religionís effects on physical health lies in its connection with psychological and social functioning."

At a symposium on "Religion, Faith, and Health" held at the American Psychological Association convention on Aug. 27, in San Francisco, Dr. Sloan was asked if he had ever responded to these strong criticisms by this group of recognized scientists. He said he had not done so because the critique was published in a minor journal (Sloan, 2001). Dr. Sloan declared at the same symposium that research on the effects of religion on health is a waste of time.

Dr. Sloan goes further than Dr. Stenger in his negative critique of religious research and appears to have a personal agenda. On Nov. 6, 1999, in San Antonio, before the Freedom From Religion Foundation, Dr. Sloan applauded the organization for their courage to be atheists and drew a laugh when he expressed certainty that many of the positive findings of faith were just "the placebo effect" (Sloan, Jan/Feb 2000). He went on to suggest possible covert motives among those engaged in religious research and said that he would like to see an investigation conducted on where funding for religious research is coming from. Dr. Sloan has taken a particular interest in the work of the John Templeton Foundation and its research program (Sloan, 2000).

It appears that he believes in some sort of conspiracy connected to the research. He stated in the New Republic in July of 1999 that ". . . the majority of these studies focus on Christians, suggesting a Christian political agenda behind the work" (Easterbrook, 1999). The fact that 87 percent of Americans self-identify as Catholic, Protestant or Orthodox Christians appears to have escaped Dr. Sloanís consideration (Gallup & Lindsay, 1999).

Dr. Stenger may believe that the value of this research is doubtful and Dr. Sloan may think that studying these issues is a waste of time or worse, but most people who are suffering from medical problems place high value on their faith. Religious involvement is a common way that individuals cope with illness (Weaver, Flannelly, & Flannelly, 2000; Weaver et al., 2001). In a study of hospitalized, seriously ill patients, Koenig (1998) discovered that religious coping (prayer, depending on God for support and comfort, reading Scripture) was the most important factor in helping them to deal with illness, in nearly 90 percent of patients indicating that they used religion ó at least moderately ó as a coping behavior. Other studies demonstrate that religious belief and practice help to prevent the onset of depression (Koenig et al., 1992) and help to hasten recovery from depression among patients who suffer from it (Koenig, George, & Peterson, 1998).

Family caregivers often rely heavily upon their religious faith to cope with the burden of caring for their loved ones. Religious involvement can lower the risk of depression (Picot, Debanne, Namazi, & Wykle, 1997), which is a common issue among family caregivers. Rabins and colleagues (1990a) at Johns Hopkins University surveyed caregivers of persons with Alzheimerís disease and of those with end-stage cancer. They found that successful coping was associated with only two variables: number of social contacts and support received from religious faith.

When these persons were followed over two years to determine what characteristics predicted faster adaptation to the caregiver role, again, only the number of social contacts and support received from personal faith predicted better adjustment over time (Rabins et al., 1990). Thus, having support from oneís religious community appears to be one of the most important factors responsible for success in the caregiver role.

Finally, and most compelling to me, hundreds of scientific findings link religious involvement to prevention of "at risk" behaviors and increased prosocial values among youth (Weaver et al., 2000).

Commitment to religious communities reduces the risk of alcohol and drug abuse, premature sexual behavior, depression, suicide, and anti-social behavior, as well as enhancing positive coping strategies among teens (Weaver et al., 2000). In a society with epidemic levels of social problems that put adolescents "at risk" for psychological problems, premature death, and disability, this is no small finding.

I have one firm point of agreement with Dr. Stenger. Even if there is a beneficial relationship between religion and health, one cannot therefore prove from those findings that there is a God or divine involvement in events. Science is an honest, disciplined attempt to describe events, while faith claims make value judgments about the meaning of those events.

Dr. Stengerís internet homepage is replete with declarations of the need for honest inquiry and appeals to rational discourse. It makes me hopeful that he might listen to new information and be able to come to a more balanced view regarding the merits of religious research. I highly recommend to him the scholarly and comprehensive review of research linking religionand health in an article published by John Mueller and colleagues in the December 2001 issue of the Mayo Clinic Proceedings (Mueller, Plevak, & Rummans, 2001).

For a full text of Dr. Weaverís article with complete references, visit this website:

Andrew J. Weaver, Ph.D, is a United Methodist minister, licensed psychologist, and director of research at The Health Care Chaplaincy in New York City.

Volume 2.5 January 2002