Giving the same information to multiple scientific teams can lead to very different conclusions, a report published today in Nature shows.