Paul Bloom in Slate:
Arguments about the merits of religions are often battled out with reference to history, by comparing the sins of theists and atheists. (I see your Crusades and raise you Stalin!) But a more promising approach is to look at empirical research that directly addresses the effects of religion on how people behave.
In a review published in Science last month, psychologists Ara Norenzayan and Azim Shariff discuss several experiments that lean pro-Schlessinger. In one of their own studies, they primed half the participants with a spirituality-themed word jumble (including the words divine and God) and gave the other half the same task with nonspiritual words. Then, they gave all the participants $10 each and told them that they could either keep it or share their cash reward with another (anonymous) subject. Ultimately, the spiritual-jumble group parted with more than twice as much money as the control. Norenzayan and Shariff suggest that this lopsided outcome is the result of an evolutionary imperative to care about one’s reputation. If you think about God, you believe someone is watching. This argument is bolstered by other research that they review showing that people are more generous and less likely to cheat when others are around. More surprisingly, people also behave better when exposed to posters with eyes on them.
Maybe, then, religious people are nicer because they believe that they are never alone. If so, you would expect to find the positive influence of religion outside the laboratory. And, indeed, there is evidence within the United States for a correlation between religion and what might broadly be called “niceness.”
More here.