You write in the book that data sets have shaped con cets, theories and policies more profoundly than we'd like to admit. It's much easier to condemn corruption among the poor, very difficult to talk about influence politics among the rich. I see this practical reality that people will pick agenders that are easy. So we need to have a different set of methological tools. You know, people were so angry at the unbundled corruption index. Why? Because i'm doing something that itot impinges on something personal to them.