“A schizophrenic homeless man who once threatened President Trump and Hillary Clinton was arrested for swinging a machete in Times Square — one of five blades he was carrying, cops said Friday.” This sentence from a recent article in a tabloid newspaper is typical of what is seen in news reports throughout the English-speaking world. The psychiatric diagnostic history of the perpetrator of a violent act is casually mentioned; although an explicit connection between the act and the diagnosis is not made, the reader is encouraged to draw a connection between the two (why else would it be mentioned?). Notably, the race/ethnicity of the alleged perpetrator is not stated, in accordance with current practices endorsed by organizations such as the Society of Professional Journalists, which recommend that race not be mentioned unless it is central to the story.
The tendency for news reports to disproportionately discuss the mental health status of perpetrators of violent crime has been documented for many years in a painstaking form of research known as “content analysis.” In 2005, Pat Corrigan and colleagues published a content analysis of the main focus of all stories related to mental illness in major US newspapers (defined as having a circulation greater than 250,000) during six week-long periods. They found that the largest category of stories (39%) concerned dangerousness, and that these stories were frequently highlighted in the front section of the newspaper (37% of all “front” section mental health stories concerned violent crime). More recent analyses of reporting in Canada and the UK found that roughly 40% of reports on mental illness continue to focus on violence and that such reports continue to be featured prominently.
Skeptical readers may have two questions about why the above should matter. First, they may ask if this reporting is not accurately reflecting the reality that people with mental illness are violent, and that their violent acts are associated with their symptoms. Second, they may ask why should it matter what is reported, since readers are free to make up their own minds after reviewing “the facts.” Below, I respond to both of these questions and demonstrate why media reporting practices are major issue that should be addressed through recommendations similar to those that have been made regarding the reporting of race.
Doesn’t reporting about violence among people with mental illness accurately reflect reality?
There are two main reasons why the answer to this question is no. First, there is the magnitude of the association between mental illness and violence. Overall, research (both clinical studies using self-report and research from countries such as Denmark and Australia with population-level “case register” data) suggests that there is an association between mental illness and violent behavior, but the magnitude of the link is much less than what the general public believes. In general, research indicates that roughly 10% of people with severe mental disorders (like schizophrenia and bipolar disorder) engage in violent behavior at some point in their lives, and this proportion is definitely elevated compared to the population base rate for engaging in violent behavior (approximately 3%). However, general population surveys indicate that most people assume that the majority of people with severe mental disorders are violent. For example, findings from the 2006 General Social Survey indicated that approximately 60% of the US general public expected a hypothetical person with symptoms of schizophrenia to be violent (even though the vignette makes no mention of violent behavior). Thus, there is a large gap between the public perception and the clinical reality.
Second, there is the nature of the association, to the extent that it does exist. When a news report states that a person with mental illness has engaged in a crime, one assumes that it is stated because it is somehow related to the crime (for example, because the person was responding to a hallucination). However, research indicates that, in the majority of cases, there is not any discernible connection between symptoms and criminal acts committed by people with mental illness. Researcher Jennifer Skeem and her colleagues have found this in a series of studies (including thorough reviews of criminal case reports), and have concluded that symptoms are only plausibly associated with criminal behavior among people with mental illness about 10-15% of the time. What is going in the other 85%-90% of cases? In those cases, crimes are usually explained by the factors that criminologists have consistently found to explain most crime in the general population- “antisocial attitudes” (e.g., beliefs that everybody is corrupt), substance use, and antisocial peers. Similarly, other research finds that the association between mental illness and violence disappears when we statistically “account” for other factors such as substance use. The finding that crime among people with severe mental disorders is explained by these factors is consistent with the view that crime relates to the overwhelming poverty experienced by people with mental illness in the United States, a poverty that is in turn related to risk factors such as “antisocial attitudes.”
Why does it matter what news media report?
Media reports of the diagnostic status of alleged perpetrators of violent crime are important for two reasons. First, they influence the general public’s perception that mental illness explains violence by only noting mental illness when it is present. One way to make sense of this is to imagine how it would be perceived if, every time a violent act was committed by a person without a psychiatric history (the vast majority of violent acts), it was reported that “the alleged assailant has no history of mental illness or psychiatric treatment.” Of course, that would not be done because reporters would see no reason to state it (as if there were a presumed causal link between not having a mental illness and violence).
The decision to only report mental illness when it is present matters, though. A recent study with a large representative sample of the Australian population found that people who recalled reading stories about violence perpetrated by people with mental illness were significantly more likely to endorse beliefs that people with schizophrenia were dangerous.
The second reason why media reports of the psychiatric diagnostic status of alleged assailants are problematic is because they lead the consumers of these reports to infer a causal link even if one is not stated. This was found in a recent study my colleague Ginny Chan and I conducted, where we randomly assigned people to read one of two versions of a vignette of the same violent incident, with the only difference being that, in one version, the sentence “Mr. John Doe has a history of schizophrenia” was added. We found that, immediately after reading the article, roughly 40% of people assigned to the vignette mentioning schizophrenia (in contrast with only 2% of those in the control condition) stated that schizophrenia was the cause of the incident. We also contacted participants a week later, and by then 64% of those exposed to mention of schizophrenia (and 5% of those not) indicated that schizophrenia was causally related to the crime. Furthermore, when asked to freely report what they thought had caused the incident, they described specific symptoms which were in no way suggested by the vignette. For example, one participant stated “Delusions stemming from schizophrenia appear to have caused this attack” while another stated “I believe that since Mr. Doe suffers from schizophrenia, he might of been hearing voices in his head … that led him to attack this man.” These statements occurred despite suggestions in the vignette that robbery may have motivated the attack.
Why would members of the public infer such detailed explanations for an incident when none was stated? We hypothesize that this is explained by a mechanism called “stereotype activation.” There is no need for an explanation for a connection between mental illness and violence to be articulated, because once mental illness is mentioned in the context of violence, the stereotype of a violent “schizophrenic” is activated. In other words, people “fill in the blanks.”
Law and Crime Essential Reads
When Parents of School Shooters Face Legal Liabilities 
Why Gen X Parents Are Afraid of the Wrong Types of Crime What can be done?
Media outlets were able to change their practices with regard to reporting race/ethnicity, and the same can occur with regard to reporting psychiatric history. That is, if a crime has an obvious connection to psychiatric symptoms (e.g., an individual commits a crime while acting on a delusion) then it can be reported, but when there is no evident connection, then it should not be. This is actually now being recommended by some media organizations. Shifting to this type of approach would have definite benefits for reducing stigma, but it would mean changing long-standing practices that assume that any story that stokes fear of the unknown is likely to draw more readers (or clicks, as the case may be).