The New Yorker article

The follow-up on Wired

An interesting article on how statistical randomness, unconscious confirmation bias, publication bias, and lack of sufficient replication can affect reported data and therefore scientific conclusions. I'm curious as to what LessWrongians think about it.