Never Trust the Science - On the need to identify bias & interpret data yourself

How about a big vetted database like arxiv of all hypotheses, all proposed experiments to test them, and all experimental results?