The argument about whether foreign aid "works" rages on. Recently, Paul Collier sought a practical middle path between William Easterly's development pessimism and Jeffrey Sach's development boosterism. How can smart people draw such contradictory conclusions from the same data? This new working paper by CGD research fellow David Roodman answers this question by describing consensus where it exists and identifying sources of controversy. Roodman concludes that, while aid has eradicated diseases, prevented famines, and done many other good things, given the limited and noisy data available, its effects on growth in particular probably cannot be detected.