The likelihood that a new scientific finding is true varies inversely with your desire to talk about it at parties.