Eduwonkette gives a clear examination of data that is generally anything but.

This post is more about public policy than charity, but it shows – at a glance – a lot of the problems with the traditional approach to charity (giving as an act of the heart without the brain; trusting charities that mean well, without carefully studying their outcomes). The fact is that we don’t know what works in education. Blind faith won’t cut it, and neither will flimsy data.