I enjoy the Econtalk podcast, and one of the concepts I’ve taken away from it is the notion of scientism – defined as the notion that whatever science is available about a topic explains the topic.
Some things are inherently difficult to study, like economics or medicine (complex systems) compared to traditional scientific topics like chemistry or metallurgy. It is harder to design studies to measure the impact of a minimum wage increase than it is to measure the impact of 55% humidity increase on exposed copper.
Seemingly our understanding increases in it’s certainty with the more complex the system.
At what point are you safer using a priori logic to make decisions than evidence? How many studies do you need to model complex behavior?