46% of film characters with mental health conditions depicted as violent
August 6, 2020 7:43 AM   Subscribe

The Annenberg Inclusion Initiative study on mental health in popular storytelling (pdf) examined 100 top-grossing films and 50 popular TV series to understand the prevalence and depiction of mental health conditions in entertainment.

Using a purposefully broad definition, the prevalence of mood disorders, anxiety, PTSD, addiction, suicide, autism spectrum disorders, and other conditions was evaluated and investigated to understand whether mental health conditions are dehumanized, stigmatized, and/or trivialized in popular media.

The study assessed whether characters with mental health conditions were shown with a job (only characters 13 years of age and older were included). A total of 36 out of 81 characters (44%) with a mental health condition were depicted with an occupation.

Occupations of Characters with Mental Health Conditions
Occupation Category --- Frequency Count
C-suite (e.g., CEO) --- 2
Law Enforcement/Legal (e.g., police, prison guard, military) --- 10
Ruling Royalty (e.g., queens) --- 3
Healthcare (e.g., doctor, nurse, therapist) --- 3
Financial Services (e.g., accountant) --- 2
Service/Trades (e.g., nanny, janitor, mechanic, performer) --- 10
Education (e.g., teacher) --- 1
Other (e.g., criminal, pilot) --- 5

Another AII study "Inequality in 1,200 Popular Films" examined the gender and underrepresented racial/ethnic status of leading characters across the 1,200 top films from 2007 to 2018.
posted by spamandkimchi (19 comments total) 20 users marked this as a favorite
 
During my undergrad, I did something similar but much smaller in scope and professionalism with Law & Order. It was mostly an excuse to watch a lot of Law & Order and to read Law & Order episode summaries on the Internet, which were two things I was doing anyway. I don't have the paper anymore, but if you discounted alcoholism, almost all of the mentally ill people on law and order are depicted as perpetrators of crimes, rather than as victims. Including alcoholism, it still held true, but there were enough alcoholics on the show that it balanced things out somewhat.
posted by jacquilynne at 8:28 AM on August 6, 2020 [8 favorites]


46% of film characters with mental health conditions depicted as violent

While I’m sure the real gist of this figure stands, and that Hollywood really does treat mental health conditions unfairly, to deleterious effect on real people with mental health conditions, I feel it’s worth pointing out that many top-grossing films depict every character as violent. Also almost certainly to deleterious effect on real people.
posted by aubilenon at 8:57 AM on August 6, 2020 [10 favorites]


I'm honestly surprised it's only 46%.
posted by teh_boy at 9:12 AM on August 6, 2020 [3 favorites]


Taking the top 100 films of each year gets you pretty far into the weeds. Here's the #100 movie for each of the second study's years according to IMDB:
  • 2007: The Last Mimzy
  • 2008: Defiance
  • 2009: Fighting
  • 2010: Leap Year
  • 2011: The Iron Lady
  • 2012: Man on a Ledge
  • 2013: The Incredible Burt Wonderstone
  • 2014: Draft Day
  • 2015: Mr. Holmes
  • 2016: Resident Evil: The Final Chapter
  • 2017: The Circle
  • 2018: Overlord
posted by Hatashran at 9:24 AM on August 6, 2020 [1 favorite]


Media depicts mentally ill persons as violent. The general public thinks of mentally ill persons as violent because that's what the media tells them. The general public is then more likely to call police on mentally ill persons. Mentally ill persons are 16 times more likely to be killed by police.

Before you go all hand-wavy about this, remember that there are real world consequences.
posted by LindsayIrene at 9:42 AM on August 6, 2020 [28 favorites]


Media depicts mentally ill persons as violent. The general public thinks of mentally ill persons as violent because that's what the media tells them.

I don't think it's as linear as this. Media depicts things the way they do to a large extent because that's how they see them, because they are part of society. They reinforce and perhaps even exaggerate societies harmful stereotypes but so does a lot of the rest of societal interaction, and it takes a significant amount of effort to confront and mitigate your own biases. That said, with media's extensive reach their choices have a larger impact, and it doesn't feel unreasonable to hold them to a higher standard. They certainly put plenty of effort (of a different kind) into a ton of things that don't have important social implications.
posted by aubilenon at 10:40 AM on August 6, 2020


I'm curious about their methodology. How did they identify the mental illnesses? Was it only when it was overtly stated in the text? Or when it was read/assumed by the audience?

American media is overwhelmingly violent, so it stands to reason that all characters (not just mentally ill characters) are more violent than real life.

Mental illness doesn't get a fair shake in general, but part of the problem here is that positive representation is often overlooked because of what positive representation *looks like*. Autistic friends have read Zuko (Avatar TLA) and Entrapta (She-Ra) as autistic, but it's not stated as such in the actual show. It's entirely possible that positive representations are being undersampled.
posted by explosion at 11:39 AM on August 6, 2020 [2 favorites]


I'm curious about their methodology. How did they identify the mental illnesses? Was it only when it was overtly stated in the text? Or when it was read/assumed by the audience?

From the linked PDF:

Drawing from multiple resources, a character was coded as possessing a mental health condition when a significant and/or persistent negative reaction (e.g., adverse thoughts, emotions, behaviors) was evidenced by internalizing or externalizing symptoms. Characters with evidence of major psychiatric conditions such as mood disorders, anxiety disorders, psychotic disorders, addictions and eating disorders were included in this definition, as were suicidal ideation or behaviors intended for selfinjury (though those motivated by political or ideological reasons were excluded). The definition was broad and included substance use disorders, as they may co-occur alongside other conditions.
posted by gusottertrout at 12:39 PM on August 6, 2020 [2 favorites]


Interesting article, that's worth talking about. Thanks for finding and sharing this. I especially liked the Strategic Solutions to Change Storytelling tool. Useful questions.

I have a few counterpoints, though I do agree with the big thesis of the article. Mental health disorders are rarely well depicted and as the authors claim, often stigmatized or trivialized.

1. I read the article but couldn't find a methodology or abstract section. That might be included in another PDF; maybe this is more of an executive summary. Tried to see if there was an academic paper, but it didn't come up on Google Scholar. Despite the use of numbers and percentages we actually have no ideas how these figures were produced. (Again, not dismissing the conclusion, I'm sure it's pretty much correct.)

Although this looks like a quantitative study, it's really qualitative with numbers used somewhat arbitrarily. There were a lot of value judgements made and I'm pretty sure the researchers had a set outcome they wanted. Would be interested to see what sort of tools they developed to ensure consistency/validity. If another group of psychologists replicated this, would they get similar numbers? (I'm only making this point because of the presentation of statistics. Otherwise I won't impose quantitative standards on a qualitative research project.)

I'd also like to see their criteria for inclusion and exclusion. Which movies did they choose? How did they define that population of movies to study? Maybe this was explained but after reading and using ctrl+f, I couldn't determine it.

There are bits of methodology embedded in the end notes, which is strange. They do reference qualitative coding practices, which are great but seems a bit dishonest to attach hard numbers to a qualitative research paper. Of course people do it all the time, but I feel like that's a methodological stretch. I am all for mixed methods research, but I'd like to see more of how the sausage was made.

2. Unfortunately mental illness is often used as a plot point to move the story forward. Many movies are violent and they might imply a mental illness in order for the action to unfold. IE James Bond villains usually fall under maybe 1 of 3-4 archetypes: Foreign Bad Guy (Russian in the 1960-80s etc), Financial Bad Guy (Goldfinger or Le Chiffre) or Mentally Ill Bad Guy (Robert Carlyle or Javier Bardem's characters.) I suppose there is a 4th archetype of rogue agent, but they're often depicted as having PTSD or other traumatic issues so it's a messy venn diagram. I think this could be applied to characters in most action movies, which presumably make up a large portion of the violence being studied.

There is a sort of availability bias in the structure of the research as well. We're obviously only counting people with overt, visible mental illnesses. Just as in real life, there are probably characters with various level of autistic spectrum issues, low level eating disorders, personality disorders, various depressive issues etc that do not pass the threshold of being counted and therefor are giving a false number. The population isn't sufficiently defined so all the numerators and denominators will be off.

Since we don't know their inclusion/exclusion criteria we can't know how some characters were diagnosed. It looks as though TV shows tend to have a higher proportion of characters with mental illness, but that's probably because as the researchers mention, they scoured wikipedia for plot lines and did text searches for references to what I guess was a set list of mental illnesses. What I mean is that there are probably lots of characters who might be written with disorders but the threshold isn't sensitive enough so we get a false negative, they're under counted and only the over top mental illness depictions are counted and then reinforce the thesis without any mechanism for disconfirmation built into the study.

I'm bringing this up because I have friends who work as directors and when they go over the script, they hire mental health experts to create psychological profiles/motivations of the characters to use in the process. Granted, these are usually Freudian psychologists because I guess Freudian models make for better drama. (So it's the same problem all over again maybe.)

3. I think it might be more productive in some ways to try and figure out why so many movies and shows are violent. People want to create a violent product then reach for mental health as a possible plot point to enable this depiction of violence.

Sometimes when people want to make a violent gun movie, they make a Western. If they want melee, make a medieval movie. If they want to depict violence in modern times then they end up leaning on one of several tropes. (Crime, Terrorism, Mental Health etc)

If we made better movies, there'd be less violence over all. In the end, I think this report is a net good, and I'm sure its conclusions are more or less correct, but the methodology seems suspect to me. I think they put the cart before the horse. I'd be wary about confidently quoting any of the stat in here.
posted by Telf at 12:56 PM on August 6, 2020 [3 favorites]


I saw explosion's comment as I was proofreading my own. I definitely agree that the sampling methodology as the the authors explain it does not justify their confident use of numbers.
posted by Telf at 12:59 PM on August 6, 2020


It's interesting to me how much the methodology irks people. I read the findings not as "scientific" but as illustrative of current problems with representation and portrayals, a snapshot from a particular set of media, that provides guideposts for people to think about and change their projects. But I am guessing all the bar and pie charts and BIG FONT%%% gives the vibe of precise statistics? I appreciated their regular reference to the sample sizes, e.g. on page 9
The most frequently portrayed mental health condition on screen was addiction, with 29 characters: Characters were addicted to alcohol (n=15), drugs (e.g., opioids, crack/cocaine, cannabis; n=13), and gambling (n=1) in the sample of movies.
posted by spamandkimchi at 1:28 PM on August 6, 2020 [4 favorites]


I also want to amplify LindsayIrene's comment on how expectations of violence have fueled the terribly high numbers of people with mental illness who are shot and killed by police and other law enforcement. In a 2012 joint report from two Maine newspapers
42 percent of people shot by police since 2000 — and 58 percent of those who died from their injuries — had mental health problems, according to reports from the Maine Attorney General’s Office.
posted by spamandkimchi at 1:37 PM on August 6, 2020 [9 favorites]


I'd like to see somebody put together a list of, well, good portrayals of mental illness on film or on TV? Even when it's sympathetic it tends to be oversimplified, and often a little condescending.
posted by atoxyl at 4:52 PM on August 6, 2020 [4 favorites]


Addiction actually gets better depictions than many conditions, IMO, because there are so many writerly first-person addiction stories. Of course a lot are mired in kind of old-fashioned ideas about treatment.
posted by atoxyl at 4:56 PM on August 6, 2020 [1 favorite]


Mental illness makes a really cheap source of dramatic tension. At its worst it’s just a lazy shortcut around giving a character cogent motives. Mentally ill characters are often just plot spackle.

It’s been years so I couldn’t possibly guess where, but I recall reading an interesting take years ago that inferred the slasher movie craze of the late 70s and early 80s, most especially Halloween, owes its existence mainly to Reagan’s general hostility to mental health care, first as governor of CA and then as President, and the resulting anxiety that people who needed care were being turned out into the community without support.
posted by gelfin at 8:19 AM on August 7, 2020 [1 favorite]


American slasher movies were also inspired by Italian giallo films, which tended to be rather Freudian in their explanations for the killers' motives.
posted by LindsayIrene at 10:14 AM on August 7, 2020


I remember liking Monk when it came out because of the portrayal of a protagonist with OCD, but it quickly got cartoonish. I didn't have a problem with him being the butt of jokes, but the way he acted as the series went on just wasn't very realistic to my experience.

(I like the show anyway for Ted Levine and the really creative murder plots.)
posted by atoxyl at 12:29 PM on August 7, 2020


Could I please ask people posting that "movies are just violent!!!" to stop with the "it's just a trope/plot point"; your point has been made, and it's kind of a shitty one.

Not just because of the very real issue around being shot by police but because it's not that people with mental ilnesses/differences are holding every role in overall violent movies. I mean, be my guest and read this brief roundup of movies featuring characters with DID. You can be the psycho murderer/manipulator male or the wounded crazy female, with United States of Tara or Mr. Robot perhaps teetering vaguely on the edge of something else and then that movie with Halle Berry I guess.

"Mental illness makes a really cheap source of dramatic tension. At its worst it’s just a lazy shortcut around giving a character cogent motives. Mentally ill characters are often just plot spackle."

Substitute other stereotypes in that sentence and see where they get you. I won't type it out, but as someone who has to think about this a lot, it sucks.

Precisely zero of the friends I have made as "mom friends" have any idea that I'm multiple, even though I came out to most of my other friends years and years ago, because there is a huge chance that if I mentioned it people would worry that I would be violent with or abusive towards their kids. Around the schoolyard it's okay to discuss anxiety and depression, maybe OCD or at the outside BDD but you can bet your pants that women with bipolar disorder, BPD, or who are dissociative are hiding it. It is not a benign plot choice.

I will link once again to this paper that discusses much of what some of you are about to say to yourselves about it being a fad or made up any, and point out that DID is found in approximately 1.1%–1.5% of representative community samples but...to look at media there are 6 women cowering in a corner while 8 men are out rampaging, or maybe, if they're being nice, manipulating other men into expressing their testosterone.

And no, it's not "just violence" or "a lazy plot choice," it is a demeaning and harmful stereotype which costs people who:
- don't seek treatment
- are misdiagnosed, on average, FOR YEARS.
- are stigmatized not only in seeking mental health treatment but in seeking medical treatment*
- who have to hide core aspects of themselves
- who force themselves, quite often, to live extremely "normally" so as not to fall into a stereotype either in front of others or indeed, for themselves.

* Fun fact: If you can't feel pain in a medical setting, you can get into incredibly bad health situations really easily, especially if you are taking pains to keep the reason for that off your chart.
posted by warriorqueen at 8:29 PM on August 7, 2020 [7 favorites]


I want to repeat this from the article:

Its really important to realize HOW COMMON mental illness is here is some information from NIMH . 18.9% of the US population suffers from mental illness, which is very close to 1 and 5 adults. 4.5% of the US adult population is thought to have a serious mental illness. Or about 1 in 20 people.

Note the statistics above *The assessment did not contain diagnostic modules assessing: adult attention deficit hyperactivity disorder (ADHD), autism spectrum disorders, schizophrenia or other psychotic disorders (although the assessment included a psychotic symptom screen). And in my obersation (and also because Warriorqueen's post is the last post at the moment) it also mentioned nothing of disassoicative disorders, either in the assessment OR exclusion.
posted by AlexiaSky at 1:41 AM on August 8, 2020 [1 favorite]


« Older A 90 minute interview with IC3PEAK   |   The End Of The NRA? Newer »


This thread has been archived and is closed to new comments