Mimesis and Suicide:
A Review and Suggestions
for Future Research
Dr. Steven Stack, PhD*
Director, Center for the Study of Suicide,
Professor, Department of Psychiatry, &
Department of Criminology, Wayne State
University, Detroit, USA
Conference on Mimetic Factors & Health,
University of Warwick, Coventry, UK. October
[email protected]
Research on Copycat or Imitative
• Between 1967-2009 at least 120 scientific studies were published
on the possible link between media depictions of suicide and
suicides in the real world.
• Most studies measure the suicide rate before and after a widely publicized
news story (or film) on suicide.
• If the change in the suicide rate is positive, a copycat effect is reported.
• Many investigations are based on a single story (e.g., Cheng, Hawton, et al. 2007. Influence of
media reporting of a celebrity suicide. J of Affective Disorders, 103, 69-75.)
Some investigations explore mimesis effects for a large number of 30 or more media
stories (Phillips, D 1974. The Influence of Suggestion on Suicide, Amer Sociological Rev, 39, 340-354; Kessler, et al 1988. Clustering of teenage
suicides after television news stories about suicide: a reconsideration. American J of Psychiatry, 145, 1379-1383).
• In large scale investigations, some of the time, research finds that after
controls for confounding factors (such as seasonal effects, day of the week
effects, unemployment trends, holiday effects) publicized suicide stories are
associated with increases in one or more population suicide rates.
Individual Level Evidence of a Copycat
• At least 8 studies have documented direct copycat effects at the
individual level. They include:
• Marzuk et al (1994) found that suicide by asphyxiation
increased 313% in NYC in the year following publication of
Final Exit, a guide recommending that suicide method.
• In 27.3% of the suicides, a copy of the book was found at the
(Marzuk, P. et al., 1994, “Increase in suicide by asphyxiation in NYC following
the publication of Final Exit,” Publishing Research Quarterly, 10, 62-68.
Simkin et al (1995) determined that 32% of suicide attempters at ER were
aware of a recent suicide on British television, and 14% reported it had
influenced their attempt. (Simkin, s et al 1995 Media influence on parasuicide,
British J. of Psychiatry, 167, 754-759).
Hawton et al (1999) determined that 20% of suicide attempters
reported that a recent suicide on British television had influenced their
attempt (Hawton, et al., 1999. Effects of a drug overdose in a television
drama on presentations to hospital for self poisoning. BMJ, 318, 972977.
Fekete found that 41% of suicide attempters in Hungary reported a
suicidal role model from the media compared to 10% controls (Fekete,
S & Schmidtke, A 1996. Suicidal Models. Omega, 33,3, 233-241.
Individual Level Evidence for & Against
a suicide copycat effect - continued
Cheng et al found that 39% of 438 depressive patients exposed to a media report concerning a
celebrity suicide reported that the media story had an influence on their subsequent suicidal
behavior (Heng, A, Hawton, K. et al 2007. The influence of media reporting of a celebrity
suicide on suicidal behavior in patients with a depressive disorder. J. of Affective Disorders, 103,
Tousignant et al found that 13% of suicides after the highly publicized suicide of a well-known
Quebec were influenced by the reporter’s suicide (Tousignant, Mishara, et al. 2005. The
impact of media coverage of the suicide of a well-known Quebec reporter. Social Science &
Medicine, 60, 1919-1926.
HOWEVER, Mercy found that suicide attempters were actually 80% less apt to report having
been exposed to a media depiction of suicide (a dichotomous conglomerate measure of any
exposure in newspapers, film, books, television, etc) in 30 days prior to the attempt (or
interview) than a sample of community controls (Mercy, J. et al Is suicide contagious?
American J of Epidemiology, 154,2, 120-127).
Classic Aggregate Level Example of
Copycat Effects: Actress Marilyn
Monroe & US National Suicide Count
Phillips (1974) found that there was a 12%
increase in suicide the month following the
widely publicized suicide of actress Marilyn
Note: Suicide is not mentioned in the
headline, conforms to guidelines for media
reporting of WHO (2008).
However, death/suicide is given a positive
definition as freeing the individual from
Photo of her body being wheeled out from
her home, violates a WHO guideline for
reporting (no photos of suicide scene or
body of suicide).
In general, the average increase in suicide
was 2.51% after the 34 Newspaper stories
analyzed (Phillips 1974)
Aggregate Level Findings:
Evidence Against a Copycat Effect
In the largest study to date, Kessler at al. (1988)
analyzed 87 Television News stories on suicide. No
evidence was found for an increase in suicide in 65
out of 69 regression analyses (Kessler et al., 1988,
American J of Psychiatry). (suggests research based
on TV coverage may nullify the copycat effect, for
some reason).
The first systematic meta analysis of 293 findings in
42 studies found that less than half of the findings
supported a copycat effect (Stack, S. 2000, “Media
Impacts on Suicide: A Quantitative Review,” Social
Sci. Q., 81:957-971).
The suicide of “SUPERMAN” George Reeves, for
example, was not associated with a copycat effect.
However, it failed to receive widespread page 1
Plan of this Presentation
• There is no automatic relationship between suicide in the
news/media and copycat suicides
• There is a need to search for the contexts that maximize or
minimize the probability of finding a copycat effect.
• This presentation
– PART I. first reviews theories of media effects which suggest such
moderating variables. Most theoretical leads have not yet been
systematically researched.
– PART II. Reviews the three existing meta analyses which document
several known methodological and theoretical moderators.
– PART III. Policy: reviews research on the WHO (2008) media guidelines
that focus on story moderators thought to decrease the odds of mimetic
effects. The degree of compliance and the effects of compliance are
– PART IV. Eight Suggestions for new research on moderators and other
Key Theoretical
Propositions on Media
Effects on Suicide
(A) Basic Dose-Response
(B) Symbolic Inter-actionism
( C) Social Learning Theory
Dangerous Stories and Receptive
Audience Characteristics
I (A) Dose-Response: Amount of
& Priority of Coverage
• Proposition 1. All else being equal, the greater the coverage (e.g., column
inches, seconds of TV news, number of networks or newspapers carrying the
story) of a suicide story, the greater the copycat effect.
• Proposition 2. The greater the priority of coverage (e.g., page 1, lead story in
TV news) the greater the copycat effect.
• However, The characteristics of the stories and the audience
also need substantial consideration as providing moderators
between sheer amount of coverage and copycat effects.
Sub proposition, yet untested,
Movies where the lead star suicides will generate more copycat suicides than movies where a minor
character suicides.
Movies where the motives for suicide are presented in detail will promote more copycat effects than
heir counterparts
I (B). Symbolic Interactionism:
Audience Receptivity & Mimetic Risk
A basic assumption in media effects research is that some members of the audience are
more receptive to Media advertisements or messages than others (eg., Blumer,H. 1969
Suggestions for the study of mass media effects, Symbolic Interactionism, Prentice Hall;
Stack 2005, SLTB). Media news stories affect mainly already vulnerable or high
risk individuals. They are “on the edge” already.
• Proposition 3. In suicide studies, groups with pre-existing protective
factors protective factors against suicide will be less susceptible to
mimetic effects.
– Lack of pre existing psychiatric disorders predictive of suicide
– Social protective factors including high religiosity, social support, economic security &
• Proposition 4. Groups with pre-existing suicide risk factors will be more
susceptible to mimetic effects.
– Psychiatric risk factors including major depression & substance abuse disorders
– Social risk factors such as economic strain (unemployment, home evictions), relationship
strain (e.g., divorce; domestic violence); social isolation (e.g., living alone), death of a loved
one, low religiosity, & altruism (suicide to relieve others from your being a perceived burden).
– Physical illness and disabilities (e.g., dying of cancer, disabled from car accident, blind)
– Comorbidity: risk increases with presence of multiple risk factors
e.g., Major Depression
Economic Strain
Youth more impulsive
e.g. Celebrities
Front Page
WHO Guidelines
I ( C). Social Learning Theory
Story Characteristics:
Positive vs. Negative Definition of Suicide
• Suicide Guidelines and most work on media effects to
date have focused more on story characteristics than
audience characteristics in mapping media effects.
• Media depictions of suicide are a fundamental
component of social learning processes where
individuals learn vicariously through watching the
experiences of others, as opposed to limiting their
learning solely to their own personal experience.
• Media portrayals of suicide can promote the learning
of both positive and negative definitions of suicide (e.g.,
Pirkis & Blood 2001a,2001b; Schmidtke & Schaller 2000; Stack 2000).
Positive Definitions of Suicide in the Media
Positive Definitions of Suicide may include:
Social Rationalizations or motives for suicide including ill health, marital
trouble, job loss, romantic strains
Individual Rationalizations including suicide as a implied solution to a battle
with life long battle with depression, substance abuse
Perceived Rewards. If suicidal persons desire to hurt significant others,
grieving relatives of suicide might comprise a “reward” for the behavior.
Focus on the positive aspects of victim’s life- e.g., attractive photo, high
school sports hero, honor student, father of 4, wealthy person.
Glorification of Deceased- e.g., fly the school flag at half mast
Long Term and sensational coverage- captures attention, attention getting.
Negative Definitions of Suicide
Negative Definitions of Suicide include:
– Focusing on victim’s disfigurement/pain (e.g. pictures of
rotting bodies at the Jonestown mass suicide in 1978).
– Stressing that suicide is wrong, wasteful, even stupid (e.g.
coverage involving Kurt Cobain’s widow after the rock star’s
suicide in 1994)
– Discussing alternatives/solutions to suicide (e.g., crisis hotlines,
Example: media coverage of grunge rock celebrity Kurt Cobain was often quite negative and did not spark
a blip up in the suicide rate (Jobes et al., 1996, The Kurt Cobain suicide crisis, Suicide & Life Threatening
Behavior, 26, 260-271).
Note on Suicide & Media History: The editor of the NYT thought that publicizing every single suicide in the
NYT would discourage suicide by shaming the suicide victims. After a period of hyper media coverage
during 1913-1914, he abandoned his idea since the suicide rate did not go down as he had hoped (Gundlach,
J & Stack, S 1990. The effect of hyper media coverage on suicide, New York City, 1910-1920. Social Science
Quarterly, 71,3, 619-627. Sheer coverage is not in itself a negative definition.
• Proposition 5. The greater the excess of positive over
negative definitions of suicide in the news story, the
greater the chance for finding a copycat effect.
However, media coverage of suicide has typically covered far more neutral/factual
and/or positive definitions of suicide than negative ones (e.g., Blood 2001, Fekete,
Schmidtke et al, 1998, Michel et al., 2000; for a review see Schmidtke & Schaller 2000:
Researchers studying a large number of stories often just lump them all together
without searching for negative definitions. Strong negative definitions are uncommon.
Auxiliary, sub component propositions:
• Proposition 5A The greater the negative definitions of suicide in a
news story will minimize the odds of finding a copycat effect.
• Proposition 5B. The greater the positive definitions of suicide, the
greater the copycat effect.
Social Learning, Story Characteristics:
Differential Identification Theory
• Since the work of Tarde (1903) in the Laws of Imitation, there has been a
proposition that the inferior tend to copy the behavior of the superior.
• Proposition 6. ( Vertical Identification) Persons will
tend to copy the suicidal behavior of superior people
(e.g., well known celebrities) more than the suicidal
behavior of common people. (Tarde, G. 1903. The Laws of Imitation, NY: Henry Holt.
Wasserman, Ira. 1984, Imitation & Suicide: A Reexamination of the Werther Effect, American Sociological Review 49, 427436. Stack,S. 1987, Celebrities & Suicide: A Taxonomy & Analysis. American Sociological Review, 52, 401-412.)
In vertical identification, persons high in individual and/or social risk factors for
suicide, would presumably be more apt to think “if Marilyn Monroe with all her
beauty, fame, & money chooses death over life, why should I (possibly ugly,
unknown, & poor) go on living? “ than if they hear about some ordinary person
As we shall see, widely publicized celebrity suicides are the most likely ones to
apparently trigger copycat suicides.
Horizontal Identification. In this view,
Proposition 7. Persons will be more apt to copy the
behavior of others like themselves than persons unlike
Dimensions of horizontal identification may include:
Demographic Identification:
– AGE, GENDER, Nationality, Marital Status (e.g., males more be more apt to copycat
suicides of males than those of females)
Problem or Motive Centered Identification:
– Individual Stressors. Physical & Mental Illness. Persons dying of cancer may be most apt
to copycat highly publicized suicides of similar persons.
– Social Stress (e.g., unemployed persons may be most apt to copycat the suicides of other
unemployed persons; divorced persons may copy cat the media-based stories of divorced
persons who suicide).
– Note: Vertical and Horizontal identification are not mutually exclusive but can occur
together (e.g., Marilyn Monroe was a celebrity, but also had long term marital problems,
substance abuse problems that serve as points of horizontal identification).
– Note: there is ample evidence that both celebrity and noncelebrity publicized suicide are
often associated with increases in the social suicide ate (e.g., Stack, S. 1990 A re-analysis of
the impact of noncelebrity suicides, Social Psychiatry & Psychiatric Epidemiology, 25, 269273).can
There are three quantitative meta analyses of the research on copycat
These suggest some of the key predictors or moderators that promote
finding a link (or not finding a link) between stories and social suicide
However, most propositions suggested from theories of media effects
have not been systematically tested.
42 Studies
Stack, S. 2000, “Media impacts on suicide,” Social Science
Quarterly, 81,4, 957-971.)
• Logistic Regression Analysis 293 findings from 42 studies. 14 on Fictional & 28
on real stories.
• where 1=copycat effect noted in a finding, 0=lack of a copycat effect.
• Key significant methodological/theoretical predictors of reporting a copycat
effect included
– celebrity status,
– real vs. fictional stories,
– media type (TV vs. Newspapers).
Dose-Response Theory: Television
vs. Newspaper Stories (OR=0.21)
• Research findings based on television news
stories on suicide are 79% less apt to
report an imitative effect than ones based
on newspaper stories..
• Relates to the amount of coverage. TV stories usually
last less then 20 seconds, can go unnoticed, easily
forgotten. Newspaper stories are much longer.
• Also, Newspaper stories, unlike TV ones, can be easily
saved, & re-read.
Differential Identification Theory:
Entertainment/Political Celebrity Status,
• Studies measuring the presence of either an
entertainment or a political celebrity were fully
14.3 times more apt than studies based on
other kinds of stories to report a copycat effect.
• The limited work on other celebrity subtypes (e.g., celebrity villains
such as spies or mass murderers who suicide; foreign celebrities;
economic celebrities) suggests that these other categories of celebrity
suicide do not have a copycat effect.
• The mass audience apparently identifies with entertainment
celebrities, well known movie stars, in particular. Stars represent the
celebration of the ordinary (Stack, S 1987. Celebrities and Suicide,
Amer Sociological Review, 52, 401-412; Dyer, R 1979. Stars. London:
British Film Institute)..
Differential Identification Theory:
Real vs. Fictional Suicide Stories,
• Research based on real suicide stories (as opposed to
fictional stories) was 4.03 times more likely to report
a copycat effect than research based on fictional
suicides such as those in movies.
There is a general agreement that research based on real suicide stories is more apt
to report copycat effects than research based on make-believe or fictional stories
(e.g., suicide in movies & soap operas). Gould, M. 2001. “Suicide and Media,”
Annals of the NY Academy of Science. Pirkis & Blood 2001a. “Suicide & the Media
Part I, Crisis. Pirkis & Blood 2001b. “Suicide & the Media Part II, Crisis.
Schmidtke & Schaller (2000) “Role of the mass media in suicide prevention,” Int’l
Handbook. of Suicide & Suicide Prevention, NY: John Wiley.
However, this finding may be a function of poor measurement of exposure to
fictional media such as film. The measurement of cumulative exposure to films
containing suicides of stars has not been measured.
55 Studies on Non
Fictional Stories
Stack, S. 2005. Suicide in the Media: A Quantitative Review of
Studies Based on Nonfictional Stories. Suicide & Life
Threatening Behavior. 35,2 121-132.
Similar methodology to Stack (2000)
Expands sample of studies focused on non fictional (REAL) suicides from 28 in
Stack (2000) to 55.
The 55 studies (by first author)
Ashton 1981
Baron 1985
Blumen ‘73
Bollen ‘82
Etzerdor ‘01
Jonas ‘92
Sacks ’92
Stack ’87
Stack ’89b
Stack ’90c
Stack ’90d
Stack ‘91
Stack ‘92
Wasserman’84 Wass.’92
Dose-Response Theory: Amount of
• Findings based on televised suicide stories were 79%
less likely than findings based on newspaper stories
to report a copycat effect.
• Page 1 newspaper articles are longer, have more
detail. Can be easily re-read and copies easily saved
for future reference.
• TV coverage typically last about 10 seconds.
• Some evidence that research based on one network
coverage (as opposed to multiple network coverage
that reaches a larger audience) has no impact at all
on suicide rates.
Differential Identification Theory:
Entertainment/Political Celebrity Stories,
• Findings based on
celebrity suicide news stories
were 5.27 times more likely
than findings based on other
stories to report a copycat
• The OR here is smaller than in
Stack (2000) since fictional
stories are left out (they have
no celebrity suicides).
Social Learning Theory, Negative
Definitions of Suicide (OR=.01)
• Findings based on suicide stories
stressing negative definitions of suicide
were 99% less likely than their
counterparts to report a copycat effect.
• E.g., The 12 days of coverage of the rotting
bodies, cult-like mass suicide in Jonestown,
was actually associated with a drop in suicide,
a drop of 58 suicides (Stack 1989, The effect of Jonestown
on suicide, Pp. 135-151, Rebecca Moore, New Religious
Movements, Mass Suicide, & People’s Temple, NY: Edwin
Cobain’s widow defined his suicide as a total waste, thereby
decreasing the odds of imitation.
• Audience Receptiveness: Findings based on the rate of
female suicide were 4.89 times more likely to report a
copycat effect than findings based on male suicide.
• Model Fit: 77.3% of the 419 findings were correctly
classified by the model.
Summary: Odds that a study will
report a copycat effect by
methodological Issue
TV stories
Neg defs
26 Studies on Fictional
Stories (e.g., Movies)
Stack, S. 2009. Copycat effects of fictional suicide: A Meta Analysis. Chapter 17 in Stack, S
& Lester, D. (eds), Suicide and the Creative Arts. NY: Nova Science.
Expands analysis in Stack (2000) from 14 to 26 investigations.
26 studies contain 146 findings.
Matching of Model’s Method with
Audience Method of Suicide, OR=4.40
Studies that matched the method of suicide in
the role model (usually an actor in a movie)
with that of the audience were 4.4 more
times apt to report a copycat effect than
Studies that did not do so.
Tip: Suicide of super
star Bruce Willis, Sin
City (2005)
An Unstudied image
It is often not clear if the model’s suicide
Had a significant effect on suicides by other
Possibly there was no overall increase in suicide
But a transference in methods.
However, at the aggregate level, Jamieson (2003) reports that
the Number of suicides in film/year explained the upswing in
Overall youth suicide between 1950-2000.
Jamieson, P. 2003 Changes in US Popular Culture Portrayal
of Youth Suicide, 1950-2000. PhD dissertation. Philadelphia:
U Penn.
Audience Receptivity: Youth
Suicide, OR=4.39
• Research findings based on youth suicide rates
are 4.39 times more apt to report copycat effects
than research findings based on other suicide
Consistent with the view that youth are more susceptible to suggestion
• Consistent with the view that youth are the most apt age group to watch
• Tip: There is no research to date assessing the characteristics of he model- their
motives, star status, etc.
• Tip: There is research yet on the reactions of the survivors to the suicide.
• Tip: There is no research to date measuring cumulative exposure to suicide
Part III:
Guidelines for Media Reporting
(WHO 2008)
Promoting vs. Preventing Copycat
Media Guidelines for the
reporting of Suicide
• Since the mid 1990’s various organizations
have produced guidelines for the reporting of
• Most Recent: World Health Organization &
Int’l Association for Suicide Prevention (2008)
issued Preventing Suicide: A Resource for Media
Professionals. Available at:
The Eleven WHO(2008) Media
1. Take the opportunity to educate the public about suicide
2. Avoid language that sensationalizes or normalizes suicide, or presents it as a solution
to problems
3. Avoid prominent placement & undue repetition of stories about suicide
4. Avoid explicit description of the method used in a completed or attempted suicide
5. Avoid providing detailed information about the site of a completed or attempted
6. Word headlines carefully
7. Exercise caution in using photographs or film footage
8. Take particular care in reporting celebrity suicides
9. Show due consideration for people bereaved by suicide
10. Provide information on where to seek help
11. Recognize that media professionals themselves may be affected by stories about
World Health Organization & Int’l Association for Suicide Prevention (2008) issued Preventing Suicide: A Resource for Media Professionals. Available at:
Suicide Prevention & Media:
Two Issues: Compliance & Effectiveness
• (1) Compliance with the Guidelines. Can the
Media be convinced to change the amount
and/or type of their reporting on suicide/ to
conform to the WHO Guidelines?
• (2) Do the Guidelines really work? Are the
criteria in media guidelines really predictive of
suicide, independent of the amount of coverage
(Phillips 1989). Or is decreasing the sheer
amount of coverage, of any kind, the main way
to prevent imitation?
1. Compliance. Have the Media
Followed the Guidelines?
Guidelines have been disseminated by many
organizations including WHO, CDC, AFSP, & AAS
(e.g., Satcher, 2001, National Strategy for Suicide
Prevention, 106-111).
• Cross sectional research at one point in time
often suggests a gap between elements of the
guidelines and actual press coverage.
• An Australian study of 410 press reports
determined that only 6.5% provided
information on where to seek help, e.g., suicide
hotline number as recommended in WHO
(2008) guideline #10 (Pirkis 2002).
Compliance. Evidence from
A Swiss study explored the impact of a 1992 press conference
between the media, Swiss Medical Association, &
Suicidologists. The media were urged to follow Guidelines for
reporting suicide. (Michel, Frey, Wyss & Valch 2000, “An Exercise in
Improving Suicide News Reporting in the Media,” Crisis, 21:71-79).
• A content analysis of stories over an eight month period before
and after distribution of the Guidelines illustrated a substantial
change in the quality of reporting.
• However, at the same time, the quantity of reporting on suicide
tripled from 151 to 468 stories. The media officials apparently
ignored this key guideline.
Percent of stories containing presumed dangerous
content before (1991) & after (1994) Dissemination
of Media Guidelines, Switzerland (Michel et al., 2000)
Page 1 Stories
Sensational Text
N of Stories
Australia: Degree of Compliance of the Australian Media with
WHO Media Guidelines (Pirkis et al 2002. Reporting of Suicide in the Australian media. (Australian & New
Zealand J of Psychiatry, 36, 190-197).
WHO Guideline
(2) Avoid Inappropriate language/sensationalism
(3) Avoid Prominent Placement of Story
(4) Avoid Detailed discussion of method of suicide
(6) Avoid word “suicide” in headline
(7) Avoid image of suicide scene/location/method
(8) Limit Celebrity Coverage (%who are not celebrities)
(9) Sensitivity to Bereaved (% NOT interviewed)
(10) Info on where to seek help
Other: Suicide Linked to Mental Disorder
2. Do Media Guidelines Actually
It is widely believed that the type of presentation
of suicide in news stories influences the
incidence of copycat suicide.
However, only one previous rigorous study to date
tests associations between selected aspects of
media guidelines and copycat suicide (Phillips, David
P., Carstensen, L. & Paight, D., 1989 “Effects of mass media news stories on
suicide with new evidence on story content,” Pp. 101-116 in Peiffer, Cynthia
(ed), Suicide Among Youth: Perspectives on Risk & Prevention. Washington,
D.C.: American Psychiatric Press.
Media Guidelines & Teen Suicide:
An Inconvenient Truth?
• Phillips, et al. (1989) studied five characteristics
of 32 televised suicide news stories & their
impact on teen suicide in the US.
• Controlling for the amount of coverage of the
story (seconds of TV time), none of the five
characteristics of the story actually affected the
incidence of teen suicide.
• (Phillips, David P., Carstensen, L. & Paight, D., 1989 “Effects of mass media
news stories on suicide with new evidence on story content,” Pp. 101-116 in
Peiffer, Cynthia (ed), Suicide Among Youth: Perspectives on Risk &
Prevention. Washington, D.C.: American Psychiatric Press.
T-tests for effect of Five Story Characteristics on
Teen Suicide, controlling for the amount of
(Phillips et al., 1989)
Motive stated vs. 1.41
Method Reported -0.84
Probability Level
Photo of model
“definite” suicide -1.21
Mention of
Family & friends
Dose Response Theory & Tips
• The sheer amount of coverage was a significant predictor
of teen suicide, thus supporting the classic dose-response
theory of imitation effects.
• Tip: Results controlled out the influence of the amount of
coverage. Without this control some story presentation
variables may have been significant.
• Tip: Phillips did not measure six aspects of today’s suicide
guidelines (e.g., sensationalism). Future work should do
• Tip: These findings refer only for teenagers. Possibly other
age groups would respond to dangerous story content.
Eight Suggestions for Future
Research on Suicide and Mimesis
(a) Cumulative Non laboratory Exposure Index: Example of Smoking Initiation Research & Film
(b) Motive-Identification & Feature Films
(c) Long Term Mimesis Effects: Deer Hunter
(d) What explains the Variation in Compliance with Media Guidelines?
(e) Large Sample Aggregate Level Research: little since 1990’s
(f) Work on the elderly as a vulnerable group
(g) Suicide in Popular World Literature and Copycat Effects
(h) Age/Gender matching of model and Audience
IVa. Cumulative, Non Laboratory Index of
Media Exposure: Smoking Initiation
• Some of the best work on mimesis and risk behavior has been done in smoking
initiation research.
• Therein researchers developed a new measure of exposure to media messages
regarding smoking: the number of films which respondents reported having
watched where the stars of the film smoked.
• A meta analysis of 40 studies concluded that the greater the number of such
films watched the greater the likelihood that teens imitated smoking at followup (Charlesworth & Glantz, 2005, “Smoking in the movies increases adolescent smoking”, Pediatrics, 116:15161528).
Importantly, in some of the best designed studies, such cumulative exposure to film was the most
single important predictor of teen smoking initiation. The film exposure index was more predictive
of smoking than traditional predictors including peer smoking, parental smoking, GPA in school,
and others. (Dalton et al., 2003, “Effect of Viewing smoking in the movies on adolescent smoking initiation: a cohort study,” Lancet, 362:281-285.
The strong results of this body of research resulted in Walt Disney Productions decision to take
smoking behavior out of their films (Marr, M. 2007 Family-Friendly Disney studio turns silver screen into ‘no smoking area,’” Wall Street Journal,
July 26, 2007, p. B-2).
Media work on suicide has not yet adopted this lead. Much work on film’s effects is still done in the lab, and, as
such, highly suicidal youth, the most likely to be responsive to suicide films, are often screened out of the study
for their own protection. This can minimize the chances of finding copycat effects.
IVb. Film Suicides: Motive-Identification
with Star Suicides in Film
• Watching movies is the number one leisure time pursuit in the
US. As such suicides in film may have an important impact, if
measured properly. Recently I have finished analyzing 1,100
suicides appearing in American feature films from 1900-2009.
• Seven broad categories of motives for suicide appear in the films.
• Possibly different audiences identify with some motives more than
Work will be needed to ascertain the extent to which, if any, various audiences (sex/age/race)
differentially identify with which of these 7 motive categories.
I suggest that the suicides of major film stars of each generation (e.g., James Cagney, Clint
Eastwood, Ben Kingsley, Kris Kristofferson) might be especially apt to foster identification and
mimesis among persons of that generation.
Examples of the Seven Suicide Motives in American Film
Stack & Bowman (2011) Suicide Movies: Social Patterns, 1900-2009.
Psychiatric/addiction Kris Kristofferson, A Star is
Born. (1976)
Economic Strain: John Travolta,
Mad City (1997)
Social strain, Rejection in Love,
Officer & a Gentleman (1973)
Terminal Illness: John Wayne,
suicide by outlaw, The Shootist
Death/grieving: Ben Kingsley, House
of Sand & Fog (2004)
Psychiatric/ The Psychopath:
James Cagney White Heat (1949)
Altruism: Bruce Willis, Sin City (2005)
Altruism: Clint Eastwood, Gran Torino
IVc. Long Term Effects of Specialized Media
Exposure: Deer Hunter, Assisted Suicide
• There is almost no work on the possible long term effects of media depictions of
suicide, although time series analyses of large samples usually finds the largest
copycat effect within ten days of a publicized suicide.
• The film Deer Hunter, which was released in 1978, influenced at least 33
copycat suicides during 1980-1985, and additional suicides after its release in
DVD format in 2002 (Coleman, L. 2004 The Copycat Effect. New York:
• Frei et al (2003) reported that the highly publicized assisted suicide (sponsored
by the right to die organization, EXIT) of a prominent couple in Switzerland
affected assisted suicides for two years (Frei < a. et al 2003. The Werther effect and assisted
suicide. Suicide & Life Threatening Behavior, 33,2, 191-200).
It is likely that the suicide images and stories from the narratives in film, books, and music contribute to some
suicides over the long term. The nature & degree of this association need exploration. A first step is for
suicidologists to compile an exhaustive list of suicide depictions in popular cultural media.
IVd. Have/ How Much have the WHO Media
Guidelines Affected the Content of Suicide
Reporting, & What Explains Cross-National
• Suicide prevention strategies generally have included a media
prevention component.
• Suicide guidelines have been available for 15 years. It is not very
clear what impact, if any, they have had in most nations world.
• Are news reports, films, and other media presentations of suicide
any different today than the pre guidelines era?
• Evidence based research from Austria, Australia, Hong Kong,
New Zealand, and Switzerland is mixed. The variance in media
compliance with WHO guidelines needs an explanation.
• For example a survey of 57 journalists in the US found that most
were not aware of the Suicide Guidelines for the Media (Jamieson,
Jamieson, & Romer 2003, The responsible reporting of suicide in print journalism, American Behavioral
Scientist, 46,12, 1643—1660).
IVe. Need for Large Sample
Aggregate Research
The three meta analyses of over 100 research studies have concluded that most
research findings do not support a mimesis effect. However, many studies continue
to explore mimesis using a single media depiction. This is dangerous to
generalizability. Most works on mimesis and suicide that I have reviewed for
journals over the past five years and most work published over the last 13 years
focus on the highly publicized news coverage of a single suicide.
Large samples (20-90) of depictions and audiences are needed to explore the
conditions under which a story is apt to trigger a copy cat effect.
This may be facilitated by aggregate level research of a large number of stories,
large number of audiences (by age/sex groups) over a long period of time.
Such work largely ceased after the publication of seminal works by Phillips (1974),
Wasserman (1984), Kessler et al (1988) and a series of papers by Stack (1987
through 1996).
Such work is extremely time-consuming, but without it, we may alternatively
overlook or exaggerate the mimesis effect, and fail to discover key moderating
IVf. A Rather Neglected Target
Audience: The Elderly & the Coming
Surge in Elderly Suicide
• The work on mimesis and suicide has focused alternatively on either the
total suicide rate or the youth suicide rate, youth comprising a group
thought to be most susceptible to imitation.
• Work is needed to explore the links, if any, between media exposure and
suicide in mid life and old age.
A study by Stack (1990) determined that the elderly were especially vulnerable to
suicide when stories of suicide among the elderly were widely publicized (e.g.,
elderly ex film star Charles Boyer who suicided the day after his wife of many
years died). The elderly suicide increase after general suicide stories was 10 per
month, but 19/month when the stories concerned elderly suicides (Stack, S. 1990
Audience receptiveness, the media, and aged suicide, 1968-1990, J. of Aging
Studies, 4,2, 195-209).
The baby boom cohort is turning 62. The %over 65 is expected to double in the
US & elsewhere over the next 20 years. The social security safety net for this
generation is expected to weaken, thereby increasing their vulnerability to suicide
and media impacts.
IVg. The Impact of the Literature
Curriculum on Copycat Suicide
Recent analysis of 239 widely read popular works of
world literature determined that 25.5% contained one
or more suicides (Stack S., 2009, Suicide Motive in 61 Works of
Popular World Literature with a comparison to feature films, Chapter 8
in Stack & Lester, (eds), Suicide and the Creative Arts, New York:
Nova Science.
It is essentially unknown to what extent the mass
population is affected by exposure to the suicides in
such works as Shakespeare’s plays, Greek tragedies and
those in 19th and 20th century widely read novels.
There are substantially more suicides in the works read
in the American English curriculum than appear in
films (less than 10% of American films have a suicide).
In comparison, there are, at best, only a few nationally
publicized suicides in the news each year over the
period 1900-1983. (Stack, 1987; 2000).
IVh. Age/Gender Matching of
Model with Audience
• Nearly all the research on copycat effects does not match the
age/gender of the model with the age/gender of the audience.
• Much stronger copycat effects are often found in studies that
employ audience- model matching.
• Schmidtke & Hafner (1988) analyzed a the effects of a
suicide in the movie, Death of a Student, which concerned
the suicide of a 19 year old male. This story was associated
– 86% increase for males 15-29
– 147% increase for males 15-19
• Hence, it is likely that some existing research has
underestimated the degree of copy cats effects.
Thanks for the
Opportunity to Let me
Share the Results of
My Research &
Suggestions for more
• Additional Suggestions for New Research
• Evidence on the vitality of Cumulative
exposure index applications
• Other
Appendix: More on the Significance of
Cumulative Index of Exposure to Movies:
Predicting Teen Drinking Initiation:
An overtime investigation of 4,655 adolescents
determined that the greater the exposure to
movies with drinking (from a list of 600
popular movies), the greater the odds of
initiation of drinking by follow up, controlling
for sex, personality characteristics, parenting
style, and other covariates, school performance,
and parental socio-economic status (Sargent, et al, 2006,
“Alcohol use in motion pictures and its relation with early onset teen drinking,” J of
Studies of Alcohol, 67:54-65).
Appendix. Multi-Media
Cumulative Exposure Indexes
• Essentially all 120 plus studies on mimesis and suicide explore just one media
modality either the news papers, or television news, or film, or a book, or
• Research is needed that uses a multimedia measure of suicide exposure in the
• It is possible that research focusing on just one type of media may
underestimate the full impact of media exposure as a whole.
• Individual level work could survey youth, and other age groups, on their past
exposure to suicide in a variety of media and the extent to which these
exposures have influenced them. Descriptive, preferably exhaustive, lists of key
examples of suicide in film (top 50/year), music (top 50/year), books (read in
school curriculums), and other media are needed as a first step in such
research. Some such lists are reviewed in Stack and Lester (2009) Suicide and
the Creative Arts, New York: Nova Science.
Appendix. Music AND Mimesis: An Application of
the Cumulative Exposure Index.
• Film and Music and the most widely consumed art forms.
• Little is known about the extent to which cumulative exposure to
songs with suicidal themes contribute to copycat effects.
• A cumulative exposure index to “favorite” suicidal songs may
constitute a risk factor for suicide (e.g., classics such as Fade to
Black, Suicide Solution, Whiskey Lullaby).
• A list of popular songs with suicidal content would form the
starting point for such work. Lists (with lyrics) can be constructed
from music WWW sites.
• Hypothesis: the greater number of favorite songs with suicidal
content, the higher the suicidality of people.
• Aggregate Work on country radio market share and white suicide
rates documented an association between musical preference rates
and suicide (Stack, S & Gundlach 1992. The effect of country
music on suicide. Social Forces, 71,1, 211-218).
Appendix. Macro Level Audience
Risk Factors: Period Effects
• An almost unexplored area in media and suicide research is the possible
impact of macro level audience characteristics on copycat effects.
• E.g., periods with extremely high unemployment rates such as the Great
Depression (1930-1939) and the current world recession (2008-2009) might be
marked by a stronger association between highly publicized suicide stories
and copycat effects due to a high level of unemployment/economic strain in
the audience.
• Research done in periods low in macro level risk factors (prosperity, low
divorce rates) may be less apt to uncover any copycat effects due to lower
audience receptiveness.
• Periods with weakened levels of protective factors may promote stronger
copycat effects. Religiosity levels among youth in the US declined by about
half (church attendance) in the last half of the 20th century, making them,
perhaps, more prone to copycat effects. Jamieson (2003) determined that
there was a significant association between the number of suicides in popular
films and the rise in teen suicide. This may be, in part, conditioned by a
period effect marked by the weakening of religious integration of youth.

Media Effects on Suicide: A Meta