A mathematician is a device for turning coffee into theorems, according to Renyi's famous joke. But since most mathematicians don't prove theorems nowadays, I tend to think of the natural mathematicians as those finding More or Less as gripping as an Agatha Christie mystery (whether or not they like set theory).
......On the news this morning, there was a story about researchers at Durham have discovered that coffee makes us hear voices. They conceded that maybe those who hear voices tend to be more stressed, and so drink more coffee, but that seems like an odd concession to me, as people who're stressed usually turn to, if not alcoholic beverages, then chocolate or cigarettes, or cannabis.
......Scientists have also claimed that cannabis can trigger psychosis. That seems more plausible, as cannabis is the famous hippy drug, but I wonder even about that. Many of the most obvious direct tests of that hypothesis would be rather unethical, and the indirect tests (e.g. statistical correlations) would be vulnerable to selection biases. It's not just that schizophrenics might be more likely to disobey the law. There is apparently a part of the brain that is involved in religious experiences, e.g. Richard Dawkins had his stimulated and experienced nothing, apparently.
......If some people are more disposed to such things (again, whether they're born that way, or whether the brain adapts to their chosen way of life, is less obvious) then that would be a relatively obscure but effective source of such a bias. Incidentally, such studies need worry religious people surprisingly little. The traditional view of God has him timelessly creating us, and so faces similar problems, e.g. from our choices to turn to him etc. And Open-theistic views must face the facts of life anyway, e.g. that some of us are born richer, or better looking, or with better brains in other ways. If they can do that, they'll be able to live with similar facts.
......Anyway (oh how tangents attract the active mind), it occurs to me that those with such a religiously inclined brain might be more inclined (statistically, not each of them of course) to hear voices and also more inclined (similarly) to drink coffee, whether because they don't drink alcohol for religious reasons, or because they like to be awake to creation, or whatever. They may also be a little more inclined (statistically) to smoke cannabis, insofar as that's associated with the mystical side of hippies, or the religious side of Rastafarians, etc. If anything, we'd expect a greater corrolation with coffee, of course (and it does seem more plausible that coffee is not actually causing us to hear voices).
......Anyway, that's one possible explanation: a common cause leading to interlinked tendencies. Another explanation is that coffee in large quantities makes some of us irritable and tense, and perhaps over-sensitive, and so maybe more likely to hear voices insofar as we're already slightly inclined towards that (although I'm not ruling out the possibility that people take coffee because they're feeling stressed), but even in that possibility there's room for biases of the former kind—a common partial cause—e.g. a weak willed person might be more likely to over indulge in coffee, and less able to resist hallucinating. Similarly, they might drink more and get into fights, for such a reason. Or they might (also) react to the drink by getting more aggressive themselves. Note that that would be no reason for those who drink to relax and socialise to drink less (you may have guessed that I drink a lot of coffee:)