Protestants who are reminded of their religion do more good deeds and trust that other people will do so, as well. But when Catholics are reminded of their religion, they do less public service and are less trusting that others will do so.

These are among the findings of a new Cornell study that explored the connections between religious affiliation and economic behavior.

Previous work on the effect of religion was based on correlations. For example, "People find [that] a sample of Protestants is more hard working than Catholics, and they want to conclude from that that it's some sort of effect from Protestantism," said Daniel Benjamin, assistant professor of economics and co-author of the study, which is published as a National Bureau of Economic Research working paper.

"But Protestants, Catholics and Jews differ in all sorts of ways besides their religion, and those other factors -- some of which might be very hard to measure -- could in fact be what's responsible for the difference in behavior," he said.

To isolate the effect of religion, Benjamin and co-authors James Choi of Yale and Cornell undergraduate Geoffrey Fisher '10 used a method developed by psychologists called priming in which 827 Cornell students unscrambled sentences. One group's sentences contained words associated with religion -- sacred, God, prophet -- and the second group's did not.

"Each of us has multiple identities. I'm white, I'm Jewish, I'm an economist," said Benjamin. "But the extent to which my behavior is affected by any of those identities depends on how salient that aspect of my identity is at a particular moment."

Belief-related words made the subjects' religious identity salient by "activating neural connections related to whatever sort of norms of behavior the subjects associate with their religion, in ways they might not be conscious of," Benjamin said.

The researchers then observed the behavior of the primed and unprimed subjects in a series of economic games, which allowed them to infer that whatever differences they uncovered were the causal effect of the religious identity norms. They then analyzed the data separately by stated religion of the participants, and tested various hypotheses.

"We found support for the hypothesis, for example, that Protestants who were exposed to the religious prime would give more money than Protestants [whose] religion was not made salient," said Benjamin, and "that the primed Catholics contributed less to the public good. That was exactly consistent with the existing hypothesis literature."

However, the researchers "found no effect of priming religion on any measure of work ethic [or] for the idea that religion makes people more forward looking, more willing to delay gratification," Benjamin said.

The study produced mixed evidence about risk taking: Primed Protestants were less inclined to take risks, while primed Catholics took more risks. Interestingly -- and unpredicted by any pre-existing hypothesis -- Jews who were primed were more reciprocal, repaying kindness with kindness, than Jews who were not primed.

"Another interesting thing we found is that Protestants who believe more in divine punishment give more to the public good when primed," Benjamin said.

As his role in the study grew, Geoffrey Fisher '10 -- who helped program the experiment, analyze data and edit the paper -- was promoted from research assistant to co-author. "Having the opportunity to co-author a paper with Dan and James has been a wonderful experience to have at such an early point in my career," said Fisher, who will be a graduate student at Caltech in the fall studying how the brain makes economic decisions and is affected by environmental stimuli, such as primes. "They gave me a glimpse into what a real research project is like."

The research was funded by the National Institutes of Health's National Institute on Aging in two separate grants, one to the National Bureau of Economic Research and one to Cornell's Institute for the Social Sciences.