Finding New Information for Ecological Inference Models: A Comment on Jon Wakefield, 'Ecological Inference in 2X2 Tables'