[yellowstuff]

Social sciences do generate testable predictions, but they often run poor studies that fail to falsify untrue predictions. 538 recently had a good explanation.

http://fivethirtyeight.com/features/science-isnt-broken/