Most religions, from Anabaptism to Zoroastrianism, feature some version of Christianity’s “Golden Rule”: Do undo others, as you would have them do unto you. Aside from being a nice concept, how do we benefit from helping others? From an individual perspective, wouldn’t we be better off if we just let others help us out, without giving anything in return? A 2003 study by a team led by Stephanie Brown indicates that we wouldn’t (Stephanie L. Brown, Randolph M. Nesse, and Amiram D. Vinokur, University of Michigan; and Dylan M. Smith, VA Ann Arbor Healthcare System, “Providing Social Support may be More Beneficial than Receiving it: Results from a Prospective Study of Mortality”, Psychological Science, 2003).

Brown and her colleagues analyzed data for 423 older married couples in the Detroit area who had responded to a 1988 survey on lifestyles. Further, they combed local mortality records for the next five years to see which individuals who participated in the survey had died. The goal was to correlate the deaths to social behavior in the individuals based on their survey responses. The surprising result of the analysis is that people who helped others were less likely to die in that period than people who only received help.

How much less likely? It’s more difficult to compute the probabilities than you might think. For example, women tend to live longer than men and perhaps are more likely to help others as well. Could this account for the difference? Healthier, more robust people might be the ones more likely to help others, since they are more able. Wouldn’t they also be more likely to live longer—not because of their generosity, but just because they are healthier? And how do you measure whether someone’s actually giving or receiving support anyway?

The study measured two types of support, instrumental support and emotional support, and two dimensions: giving and receiving. Giving support was determined by asking whether respondents had helped others outside the family with shopping, housework, child care, or other tasks. Receiving support was measured a bit differently: respondents were asked whether they “could count on” others including family members to help if they needed it. Emotional support was measured only for spouses: respondents were asked if they felt loved and if their spouse listened to them, and vice versa.

Next Brown and her colleagues computed the likelihood of mortality without consideration of these social factors: just considering the effects of age and gender. The results for the survey corresponded to known national averages. They then considered social contact—whether the individuals socialized outside the home or using the telephone, a factor known to decrease mortality based on previous research. Brown et al.’s results corresponded to the previous research, and established a baseline from which to assess the factors they were considering in this study.

Working from this baseline, they found that giving social support reduced the odds of mortality by over 40 percent, and receiving social support increased the odds of mortality by 30 percent during the five years following the survey. The effect of social contact, by contrast, was only a 19 percent reduction in mortality. Similarly, giving emotional support to a spouse was also associated with a 30 percent reduction in mortality after controlling for other factors, but receiving emotional support did not significantly alter mortality.

Of course, these results are all just correlations: for a true experimental study, groups would have to be randomly assigned to giving or receiving conditions—indeed, it’s difficult to conceive of an ethical experiment to duplicate Brown et al.’s results, which is why we must sometimes rely on correlational data. In any case, this study does offer compelling evidence that philosophical statements like the Golden Rule aren’t just platitudes, they may also benefit those who try to live by them.