What really helps the poor?

It’s difficult to test whether poverty relief actually works. Do randomised controlled trials provide a scientific measure?