Peter Singer is a Yale philosopher with an interesting take on altruism - namely, how to make it effective. He proposes the following scenarios to make his point:
You are walking to a meeting where you will earn $500. On the way to the meeting you pass a child who is drowning. You can rescue the child but lose the $500. Do you rescue the child?
If you do, there's a cost to rescuing the child.
Now consider that immunizations cost $500 in India. Do you feel the same obligation?
Why do you have a different reaction to these two scenarios?
According to Mr. Singer, there is a psychological difference - salience. You can see the child and take direct action and you will get immediate feedback. The organization in India is more abstract - there is no individual, no identifiable victim. This changes then how you think of the scenario.
He points out that there is also some economic difference rooted in uncertainty - knowing something will happen versus how will I know my $500 will be effective if it's given to a immunization organization.
He asserts that uncertainty is not the main reason people have different responses to the two scenarios. True that skepticism causes a delay in decisions and a trend to prove what works is born out of this delay. But he insists that salience and identifiable victims matter for getting people to give, not evidence of what works.