“There are three kinds of lies: lies, damn lies and statistics.” Most of us can relate to Mark Twain's sentiment quite well; even as scientists, we exhibit a notable sense of apprehension towards the subject. The causes are not difficult to find: in our undergraduate lectures in Statistics, unlike those in Physics, for example, we are not taught any underlying general principles; rather, we are presented with an array of seeminglyad hoctests and procedures with no obvious rationale for deciding which one to use. But does statistics have to be this way?