Another interesting finding, from the GSS, is that both atheist/agnostic and "no religion" Americans have more confidence than others in the scientific community, but not in medicine. This seems to indicate a greater trust in science as a source of truth or value for the less religious than for the more religious. In a 2002 VCU survey, those who said religion was an important part of their lives were more likely than those who said religion was not important to believe that scientific research does not pay enough attention to moral values.

Less-religious Americans were also 15-31 points more likely than others to worry that the country will become intolerant of behaviors that do no harm and to think the federal government should not be involved in promoting moral values in our society.

Areas where we have only a single way of defining less-religious Americans also showed significant differences:

• Those who said religion was not an important part of their lives were 24 points more likely to favor human cloning, provided it is used only to help develop new medical treatments.
• Low-religiosity Americans were 19 points less likely than others to consider it unacceptable to have a child without being married.
• Low-religiosity Americans were 26 points less likely than others to oppose casino gambling in their state.

No doubt, many more such differences are waiting to be found by further investigation in many other areas.

In an era where religious values are likely to garner a great deal of attention, the lesson is clear. If need be, struggle over which measure of religious or nonreligious best fits your survey's purpose. But for heaven's sake, make sure you include one of them in any survey that measures political or social attitudes. There's no excuse not to.

John Benson is managing director of the Harvard Opinion Research Program, Harvard School of Public Health, and associate editor of Public Opinion Pros.

Additional data

Attitudes About God and Atheism
Demographics of "no religion" and "low importance of religion" respondents
Additional readings