Very religious Americans are more likely to practice healthy behaviors than moderately religious or nonreligious people, according to findings released Thursday from the Gallup-Healthways Well-Being Index.

The difference in habits was most pronounced in terms of smoking, with the nonreligious 85 percent more likely to be smokers than those who are very religious. Very religious Americans also outpaced the other groups when it came to eating healthy and getting frequent exercise.

The survey defined very religious people as those who report that religion is an important part of their daily life and that they attend religious services every week. The researchers note that religious doctrine may in part explain the relationship:

Seventh-Day Adventists, for example, strictly adhere to vegetarian lifestyles free of alcohol and smoking, while orthodox Mormons and Muslims do not drink alcohol. In some Christian denominations, gluttony and sloth are considered two of the seven deadly sins, and many evangelical faiths frown on drinking and smoking. The Bible indicates that one's body is the "temple of God," which could in turn help explain the relationship between religious orthodoxy and exercise and certain types of food consumption.

But there are other implications, they add:

It is possible, of course, that the noted relationship between health and religiosity could go in the other direction--that people who are healthier are the most likely to make the decision to be religious. This could be particularly relevant in terms of church attendance ... Healthier people may be more likely and able to attend religious services than those who are less healthy.

It may also be possible that certain types of individuals are more likely to make healthy lifestyle choices and more likely to choose to be highly religious. The most parsimonious explanation, however, may be the most intuitive: Those who capitalize on the social and moral outcomes of religious norms and acts are more likely to lead lives filled with healthier choices.