Recall that, in Singer's thought experiment, if you'd be willing to ruin your $500 iPhone by jumping into a pool to save a drowning child, you should also be willing to spend $500 to save a child's life. Since there are plenty of charities in the third world that can save lives at fairly low cost, people are not consistent if they would do the former but do not do the latter.
Jason ably points out a problem:
I'm reminded of a time I was walking our then four year old to daycare. We came upon an earthworm on the sidewalk. The rain had ended and the worm would likely die without intervention. We picked up the worm and put it onto the lawn. Then we turned the corner and saw several hundred worms on the sidewalk. She prepared to start picking them all up, and I instead brought her on to daycare. There are only so many hours in a day.