Design a site like this with
Get started

Correlation and Causation

Correlation and Causation

Recently a story came out with the sensationalized headline “Teen Marijuana Use Linked with Schizophrenia”. The informative post below describes why this headline was misleading, and give perspective on the sociological problems associated with media coverage of mental illness.

This post was chosen for #SoG+CuratorsChoice  by Buddhini Samarasinghe who enjoys seeing #ScienceMediaHype debunked so well 🙂

Originally shared by Zuleyka Zevallos

Why Correlation is not Causation: Cannabis Use & Schizophrenia

As so often happens, a post from Science on Google+, a community I help moderate, has got me thinking about how easy it is for headlines to quickly lead to #ScienceMediaHype . A post with a link to a news story has the headline, “Teen Marijuana Use Linked with Schizophrenia” ( As a sociologist with an interest in mental health, this sets off alarm bells. The discussion on our community quickly turned into a debate about the correlation presented in the headline. As a few of our community members pointed out, correlation does not equal causation.  My post provides a summary of the actual study and I discuss the sociological problems associated with media coverage of mental illness.

Study on Working Memory

The linked article does an okay job of describing the study but the headline and its focus over-extends the study’s findings. This is the problem with media stories: headlines can shape the way the public understands scientific findings. Journalists present quotes from scientists that fit the angle of their story, putting less emphasis on other aspects of the research. 

The study is published in  Schizophrenia Bulletin. The sample includes 44 healthy controls, 10 people with a history of cannabis use disorder (CUD), 28 schizophrenia participants with no history of substance use, and 15 schizophrenia patients with a CUD history ( Ninety percent of the participants who had schizophrenia already had a CUD history prior to their mental illness. Most of the cannabis users were heavy users, smoking cannabis daily or at least weekly, and most also smoked cigarettes heavily, a variable that the researchers wanted to test. 

The study actually tests working memory deficiency not the cause-effect relationship between schizophrenia and cannabis use. Participants were given memory activities and then brain imaging was used to see their brain patterns. I’ve included the three images from the study. Neuroscience is not my area of research, but I include the diagrams in case other researchers should be interested.

The researchers note that without substance abuse, schizophrenia inhibits cortical development. They note less is known about how cannabis affects brain symmetry, though it is known to disrupt the hippocampus,  which is related to our limbic system. The hippocampus is linked to information retention for both short and long-term memory as well as other functions like spatial navigation. 

The study finds that use of cannabis at an early age impacted memory function. The sample had an average age of 24, so patterns associated with longer-term development need further study. 

The study observes that cannabis users and the participants living with schizophrenia both have problems with memory tasks. At the same time, the researchers not that the brain asymmetry observed when carrying out memory tasks may be linked to a” neurobiological vulnerability” among schizophrenia sufferers. That is, that the observed pattern may be the outcome of a predisposition to substance abuse. For example, the researchers note that similar brain patterns are found amongst cocaine users. So, to put it another way, schizophrenia users may be drawn to substance abuse. More on this below.

Finally, the authors conclude what many of our community members had been discussing: that there is no direct cause and effect relationship. “Although our data may be compatible with a causal hypothesis, the cross-sectional data do not allow us to test causal relationships or reject alternative explanations. Thus, the shape differences could be explained as either due to the effects of chronic cannabis abuse or the presence of biomarkers that characterize a vulnerability to the effects of cannabis.”

Their research notes that with laws changing, cannabis may be more readily available to youth with a predisposition towards schizophrenia. This makes their research all the more pivotal. This is both in terms of better understanding how brain development is affected by schizophrenia and the social, health and subjective reasons why youth may engage in cannabis use at different stages of their disease. 

Research on Correlation

There are many studies that have linked self-reported cannabis use in early adulthood to an increased risk of developing schizophrenia later in life. One of the most widely cited studies involves Swedish conscripts of 1969, with a follow up study confirming the results ( Nevertheless, the direct association between cannabis use and schizophrenia is disputed. For example, there is more to be learned about the relationship between the time that someone starts using cannabis and their first schizophrenic episode ( 

There are three hypotheses to explain why young people with schizophrenia use cannabis ( 

1) Cannabis triggers the disease in people with a predisposition, as is suggested by the study at hand.

2) People with schizophrenia use cannabis as a way to self-medicate or manage their experience of the disease. The research does not support this strongly, though cannabis use may give sufferers a perception of control over their disease. This is not something to be dismissed and requires further research. One study from 2012 suggests that one component of marijuana, cannabidiol, may be used to treat schizophrenia. This component does not contain THC which is responsible for the intoxication effects of traditional marijuana. Instead, cannabidiol may reduce the symptoms of psychosis, but further research is needed to fully test this treatment (

3) Cannabis use can trigger schizophrenia through confounding variables. This is the prevailing medical view, though the association is not so neatly weaved together. 

Problems with Media Coverage of Mental Illness

The article that covered this story went for a shock value headline. Headlines prime audiences about what they should expect from a scientific article. Sensationalised headlines invite personal opinion based on individual experience. This may range from “I smoke pot and I’m fine” to disparaging comments about “crazy” people. This is the problem with the way in which news headlines shape public discussions of science. Should people read an article before discussing the soundbite? Of course. Does this happen in practice? Not as much as it should. The idea that sensational headlines “sell” is flawed. Shock headlines sometimes get people to click on a link. On a social media site like Google+, some people will go off the headline and the text in a post. In fact, most of the social media research shows that people rarely want to click away from the social site they’re currently on (more on this in another post). This is why we stress in our community that posts should summarise the science behind an article in detail. The link should be there for people who want to read further (and if you write a good post, people will be more willing to click on a link!).

While research suggests that people are sceptical of media reports, not everyone is trained to think about research the way scientists do. Paywalls also stop people from reading the study for themselves (as well as the technical language used in academic journals). This is why scientists need to step up and debunk bad science journalism.

Moving Beyond Individual Speculation

The research has established a correlation, but causation is disputed. The evidence strongly suggests that cannabis compounds schizophrenia and the gravity of this finding cannot be reduced. Nevertheless, media stories that run with a causation headline only serve to spread misinformation.

Bad science writing invites cannabis users to say: “I use it and I’m fine!” It also serves to reinforce a cultural stereotype that people with mental illness may have avoided their condition if they’d only stayed away from recreational drugs. In the end, the causation narrative does more damage and serves only to stigmatise both cannabis users and schizophrenic sufferers as deviants. People living with schizophrenia are doubly stigmatised as their cannabis use makes it seem as if they are wilfully contributing to their illness. Research is seeking to better understand why some people with schizophrenia rely on cannabis. 

Mental health is a serious matter. It shouldn’t be reduced to an alarmist headline. Mental illness is part of the human condition. Addressing the treatment of schizophrenia requires compassion, not judgement. Most of all it needs solid scientific research, not dismissive condemnation based on personal or social conjecture.  

#sociology   #socialscience   #schizophrenia   #cannabis   #science   #research  


Join the Conversation


  1. Joseph Grant I would agree that as people who promote science we should keep an open mind. However, I don’t think that having another correlational study refutes any of the ideas presented by Zuleyka Zevallos . If the research is correlational, you could easily say if you smoke pote you have a 40% increase in the risk of developing schizophrenia (what the guardian reported) as saying if you have schizophrenia you are __% more likely to smoke pot. You can’t infer causality in either direction. In order to effect causality you need either a true experimental setup (randomly assign people to a group that smokes pot or doesn’t smoke pot with a representative sample; this won’t happen because you are assigning people to a risky category) or have a longitudinal design where you can measure pot smoking at a particular time and show that that smoking predicts psychotic symptoms at a later date. Even the latter is difficult to truly draw causal inferences but it is at least closer to being able to do so.  


Leave a comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: