Reports of the National Center for Science Education

Do Scientists Really Reject God?

In a recent issue of RNCSE, Larry Witham reported on research he and historian Edward Larson carried out to investigate the religious beliefs of scientists.They had surveyed a sample of 1000 individuals listed in American Men and Women of Science, (AM&WS), using questions originally asked by the Gallup organization in a series of polls of American religious views.The report, entitled "Many scientists see God's hand in evolution", concluded that although scientists were quite different from other Americans in their views of "extreme" positions— such as young earth creationism and atheism—they were very similar to other Americans in the "middle" or "theistic evolution" position.

In the table below, the full wording of Gallup's question 1 is, "Humans were created pretty much in their present form about 10 000 years ago." The difference between scientists and other Americans is striking. Scientists also respond quite differently to the third question, "Man evolved over millions of years from less developed forms. God had no part in this process." But scientists' responses to Gallup's "theistic evolution" question—"Man evolved over millions of years from less developed forms of life, but God guided the process, including the creation of Man"—directly mirrors that of the general public. The "middle ground" is apparently equally attractive to scientists as it is to the general public.


Question Scientists Public
1. Special Creation, 10 000 years 5% 46%
2. Evolution, God Guided 40% 40%
3. Evolution, God had no part 55% 9%

Larson and Witham also asked the AM&WS sample a second set of questions, repeating a survey performed in 1914 by sociologist James H Leuba. Leuba had found that, in contrast to the high levels of religious belief in the general American public, scientists exhibited low levels of belief in God. He predicted that over time, more and more scientists would give up their belief in God, as scientific knowledge replaced what he considered to be superstition. Larson and Witham found to the contrary that disbelief among scientists remained stable: 58% in 1914 and 60% in 1976 (Larson and Witham 1997).

Leuba had taken a subsample of more prominent or "greater" scientists in the AM&WS sample and reported that they exhibited a higher rate of disbelief (70%) compared to less prominent AM&WS scientists. Recently, Larson and Witham asked Leuba's questions of members of the National Academy of Sciences, since AM&WS no longer lists "greater" scientists. They claimed to find that NAS scientists had higher levels of disbelief and agnosticism, reporting "near universal rejection of the transcendent by NAS natural scientists" (Larson and Witham 1998).

Are you confused? How can scientists be so like other Americans in one survey and so different in another? We can find part of the explanation in the considerable differences between the questions asked by Gallup and those asked by Leuba.

The wording of questions in any survey can influence the results. Gallup's questions are quite straightforward, well designed to reveal people's attitudes towards evolution. For reasons that will become important later in this article, a question that requests an opinion on only one issue is superior to one which queries attitudes about two or more.

First, let's look at Leuba's questions, which are, to be charitable, ambiguous. The "personal belief" question attempts to ascertain belief not just in some sort of God, but a very specific kind of personal God.
1. I believe in a God in intellectual and effective communication with humankind, i.e., a God to whom one might pray in expectation of receiving an answer. By "answer", I mean more than the subjective psychological effects of prayer.
1. I believe in a [personal] God...
1914 1998
27.7 7.0

Indeed, the percentage of "yes" answers in 1998 is strikingly lower than that in 1914. Does this mean that fewer scientists believe in God? Not necessarily. Consider how specific this question is. To answer "yes" to this question, one would have to believe that God is not only in communication with humankind, which many religious people do believe, but that God is in both intellectual and effective communication. What is the meaning of "intellectual" communication? "Effective" communication? Someone who believed that God communicated with humankind but not "intellectually" (whatever that means) would have to answer "no." Is "effective" used in the modern sense of the word meaning "something that works well", or in the more archaic (1914) use of the term meaning "to bring about"? Do scientists reading this question today interpret it in the same way as those in 1914?

The clause about answering prayers is also problematic.There are schools of theology that hold that God is personal in the sense of watching over and caring for humankind, but nonetheless, does not answer prayers. We do not know whether members of the general public would respond similarly or differently than scientists do to this definition of God: we do know that there is a wide variety of definitions of God.

Not only have there been changes in theology since 1914, which may be reflected in different Americans' definitions of God, but there have been improvements in survey research techniques. Experienced pollsters simply do not ask paragraph- long questions anymore because they know that they elicit contingent (and therefore difficult to interpret) answers!

Most educated, late 20th century Americans are "test wise" and know that the more components to a question, the more likely it is that the question is "wrong". I doubt that this was the case in 1914, when citizens 'were exposed to far fewer surveys than they are today. I surmise that modern survey-wise scientists would be more likely to answer "no" to a multi-component question like Leuba's number 1 than "yes".

What about Leuba's second question?

2. I do not believe in a God
as defined above.
1914 1998
52.7 72.2

How might this question be interpreted? There is more than one way—which means it's not a good question.You might answer "true" if you did not believe in God at all, which is how Leuba, and apparently Witham and Larson, interpret the question; they describe these answers as demonstrating "personal disbelief." But you might answer "true" if you believed in a different kind of God than Leuba defined! A "yes" on question 2 would include both non-believers and those who believe in a less personal God than that of question 1.

Leuba's third question also allows for multiple interpretations.

3. I have no definite belief
regarding this question.
1914 1998
20.9 20.8

Well, there has been no change in the number of "yes" answers over time, but what does the question mean? To me, a "yes" means "I don't think much about religion in general" rather than meaning, as Leuba, Larson and Witham conclude, "I have 'doubt or agnosticism'." Nonbelievers might very likely answer this question "false", because they do have definite views on this question! Most of the atheists and agnostics that I know have quite definite views about belief in God! Just as with the other Leuba questions, a "yes" answer reflects more than one possible opinion. Positive answers to this question include those who do not believe, as well as those who are not especially interested in the topic.

What one might conclude from the 1998 Larson and Witham study of NAS scientists is that belief in Leuba's definition of a personal God has decreased over time among scientists. The main problem, however, is that Leuba's questions are not well designed for investigating the religious views of scientists (or anyone else).

The Gallup questions, which deal with views of God's role in evolution, rather than general belief or disbelief in God, are far less ambiguous. When these questions were used (Larson and Witham 1997), the answers showed that a large proportion (40%) of prominent scientists believe in a God that is sufficiently personal or interactive with humankind that human evolution is guided or planned.

The title of the recent Larson and Witham article in Nature, "Leading scientists still reject God" is premature without reliable data upon which to base it.

By Eugenie C Scott, NCSE Executive Director
This version might differ slightly from the print publication.