Group-level Consequences of Psychological Problems — LessWrong