Effective Altruists should consider religions as possible conduits for effective altruism, and religious faith as a means of increasing their own rationality. ^[1] 

To this end, I propose an experiment among (non-religious) effective altruists, where effective altruists are randomly selected to adopt the tenets of different faiths for a significant period of time (90 days to a year). The analytically ideal experiment would randomly select effective altruists to adopt different faiths, leaving some non-religious effective altruists as a control group. In fact, in the analytically ideal experiment, even religious effective altruists could be randomly selected to change their faith or adopt non-faith. Clearly, this has ethical implications that almost certainly eliminate this as an option. 

At the risk of introducing selection bias, an ethical experiment could ask for non-religious effective altruists to volunteer to be randomly selected into different religious faiths or no religious faith at all. There would have to be the willingness to select into beliefs that volunteers would find incorrect or even abhorrent at first. Ideally, an effective altruist randomly selected into the faith would do their best to sincerely practice the faith and adopt its tents for 90 days or more.

Different religions would then be measured in their responsiveness to effective altruism when it is discussed by their newfound believers. To what extent does the religion find the idea compatible with their beliefs? How does this vary by the gender, race, or other characteristics of the newfound believer? To what extent do religious faiths try to redirect altruism efforts into social insurance for their own believers? To what extent does the religion try to redirect EA towards expanding the religion itself or to pet political causes? To what extent do religious faiths then incorporate EA beliefs into their own charitable works? 

Potential religions include: Catholicism, Sunni Islam, Progressive Christian (and/or Quaker), Shi'ite Islam, Eastern Orthodox, Orthodox Judaism, Latter Day Saints, Conservative Protestant, Buddhism, Reform Judaism, Hinduism, Liberal Protestant, and others, with no religion left as a control. The number of Effective Altruists willing to volunteer clearly determines the number of religions in the treatment.

There are some reasons that this experiment may prove to be revolutionary, and will certainly prove highly informative. 

First, I introduce my main assumption, which also serves as my idealized, stylized understanding of effective altruism: Effective Altruism means that one does the best one can to make as much income in an ethically responsible way, one lives far below one's means, and one donates to causes that offer the greatest clear (likely measurable) benefit to humanity.

I am not part of effective altruism, but I find this idea very useful as a standard of very good behavior. It is useful partly because I fall short of it and I can see how I fall short. I admire this exacting standard, and I admire the standard precisely for its disregard of my desire to feel good about myself. It doesn't coddle me or let me off the hook. In short, insensitivity is a virtue in this context.  Effective Altruism is of course not intended to cause discomfort, it is in fact very moral in my view, but honest statements of the highest moral behavior will cause discomfort for those who wish to feel morally excellent without actually being morally excellent, and in my view that covers a lot of people. In short, the very high quality of EA's core idea may limit EA's appeal for many people.

For this reason, attempts to broaden the appeal of effective altruism carry the downside risk of trying to attract people who want to dilute the challenging idea of EA in order to feel more comfortable or to have EA comport with their more social orientation. This may express itself as a belief in giving money to the right people instead of a focus on helping humanity in clear, often measurable ways. 

Consequently, I reframe the issue: Rather then modifying EA to broaden appeal, instead find the people and cultures who can appreciate EA as-is.

There are reasons to believe that some religious cultures may be better able to appreciate the virtues of moral truths that express standards in ways not designed to assuage people's feelings. Indeed, expressing truths in ways that do not assuage people's feelings is likely the only way to express many important truths.

First, and I believe a literature review is not required here, there is a myriad of research in cognitive psychology that shows that a large part of human irrationality can be summed up as an instinct towards irrational collectivism.^[2]  In addition, recent research shows that so-called autism spectrum disorder is also an expression of advanced rationality. This indicates that most people, in the natural state, are simply unable to absorb certain truths expressed by those with enhanced rationality.

Second, religions are myths with multiple complementary parts that form cultures that have proven adaptive over time. We must consider, then, that religions may have proven adaptive because they found ways for neurotypical people and perhaps even psychotics to better appreciate more socially insensitive people with enhanced rationality, and perhaps also aided socially insensitive people with advanced rationality to get along with neurotypicals and psychotics, and helped people with enhanced rationality to assert themselves when needed. 

Of course, religions also exist to create group cohesion, conquest, and social insurance, so religions may vary in the way they handle cognitive diversity. Certainly some religions could enforce irrational collectivism to a greater degree, while other religions may correct the excesses of irrational collectivism. That is one question this experiment can help answer. 

For this reason, selecting effective altruists to randomly join different religions and see how these religions grapple with the ideas of effective altruism would certainly prove to be deeply informative, and may represent a new, productive social vector for the spread of effective altruism. 

Thank you all very much for your time.