Oh boy, here we go again.

Today’s story comes from PBS, with the eye-grabbing title:

Heavy marijuana use may cause poor memory and abnormal brain structure, study says

Cue college students and worried mothers clicking away. A study by Northwestern University looked at long-term effects of cannabis in teenagers and compared the size and shape of subcortical structures of the brain in users vs. non-users. Interestingly, schizophrenic subjects who used cannabis had more exaggerated differences, suggesting a possible biological link between cannabis use and the condition.

But tellingly, the study was not reported quite like that. Indeed, the PBS report focuses on the effects on cannabis on normal teenagers, and its title was originally more strongly worded:

Correction: The title of this post was corrected to indicate that researchers have not concluded a direct link between heavy marijuana use and abnormal brain structure or poor memory, but to reflect that the study shows a possible association between the two.

And this is exactly the issue we see with these kinds of reports, and what I would like to discuss. You could argue there is a fundamental incompatibility between piecemeal research and the journalistic demand for headlines. In order to make good news, a science story must be gripping – it must tell us something new that is interesting, or relevant to our lives. It must have a beginning, middle and end. Good story telling is what a good journalist does.

The problem is that this leads to de-contextualisation, with studies looked at in isolation and generating the impression that each of them is a single morsel of the ‘truth’. This is why we feel flabbergasted when yesterday’s Guardian tells us coffee causes cancer and today’s CNN report says is cures cancer. In reality, research studies form continuous lines of evidence that must be interpreted together, approaching something closer to the truth.

Let us examine the specific case of cannabis usage reported here. While this report may lead you to think that THC will destroy your brain, there is evidence that cannabis use is not nearly as harmful as heavy consumption of other psychoactive substances such as alcohol. Of course, cannabis use remains an unsafe recreational activity, solidly linked to cardiovascular, respiratory and mental disorders (here, here and here). It’s all a matter of relative risk.

How we read a scientific paper and a newspaper article are two different things. We cannot expect our audience to know the full background to the story, or the current consensus on a given topic. But it can guide us towards learning more – a simple statement of caveats leads many readers thinking “well hang on, why?”.

Science journalism can be a powerful window into the vastness of science, and encourage people to explore, learn and wonder. So let us encourage this, show how exciting a finding or theory is, and leave the door open. If enough people are left in suspense, wondering why, they will go and find out – and that is a good thing.


Press release from Northwestern University and original paper published in the journal Schizophrenia Bulletin.