The Bloomberg headline reads “Seven cups of coffee a day may lead to hallucinations.” The study cited is by researchers at Durham University that links what appears to be heavy consumption of coffee to hallucinations. But there is some strange stuff going on with this research, visible only if you dig a bit deeper.

First of all, the 7 cups is actually instant coffee, which has a very low caffeine content. The news account states that…

“People who drink at least 330 milligrams of the stimulant a day were three times as likely to have hallucinations as those who consumed less than 10 milligrams a day, Durham University researchers found in a study of 219 college students published today in Personality and Individual Differences."

So how much is 330 mg of caffeine? Well a “tall” (12 oz) coffee of the day at starbucks has 260 mg. Have one “Grande” and, well, you ring the bell at 330 mg. So let’s stop and think about the inherent prevalence of caffeine induced hallucinations now that we know that they are really talking about one 16 oz coffee. It can’t be high.

But there is even more strangeness if you actually read the published scientific article. The short of it is that the authors never report an analysis that divides risk of hallucination to specific caffeine consumption. 330 mg never appears in their report. Comparing 330 mg drinkers to 10 mg drinkers does not happen within their report either. In fact, they never directly relate any consumption amount to hallucinations. Instead they conducted analyses that correlated consumption per kg (i.e., the person’s weight) to hallucinations after controlling for stress and other variables.

So one is left to wonder where the news story got the numbers they are reporting. And, of course, the findings the actual report does present are in no way indicative of a causal relationship. This is just a correlational study.

And finally, looking at the measures, it’s clear that no hallucinations were actually ever observed within their sample. They did not measure hallucinations; instead they measured hallucination-proneness, which they describe as a measure of “predisposition to hallucination-like experiences.” These were, after all, college students and not a clinical population of schizophrenics. So how many hallucinations would one really expect?

So, what’s a person to do? My best advice is avoid reading sloppy news about research studies, and drink your cup of coffee in peace.