In 1914, psychologist James Henry Leuba surveyed 1,000 randomly selected scientists in the United States, asking them if they believed in a personal God, defined in this way: a God in intellectual and affective communication with man... to whom one may pray in the expectation of receiving an answer. Among those who answered the survey, 41.8% answered the question affirmatively, 41.5% negatively, the remainder did not know or refused to answer. From these data, Leuba drew the conclusion that faith in God would decrease with the advance of science, and predicted that by the late twentieth century virtually all scientists would be atheists.
In 1996, Larson and Witham repeated Leuba's survey using exactly the same question, so that the results were comparable. They found that 39.3% answered affirmatively, while 45.3% answered negatively. These numbers were therefore approximately the same as eighty years earlier. As the authors say in their paper, if in 1914 the high number of atheists was surprising, what was surprising in 1996 was the high number of believers.
These two surveys present a problem: Leuba and his followers consider atheists those who answered their question negatively. But who would answer negatively such a question? Not just atheists, also agnostics and indifferent, plus those who believe in a non-personal God.
In a more recent study (2009) conducted by The Pew Forum among 2500 members of the American Association for the Advancement of Science (the main American scientific association, which publishes the journal Science and co-sponsored the survey), more detailed results were obtained with a different question. 33% of those scientists who answered the survey state their belief in a personal God; another 18% believe in a universal spirit or higher power of some kind; 41% do not believe in any of those things; the rest did not know or did not answer. An additional finding of this study is that the negative response occurs more frequently among scientists older than 65 than among the younger, and less among chemists and biologists than among physicists, astronomers and geologists.
To avoid the problem of the previous surveys, which put atheists, agnostics and indifferent together in a mixed bag, there was an additional question that allowed respondents to ascribe themselves to more detailed groups. The result was as follows: 10% Catholic; 20% Protestant; 8% Jews; 10% belong to other religions. The total (48%) is quite close to the sum of those who chose the first two answers to the first question. Furthermore, 17% declare themselves atheists, 11% agnostics and 20% indifferent, giving a total of 48%, a little more than those who chose the third option in the first question. The difference may be due to the fact that some who did not answer the first question did answer the second.The conclusion is clear: atheism, which by 1915 thought it had won the game, seems to have stalled during the twentieth century and the beginning of the twenty-first. In fact, explicitly atheistic American scientists are still a minority compared to believers, except at the National Academy of Sciences: in another study by Larson and Witham, restricted to its members, the proportion of
believers was much lower (7%).