I am a massive, massive fan of Thinking Fast and Slow. However, I am careful about application, because a lot of those studies seem not reproducible.