Journal of Affective Disorders 134 (2011) 488496

Contents lists available at ScienceDirect

Journal of Affective Disorders

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j a d

Preliminary communication

The Werther effect reconsidered in light of psychological vulnerabilities:

Results of a pilot study
Louise Pouliot , Brian L. Mishara, Ral labelle
Center for Research and Intervention on Suicide and Euthanasia, Department of Psychology, Universit du Qubec Montral, Qubec, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Background: Findings from three decades of epidemiological studies suggest that media
Received 12 January 2011 diffusion of stories about suicide is related to increases in suicidal behaviours in the population
Accepted 27 April 2011 exposed to the media reports. However, we still know little about the psychological processes
Available online 31 May 2011 and personal vulnerabilities that prompt some people to engage in suicidal behaviours after
exposure to media presentations of suicides. This cross-sectional study explored the possible
Keywords: impact of exposure to film suicide in normal young people.
Media Methods: Undergraduates from a university (mean age 23 years) completed a questionnaire on
Suicide exposure to suicide portrayal in ctional lms, in which assessment of negative emotional and
Vulnerability
cognitive reactions resulting from exposure, as well as emotional reactivity, dissociation,
Cognition
Emotion
thought suppression, and suicidal tendencies were made.
Werther effect Results: Of the 101 participants, 70% reported being distressed by the portrayal of a suicide in a
fictional film. Among those, 33% stated they felt distressed about the portrayal for several days
to several weeks. The majority of the affected participants (71%) indicated having been
mentally preoccupied for some time by the portrayal and experienced intrusive memories
(68%). Emotional reactivity and dissociation tendencies were significant predictors of the
negative reactions to the suicide film they viewed. Participants who reported that the idea had
crossed their mind to imitate the suicidal protagonist in the film were 3.45 times more likely to
be suicidal and tended to present higher dissociation and thought suppression propensities
compared to those who did not report these thoughts.
Limitations: The results showing possible influences of suicide portrayal in fictional film on
suicide related cognitions were based on a survey methodology.
Conclusion: Results suggest that fictional suicide portrayals in the media may have a deleterious
impact on viewers, and such impacts do not appear to be limited to people having a clinical
profile of mental disorders, as previously assumed by researchers in the field.
2011 Elsevier B.V. All rights reserved.

1. Introduction

This study examined in a direct manner the affective and

This paper is based upon an oral presentation at the 13th European
Symposium on Suicide and Suicidal Behaviour held in Rome in 2010.
Corresponding author at: Center for Research and Intervention on
Suicide and Euthanasia, Universit du Qubec Montral, C.P. 8888, Succ.
Centre-Ville, Montral (Qubec), Canada, H3C 3P8.
E-mail address: pouliot.louise@uqam.ca (L. Pouliot).
cognitive inuences of media portrayals of suicides in lms
and television on audiences as well as the role of psycho-
logical vulnerabilities in the impact of exposure to suicide in
these media. A search using the web based Internet Movie
Data base (IMDb, 2010) with the key word suicide
appearing in the synopsis of lms and television series
identied more than 4000 results. Therefore, audience

0165-0327/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jad.2011.04.050
 L. Pouliot et al. / Journal of Affective Disorders 134 (2011) 488496
members in the general population often nd themselves
exposed to a suicide story diffused by electronic mass media.
Over the past three decades, numerous studies have
reported an association between mass media diffusion of
suicides and a subsequent increase in the number of
suicides in the general population (Berman, 1988; Bollen
and Phillips, 1982; Etzersdorfer et al., 2004; Fekete and
Macsai, 1990; Frei et al., 2003; Gould et al., 1988; Gould and
Shaffer, 1986; Hassan, 1995; Jonas, 1992; Niederkrotentha-
ler et al., 2010; Niederkrotenthaler et al., 2009; Phillips,
1974; Phillips and Carstensen, 1986; Platt, 1987; Romer et
al., 2006; Schmidtke and Hfner, 1988; Schmidtke and
Schaller, 1998; Stack, 1996; Tousignant et al., 2005). The
association is more frequently observed in the context of
studies examining press media reporting of real suicide
cases than with other types of media or ctitious cases
(Pirkis and Blood, 2001a, Pirkis and Blood, 2001b; Stack,
2003; Westerlund et al., 2009).
Phillips (1974), who was among the rst researchers to
document the phenomenon, used the term, Werther effect
to describe this relationship, in reference to Goethe's (1774
1968) novel telling the story of the young man Werther, who
ended his life after an unhappy love affair. Goethe's novel had
been banned in Europe on the grounds of its putative link
with a series of suicides (Thorson and berg, 2003).
Currently, few studies provide proof that individuals de-
ceased by suicide after media diffusion of the suicide story
had actually been exposed to the stimulus (for discussion of
this see Hawton and Williams, 2001; Pirkis and Blood, 2001a,
2001b; Pouliot and Tousignant, 2010). Only a restricted
number of studies have offered such a proof (i.e. Cheng et al.,
2007a, 2007b; Tousignant et al., 2005; Yip et al., 2006). In
addition, these studies, because of their epidemiological
ecological approach, have been unable to increase our
understanding of the underlying psychological mechanisms
involved in the inuence of mass media on suicide (Pouliot
and Tousignant, 2010). One may assume that the exploration
of proximal psychological experiences provoked by suicide
stories presented in media could help in our understanding of
the psychological processes involved in this phenomenon.
Results showing a positive association between suicide
portrayals in the media and increased rates of suicide in the
population are commonly interpreted, on a post-hoc basis, as a
process of imitation. It is further assumed that the persons who
are affected by media depictions of suicide are already
vulnerable and at risk. Many researchers who evoke the
imitation hypothesis use Bandura's Social Learning Theory to
explain the effect. Bandura (1977, 1986) postulates that under
certain circumstances a perceiver observing the behaviour of an
individual may be prone to imitate or engage in a similar
behaviour. Researchers, who hypothesise that individuals with
personal vulnerabilities to suicide are those who are likely to be
at greater risk to suicide after exposure to media suicides,
propose this hypothesis with empirical substantiation (Chen
et al., 2010; Chiu et al., 2007; Doron et al., 1998; Fekete and
Schmidtke, 1996; Zahl and Hawton, 2004). The current state of
knowledge about the role of personal vulnerabilities in media
related suicides is limited and results are divergent (Gould,
2001). However, results from a restricted number of studies
indicate that individuals with a history of suicide attempts or
suicidal ideation appear to be more vulnerable to engaging in
489
suicidal behaviour following their exposure to a suicide story in
the media (Chen et al., 2010; Cheng et al., 2007a; Fekete and
Schmidtke, 1996; Martin, 1996; Zahl and Hawton, 2004).
Studies have consistently found that gender is an important
predictor of suicidal behaviour. Results from studies conducted
in North America and most of the world, suggest that females
are more likely to attempt suicide and report suicidal ideation,
but are less likely to die of a suicide than males (Canetto, 2008;
Mishara and Tousignant, 2004). Given that suicidal behaviour
patterns are gender related, this has been examined in the
realm of the Werther phenomenon. Pirkis et al's. (2006)
research, where the relationships between 4635 media
portrayals of suicide and subsequent suicidal behaviours in
both genders were analysed, revealed that 39% of media items
were followed by an increase in male suicides and 31% by an
increase in female suicides. They also found that media
portrayals broadcast on television were as likely to be
associated with an increase in male as in female suicides.
Explicit reference in the media item content to a particular
suicide method was not related to an increase in frequencies in
both female and male suicides.
Findings from Stack's (2005) review, based upon 55 studies,
suggest that females are more susceptible than males to the
Werther effect. Namely, research studies on the suicide rates of
women were almost 5 times more likely than others to report a
signicant increase in completed suicides after media de-
pictions of suicides. In the light of the current state of knowledge
about the effect of gender in the relationship between suicide
portrayals in the media and the gender differences in real
suicides, there is an evident need to further explore the role of
sex differences in media effect.
As demonstrated in many studies, media content has the
potential to elicit emotional reactions in audience members
(Bandura, 1977; Giles, 2003). Studies, examining the impact
of visual media content on emotional reactions, have shown
gender differences. In three studies where pictures or lms
were presented to subjects, female subjects tended to report
more intense emotions than their male counterparts, as
assessed using a self-administered questionnaire (e.g. Gross
and Levenson, 1995; Huston et al., 1995; Lang et al., 1993).
Research in other domains may help us better understand
the mechanisms underlying the effect of media presentations
of suicide on suicidal behaviours by individuals. One variable
that has been investigated is emotional reactivity. Mehrabian
(1995), considered emotional reactivity to be a psychological
vulnerability, which he dened as the propensity in some
individuals to be more emotional, to experience intense
emotions more easily, to emotionally react to events in a
more intense manner and to stay in a state of arousal for a
longer period. In particular, according to a series of studies
conducted by Mehrabian (1995), it appears that those who
have high emotional reactivity have better memories of
emotional events, and tend to be more impulsive, more
anxious and more suicidal. A more recent study conducted by
Nock et al. (2008) with 87 adolescents and young adults
under clinical treatment indicated that emotional reactivity
played a mediator role in the association between psychopa-
thology and suicidal ideations.
A third psychological vulnerability that may be associated
with cognitive reactions to exposure to media portrayals of
suicide is the tendency for dissociation. In the DSM-IV
490 L. Pouliot et al. / Journal of Affective Disorders 134 (2011) 488496
(American Psychiatric Association, 1994), dissociation is
dened as a disruption of the usually integrated functions
of consciousness, memory, identity or perception of the
environment. Research on dissociation tendency has found
that following exposure to traumatic events, those who have
such tendencies tend to be more vulnerable to intrusive
memories (Holmes et al., 2004; Murray et al., 2002; Punamki
et al., 2005). In a laboratory experiment, Murray (1997)
examined the potential association between dissociation, as
measured with a self-reported questionnaire, and reactions to a
lm depicting a series of car accidents. He found that those
participants having a high score on dissociation reported more
intrusive memories associated with the lm content.
Fourth, the personal tendency to suppress or divert from the
stream of consciousness aversive emotional thoughts (Wegner
and Zanakos, 1994) may also be an important psychological
vulnerability to consider. Multiple pieces of evidence converge
to show that this factor is involved in the development and
maintenance of many psychological disorders and is associated
with suicidal behaviours (Chawla and Ostan, 2007; Lynch et
al., 2004; Najmi et al., 2007; Pettit et al., 2009). A series of three
independent studies conducted by Pettit et al. (2009), with
clinical and non clinical samples of participants between age 15
and 22 years, indicated a positive and signicant association
between the tendency to suppress unwanted thoughts and the
intensity of suicidal ideation.
Only a small number of studies have examined in a direct
manner the cognitive and affective inuences of media
portrayals of suicide in lms. In fact, it is difcult to compare
results between these studies because of striking differences
in their methodology, participants' proles and the depen-
dent variables measured. Martin (1996) investigated the
predictive power of the frequency of exposures to suicide on
television for suicidal ideation, suicide attempts, and depres-
sion using a self-reported questionnaire with 357 high school
students. The frequency of exposures to suicide on television
did predict suicide attempts, but it was not a signicant
predictor of either depression or suicidal ideation.
Doron et al. (1998) presented to three groups of
participants under psychiatric care a 20 minute lm excerpt
showing the suicide by suffocation of two adolescents by in
reaction to their conicting relationships with their parents.
The rst group of participants was composed of suicide
attempters, the second group presented suicidal ideation and
the third group was not suicidal. Measures of physiological
reactions, anxiety and psychomotor agitation were taken
before and after the stimulus presentation. In general, the
suicide attempters had less intense reactions to the lms,
compared to the other groups. The group with suicidal
ideation had intermediate values on the same measures of
reactions, when compared to the other two groups.
Biblarz et al. (1991) studied the impacts of lms depicting
either a suicide scene, a violent scene or a neutral content
scene on attitudes toward suicide. Their study was conducted
with the participation of non suicidal university students who
were assigned to one of the three viewing conditions.
Participants' attitudes toward suicide and level of physiolog-
ical arousal were measured before and after the lm
presentation. Results failed to show a signicant difference
between groups in their attitudes toward suicide. However,
participants who viewed the suicide scene manifested greater
physiological arousal after viewing compared to the other
groups.
Using a qualitative methodology, Zahl and Hawton (2004)
interviewed 12 patients, aged between 17 and 25, who had
engaged recently in a self-harming behaviour. Participants
were asked to indicate whether they had been exposed to a
suicide story on television or at the cinema and if so to
describe the inuence that this story had had on them.
Among the 12 patients interviewed, ten reported that a
suicide story they saw had inuenced them. The inuence
reported by those patients varied greatly, with reactions
including: believing that suicide is an acceptable option,
rejecting the option of suicide, considering other methods of
suicide, conding to someone close to them about their
suicide attempts and prompting them to engage in the same
suicidal behaviour as the one depicted in the lm or television
suicide story they had seen.
More recently, Till et al. (2010), building upon Biblarz et al.
(1991) methodology, studied the inuence of ve lms, with
and without a suicide content, on non suicidal adult subjects'
mood, arousal, inner tension, self-esteem, life satisfaction,
depression, suicidality, and attitudes toward suicide. Their
results did not indicate any signicant differences on the
dependant variables between viewing conditions. However,
all lm stimuli led to adverse consequences on participants'
mood, inner tension and depression, whilst being associated
with positive effects on their self-esteem, life-satisfaction and
suicidal tendencies.
The above body of ndings produced by research on the
Werther Effect indicates that little is still known about the
mechanisms underlying the association between media and
suicide. With that in mind, this study was designed to explore
the emotional and cognitive reactions to suicide stories in
lms and in television programmes as self-reported by a
group of respondents and to examine the associations of
participants' reactions with several indicators of psycholog-
ical vulnerability.
Six research questions guided the development of our
study and the analyses: (i) How prevalent and long-lasting
are suicide story-induced emotional and cognitive reactions
in a sample of young adults? (ii) What is the nature and range
of negative affective and cognitive reactions reported? (iii)
How prevalent is the reporting by young adults of suicidal
thoughts induced by a suicide lm? (iv) Which of the
psychological vulnerabilities are related to suicide-story
induced affective and cognitive reactions? (v) Which of the
psychological vulnerabilities of suicidal ideation, emotional
reactivity, dissociation, and thought suppression tendencies
predict the intensity of negative affective and cognitive
reactions? and (vi) Which of the psychological vulnerabilities
factors predict suicidal thoughts induced by suicide lms?
2. Method
2.1. Participants
The study sample consisted of 101 undergraduate stu-
dents from the province of Quebec, Canada. All participants'
rst spoken language was French. Their average age was
23 years old (SD = 4.27). Twenty nine (29) were male and
seventy one (71) were female. The gender for one participant
 L. Pouliot et al. / Journal of Affective Disorders 134 (2011) 488496
was not available. At the recruitment phase of the study,
described in further details in the procedure section, 305
students consented to participate in the study. In a later
phone contact to arrange with them an appointment to either
one of three group sessions, 204 students desisted from
participating.
2.2. Instrument
A 127 item questionnaire, divided into three sections, was
used in the study. In the rst section, socio-demographic
information about participants' age, gender and level of
formal education were collected with 4 items. The second
section entailed 32 items asking respondents about their
exposure to suicide stories, seen on television, in music
videos, in lms broadcast on television or seen at the cinema.
Following an afrmative response about exposure, they were
invited to give the title of the lm they viewed and to
describe its content in a few lines. Then, on a scale of six
categorical choices, respondents were invited to indicate the
duration of their distress, if any, induced by the lm they
viewed (ex.: I have been distressed 1 For a few hours (less
than a day) to 6 For many months). Also, the negative
affective and cognitive reactions induced by a lm, as
remembered and reported by the respondents, were ex-
plored with a 22 item scale, inspired and adapted from Weiss
and Mamar's Impact of Event Scale-Revised (IES-R: Weiss
and Mamar, 1997) ( Cronbach = .89). For this study, the
version validated in French by Brunet et al. (2003) served as a
template in the construction of the items. The original
version of the instrument is designed to assess stress
responses to real life events. In the context of our study, it
was necessary to adjust items' wordings to the fact that
respondents would be asked to report on their cognitive and
emotional responses evoked by a suicide scene viewed in a
lm. Therefore, original items' scale where the word event
appeared was removed and replaced by the word scene. On
a further series of items, participants were solicited to
estimate retrospectively the intensity of their emotional
reaction to four types of emotions (fright, horror, anger, and
powerlessness) likely to have been evoked by the stimulus
viewed on a ve point scale (1 not at all to 5 extremely).
This section also included two items asking participants to
indicate to what extent the media presentation they viewed
made them fear they would engage in the same suicidal
gesture as the protagonist in the scene, and to what extent
they had the idea to do the same thing as the protagonist in
the scene. Answers to these items were collected on a ve-
point scale (1 not at all to 5 extremely).
The third section of the questionnaire consisted of scales on
the four psychological vulnerabilities under study (i.e. emo-
tional reactivity, tendency to commit suicide, dissociation
tendency, and thought suppression tendency). The eight item
subscale of the Suicide Probability Scale, measuring suicidal
ideation, was used as a proximal indicator of the propensity of
respondents to commit suicide (SPS: Cull and Gill, 1982/1988).
For this study, the version validated in French by Labelle et al.
(1998) of the SPS was used ( Cronbach = .71). The tendency
towards dissociation was assessed with the 38 item Trait
Dissociation Questionnaire (TDQ: Murray et al., 2002) (
Cronbach = .93). The Mehrabian Trait Arousability Scale (TAS,
491
34 items: Mehrabian, 1994) served as the measure of
participants' emotional reactivity ( Cronbach = .89). Finally,
Wegner and Zanakos' 15 item White Bear Thought Suppression
Inventory was used for the psychometric assessment of
participants' tendencies to thought suppression (WBSI:
Wegner and Zanakos, 1994) ( Cronbach = .91). Since no
version of the TAS, TDS, and WBSI validated and translated in
French existed at the time the study was conducted, these
scales were translated by our research team using the backward
translation technique (Vallerand, 1989). The reliability of the
adapted or French translated version of the scales used in the
study generally compared favourably to the English versions.
2.3. Procedure
Participants were recruited during class time in a college
and a university located in Montreal. They were informed that
a study on the perception of violence in the media would be
held at the university and that they were being approached to
take part in the study. Students who manifested their interest
to participate where invited to write down their full name
and phone number on a list in order to later be contacted by
the experimenter or her research assistant. A few days later,
they were contacted by telephone and invited to attend one
of three group sessions held at the university. Participants
gave their informed consent at the outset of the group
sessions. They then completed the questionnaire used in the
study. Approximately 45 min were necessary to complete the
questionnaire.
Ethical approval from the institutional research ethics
review committee of the university was received before the
study was conducted. Since suicidal tendencies in partici-
pants were measured in this study, participants' score on the
Suicidal Probability Scale was computed in the two days
following the completion of the questionnaire. In accordance
with the SPS administration manual, individuals who present
a score equal or higher to 53 are of clinical consideration.
Using this cut-off criterion, 44 participants in our sample
were contacted and scheduled for a 15 minute clinical
interview with a member of our research team with
experience in suicide risk assessment at the university within
seven (7) days after completion of the questionnaire. At the
end of the interview, these participants were given a list of
mental health resources. Following these interviews, none of
the participants were considered at a current risk to commit
suicide, whether low, moderate or high. This procedure was
undertaken to ensure the well-being and referrals of our
participants, if needed.
3. Results
3.1. Prevalence and duration of suicide story-induced emotional
and cognitive reactions
The rst research question asked how prevalent and long-
lasting suicide story-induced emotional and cognitive re-
actions would be in our sample. Consistent with past media
studies showing gender differences in emotional reactions,
gender specic results are presented. On the rst set of
questionnaire items, 70% of the participants indicated having
memories of a scene of suicide depicted on screen and having
492 L. Pouliot et al. / Journal of Affective Disorders 134 (2011) 488496

been distressed by the scene. When looking at gender
differences, 55% of the males endorsed having been distressed
compared to 76% of females. From this sub-group, 33% overall
mentioned that their distress lasted from a few days to less
than a month. However, males were proportionally less
numerous in claiming that their distress lasted over the rst
day of their exposure to the suicide scene (13%) compared to
females (39%). Each participant completed the questionnaire
using as referent a lm they viewed in the past outside the
study context. The lms listed by participants, following
verication by our research team, were classied, by the
Quebec Film Review Board, as being adequate for mature
audiences, available in most local video stores for rentals, and
all had been featured in movie theatres during the last
decade. Therefore, the media stimuli they reported on were in
the limits of what most adult spectators may view or be
exposed to on a day to day basis. On a subsequent item of the
questionnaire asking participants to indicate the emotions
provoked by the suicide stimulus they viewed, the majority
(60%) reported that the stimulus viewed had induced a
feeling of powerlessness. They were also numerous (51.4%) to
indicate having felt horried by what they had viewed. In
comparison, they were less likely to indicate having been
frightened (34.2%) and reporting that the scene induced a
feeling of anger (42%). As shown in Table 1, for a good
proportion of the participants, the intensity of these scene-
induced emotions was quite strong, ranging from moderate
to extreme. Table 1 shows that females across all items
covered in the questionnaire reported stronger emotional
reaction than males.
3.2. Range and types of stress reactions (affective and emotions
reactions) reported
The second research question concerns the range of stress
reactions that were reported in our sample following their
viewing of a media presentation of a suicide. On the total
mean scores derived from the scene-induced cognitive and
emotional stress scale non signicant gender differences
were found, (t (64) = .37, n.s.; males: M = 36.13, SD = 9.93
vs. females: M = 37.33, SD = 11.21). Over 71% of the sample
reported at least one stress reaction. Typically, stress re-
actions were of an intrusive nature (71%), such as being
mentally preoccupied by what they had seen in the scene. In
7% of the participants, this mental preoccupation was
extreme. Furthermore, nine participants (13%) said that
they experienced recurrent thoughts for many days to
many weeks following the viewing. Most participants also
reported having sustained waves of strong feelings about the
scene (68%), and that any reminder of the scene brought back
emotions related to it (61%).
3.3. Prevalence of suicide-related thoughts induced
by the suicide scene
The third research question asked how prevalent suicide-
related thoughts would be in our sample. We found that 13%
said that they feared that they would engage in the same
suicidal gesture as the protagonist in the scene. Furthermore,
13% reported that the idea crossed their mind to do the same
thing as the protagonist in the scene. No signicant gender
differences were found for any of the items pertaining to the
suicide ideas elicited by the suicide scene they viewed. In
other words, both females and males endorsed in the same
proportions both items.
3.4. Associations between psychological vulnerabilities and suicide
story-induced negative emotional and psychological reactions
The fourth research question was to determine which
psychological vulnerabilities considered in the study would
be associated with the suicide-scene induced stress reactions
and suicidal related thoughts. As shown in Table 2, correlation
analyses revealed that scores on the two indices of scene-
induced suicide related thoughts were both signicantly
associated with participant's history of past suicide ideation,
(r(70) = .44, p b .001, and r(69) = .38, p b .01). The gender
variable was signicantly related to two out of the seven (7)
variables under study. Males tended to show stronger
dissociation tendencies than females (r (69) = .27,

Table 1
Distribution of participants' answers on type and intensity of experienced emotions about the suicide scene they viewed (n = 70).

Item Not at all To some extent Moderately Quite a lot Extremely

I felt frightened
Total sample 37.1* 28.6 21.4 11.4 1.4
Male 43.8 25.0 25.0 0 6.3
Female 35.2 29.6 20.4 14.8 0
I felt horried
Total sample 21.4 27.1 27.1 11.4 12.9
Male 31.3 31.3 12.5 18.8 6.3
Female 18.5 25.9 31.5 9.3 14.8
I felt angry
Total sample 28.6 30.0 24.3 14.3 2.9
Male 25.0 50.0 18.8 6.3 0
Female 29.6 24.1 25.9 16.7 3.7
I felt powerless
Total sample 18.6 21.4 18.6 20.0 21.4
Male 43.8 12.5 12.5 31.3 0
Female 11.1 24.1 20.4 16.7 27.8

Note. *Percentage of participants endorsing the answer.

 L. Pouliot et al. / Journal of Affective Disorders 134 (2011) 488496
Table 2
Intercorrelations between measured variables (n a).
Gender Reactivity Dissociation
Gender .33 .27
Reactivity .30
Dissociation
Suppression
Ideation
Stress
Imitation 1
Imitation 2
Notes: Reactivity = emotional reactivity; dissociation = dissociation tendency; suppression = thought suppression tendency; ideation: suicide ideation;
imitation 1 = I have feared that I would engage in the same suicidal gesture as the protagonist in the scene; imitation 2 = The idea crossed my mind to do the
same thing as the protagonist in the scene.
a
n varies from 65 to 70 according to availability of data on the variable.
b.05.
b.01.
b.001.
p b .05). On the other hand, females tended to report higher
emotional reactivity than males (r (69) = .33, p b .01). A
strong and positive correlation was observed between
emotional reactivity and the overall score on the suicide
scene impact scale measuring scene induced stress reactions
(r (65) = .47, p b .001). There was a signicant positive
association between tendencies toward thought suppression
and stress reactions, (r (65) = .40, p = .001). The relationship
between the propensity towards dissociation and stress
reactions was also signicant, (r (65) = .40, p = . 001). The
association between the severity of stress reactions and
suicide ideation was not signicant. Finally, analyses indicat-
ed that the severity of scene-induced stress reactions, thought
suppression, emotional reactivity, and dissociation propensi-
ties were not signicantly correlated to stimulus-induced
suicidal ideation. Since gender appeared as a non signicant
correlate of the independent variables under study, this factor
was not included in further analysis.
3.5. Predictor analysis for stress reaction outcomes induced
by a suicide scene
The fth research question asked which psychological
vulnerability factors would predict the severity of stress
reactions. The independent contribution of the psychological
vulnerability factors in the prediction of the severity of stress
reactions induced by the suicide scene was determined by a
standard multiple regression analysis. The propensities
towards dissociation (partial = .22) and emotional reactivity
(partial = .35) were signicantly associated with the severity
of stress reactions, (F (4, 60) = 6.29, p b .001), and explained
54% of the variance in stress reactions. In contrast, the
thought suppression factor and suicidal ideation antecedent
did not contribute signicantly to the prediction of the stress
effect of the suicide scene (for further details see Table 3).
3.6. Predictor/Moderator analysis for scene-induced suicidal
ideations
The sixth and most important question concerns the
predictive power of the psychological vulnerability factors
used in this study on scene-induced suicidal ideation and was
addressed by conducting logistic regression analyses. Two
493
Suppression Ideation Stress Imitation 1 Imitation 2
.11 .20 .05 .04 .03
.60 .23 .47 .20 .16
.61 .37** .40 .15 .03
.24* .40 .12 .10
.14 .44 .38
.19 .07
.71
logistic regression analyses were performed using each of the
vulnerability factors, severity of stress reaction index and
interaction cross product terms as the independent factors to
predict scene-induced suicidal ideations.
The rst logistic regression showed that having antecedents
of suicidal ideation positively predicted fear of imitating the
suicidal gesture of the protagonist in the scene (M= 14.78 vs.
M = 10.08; OR= 1.62, 95% CI = 1.01 to 25.27; 2 = 17.51,
p b .05). However, none of the interaction terms was signicant.
The second regression analysis showed that participants who
reported that the idea crossed their mind to imitate the suicidal
protagonist in the scene were more suicidal (M= 15.56 vs.
M = 9.93; OR = 3.45, 95% CI = 1.18 to 836.97), presented
higher dissociation (M= 50.56 vs. M = 39.95; OR= 2.42,
95% CI = .01 to 1.13) and thought suppression tendencies
(M= 52.44 vs. 47.03; OR = 2.34, 95% CI = .76 to 140.15),
(2 = 23.64, p b .01). Again, none of the interaction terms
were signicant.
4. Discussion
This study is among the rst of its kind to explore
empirically the emotional and cognitive reactions induced by
ctional portrayals of suicide presented in movies and
television series with a non clinical sample of young adults.
It went beyond previous research in attempting to identify
psychological vulnerabilities likely to predict emotional and
cognitive reactions induced by suicide portrayals in mass
media.
In our sample, reactions of distress that are reported to
have been caused by exposure to a suicide scene on screen
appear to be quite prevalent, with 71% of our sample
reporting distress after media exposure to a suicide. In this
regard, it is worth mentioning that all lms or television
series mentioned by participants are within everyone's reach
from local video stores or the internet. For the majority, the
distress induced had a short duration (limited to a few hours).
However, more than one third of participants indicated that
their distress lasted for several weeks. This nding contrasts
with Bollen and Phillips' (1982) conclusion, derived from
their study on the effects of broadcast television news on
suicides. The authors then concluded that the effects on the
audience were limited within seven days of the broadcast.
494 L. Pouliot et al. / Journal of Affective Disorders 134 (2011) 488496
Table 3
Standard multiple regression of vulnerability factors on severity of suicide-
scene-induced stress reactions.
Predictor B SE (B) t p
Suicidal ideation antecedents .06 .13 .05 .47 .64
Emotional reactivity .39 .14 .40 2.90 .01
Dissociation tendency .33 .16 .29 2.04 .05
Thought suppression .01 .17 .00 .00 .99
Our ndings suggest that ctional portrayals of suicides, by
their graphic nature, may have a more signicant and longer
lasting impact than newscasts about suicides.
Our results also show that ctional portrayals of suicide in
electronic mass media seem to provoke stress reactions,
which in most instances take the form of a mental
preoccupation, recurrent and obsessive thoughts and waves
of strong feelings about the suicide portrayal they viewed. We
found that for 10% of participants, the stress reactions were
acutely experienced for a long time. These suicide scenes
were reported as having evoked feelings of powerlessness
and horror in most of our participants. For 13% of the
participants, their exposure to a suicide portrayal was
accompanied by a fear of engaging in the same suicidal
gesture that was depicted and thoughts about doing the same
as the suicidal protagonist in the scene. These observations
are consistent with previous research showing that viewers
are not all equally affected by media portrayals of suicide
(Biblarz et al., 1991; Till et al., 2010). However, as proposed
by some researchers, some viewers may be more at risk of
being negatively affected by media portrayals of suicide
because of psychological vulnerabilities (Chen et al., 2010;
Cheng et al., 2007a; Chiu et al., 2007; Doron et al., 1998;
Fekete and Schmidtke, 1996; Zahl and Hawton, 2004).
Despite the observation in our study that females reported
overall stronger emotional reactions, this does not seem to
translate to greater susceptibility to suicide-scene induced
cognitive and emotional stress reactions for them compared to
their male counterparts. Many clinical studies have demon-
strated that females report more symptoms of mood distur-
bance and depression than do males (for a review see Zender
and Olshansky, 2009). Notwithstanding this fact, the aetiology
of this difference and its' impact on stress responses are still
unknown (Young and Korszun, 2010). As it is the case in the
majority of clinical studies and studies in the eld of
psychology, females outnumbered males in our sample. It
follows that a replicated study with a balanced number of
females and males participants may offer a more conclusive
answer about gender issues in the Werther Effect. Our results
concur, to some extent with Pirkis et al.'s (2006) ndings that
for both, men and women, dissemination of suicide portrayals
in electronic media are as likely to be associated with suicidal
behaviour.
Congruent with prior research, our results show that
suicidal antecedents increase the odds that ones' thoughts
about suicide would be affected by a portrayal of suicide in
the media (e.g. Chen et al., 2010; Cheng et al., 2007a). People
with prior suicidal ideation and/or suicide attempts appear to
be particularly affected by suicide scenes. They are more
likely to fear that they will commit suicide like the victim in
the media presentation and contemplate that they might
replicate the protagonist's suicidal behaviour. In our sample,
nine participants reported such an inuence. This should be
of major concern since our convenience sample was drawn
from a non-clinical sample.
Results of this study also reveal that propensities towards
emotional reactivity and dissociation are signicant psycho-
logical vulnerabilities accounting for variations in the severity
of stress reactions induced by media depictions of suicide.
These ndings are in line with previous research on the
relationships between these psychological vulnerabilities and
cognitive and emotional reactions to aversive events in
general (Holmes et al., 2004; Mehrabian, 1995; Murray,
1997; Murray et al., 2002; Punamki et al., 2005).
Concerning the Werther Effect, it may be argued that
psychological vulnerabilities could partially account for the
suicidogenic effect of suicide portrayals in electronic media
on the population. Some media portrayals of suicide, as
demonstrated in this study, induce emotional and psycho-
logical stress reactions mostly modulated by propensities to
emotional reactivity. The risk that these stress reactions result
in suicidal ideation depends upon antecedents of suicidal
thoughts, thought suppression and dissociations tendencies.
Together these psychological vulnerability factors are signs of
maladaptive cognitive and emotional processing strategies to
cope with stressful events in general, and as might be also the
case with exposure to depiction of suicide in media.
Prospective and experimental studies are necessary to
determine whether or not emotional reactivity does indeed
lead to or predispose one to increased stress reactions to
media portrayals of suicide, and that suicidal, thought
suppression and dissociation tendencies act as mediator
variables.
Due to the relatively small sample size, the absence of, or
marginally signicant results for some psychological vulner-
ability variables should be viewed cautiously. Additional
studies, with a higher statistical power are needed in order to
verify if those psychological vulnerabilities are independent
predictors, moderators or uninvolved in the Werther effect.
5. Limitations
The ndings should be interpreted with caution since
they are based upon reports on exposure to variety of
different media portrayals of suicide which are probably not
comparable in terms of their content and cinematographic
characteristics. Furthermore, the results are based upon a
survey methodology. Survey methodologies cannot estab-
lish causal relationships nor determine the causal direction
between the variables under study. In light of such
limitations, future research might consider using an exper-
imental approach where participants are exposed to media
depictions of suicide where the attributes of the media
presentations are controlled, and their reactions are
assessed, thus allowing for a better test of causality. Also,
the emotional and cognitive inuences documented here
may not fully generalise to the young adult population since
our sample was composed of undergraduate students. We
suggest that future research use a more heterogeneous
sample of participants.
 L. Pouliot et al. / Journal of Affective Disorders 134 (2011) 488496
In spite of these limitations, this study suggests many new
paths of inquiry on the interplay between electronic media
portrayals of suicide, psychological vulnerabilities, and cogni-
tive and emotional reactions.
6. Conclusion
This study contests several previously held assumptions
about the Werther effect, many of which are included in media
guidelines for reporting on suicides. It shows that ctional
depictions of suicide are as potent, and possibly more
inuential, in provoking suicidal reactions as news stories
about real suicide cases. Furthermore, their negative impacts
seem to not be limited to people with a strong clinical prole.
Stress reactions to ctional suicides may extend well beyond
seven days after exposure. There is a need to replicate our
ndings with larger and more varied samples and to conduct
experimental investigations to explore more fully the factors
which inuence the impact of ctional depictions of suicides in
the media. However, given our ndings, it is reasonable to
recommend that broadcasting companies and lmmaking
agencies should be aware of the potential negative impact on
viewers in the diffusion of suicide portrayals on screen.
Role of the funding source
This research was not supported by any grant from funding sources.
Conict of interest
The authors declare that they have no competing interests.
Acknowledgements
At the stage of its design, this study has beneted from many useful
comments of afliated researchers and community partners of our research
center. We would like to thank Eveline Pilon, for her assistance in the literature
searches, and Prscilla Labelle for her collaboration in data collection and entry.
References
American Psychiatric Association, 1994. Diagnostic and statistical manual of
mental disorders, 4th ed. Author, Washington, DC. (DSM-IV).
Bandura, A., 1977. Social learning theory. General Learning Press, New York,
NY.
Bandura, A., 1986. Social Foundations of Thought and Action: A social
Cognitive Theory. Prentice Hall, Englewood Cliffs, NJ.
Berman, A.L., 1988. Fictional depiction of suicide in television lms and
imitation effects. Am. J. Psychiatry 145, 982986.
Biblarz, A., Brown, R., Biblarz, D., Pilgrim, M., Baldree, B., 1991. Media
inuences on attitudes toward suicide. Suicide Life Threat. Behav. 21,
374384.
Bollen, K.A., Phillips, D.P., 1982. Imitative suicides: a national study of the
effects of television news stories. Am. Sociol. Rev. 47, 802809.
Brunet, A., St-Hilaire, A., Jehel, L., King, S., 2003. Validation of a French version
of the Impact of Event Scale-Revised. Can. J. Psychiatry 48, 5661.
Canetto, S.S., 2008. Women and suicidal behaviour: a cultural analysis. Am. J.
Orthopsychiatry 78, 259266.
Chawla, N., Ostan, B., 2007. Experiential avoidance as a functional
dimensional approach to psychopathology: an empirical review. J. Clin.
Psychol. 63, 871890.
Chen, Y.-Y., Tsai, P.-C., Chen, P.-H., Fan, C.-C., Hung, G.C.-L., Cheng, A.T.A.,
2010. Effect of media reporting of the suicide of a singer in Taiwan: the
case of Ivy Li. Soc. Psychiatry Psychiatr. Epidemiol. 45, 363369.
Cheng, T.A., Hawton, K., Chen, T.H.H., Yen, A.M.F., Chen, C.-Y., Chen, L.-C., et
al., 2007a. The inuence of media coverage of a celebrity suicide on
subsequent suicide attempts. J. Clin. Psychiatry 68, 862866.
Cheng, T.A., Hawton, K., Lee, C., Chen, T.H.H., 2007b. The inuence of media
reporting of the suicide of a celebrity on suicide rates: a population-
based study. Int. J. Epidemiol. 36, 12291234.
Chiu, S.-H., Ko, H.C., Wu, J.Y.W., 2007. Depression moderated the effect of
exposure to suicide news on suicidality among college students in
Taiwan. Suicide Life Threat. Behav. 37, 585592.
495
Cull, J.G., Gill, W.S., 1982/1988. Suicide Probability Scale (SPS): Manual.
Western Psychological Services, Los Angeles.
Doron, A., Stein, D., Levine, Y., Abramovitch, Y., Eilat, E., Neuman, M., 1998.
Physiological reactions to a suicide lm: suicide attempters, suicide
ideators, and non suicidal patients. Suicide Life Threat. Behav. 28,
309314.
Etzersdorfer, E., Voracek, M., Sonneck, G., 2004. A dose-response relationship
between imitational suicides and newspaper distribution. Arch. Suicide
Res. 8, 137145.
Fekete, S., Macsai, E., 1990. Hungarian suicidal models: past and present. In:
Ferrari, G., Bellini, M., Crepet, P. (Eds.), Suicidal behavior and risk factors.
Monduzzi Editore, Bologna, Italy, pp. 149156.
Fekete, S., Schmidtke, A., 1996. Suicidal models their frequency and role in
suicide attempters, non suicidal psychiatric patients and normal
control case: a comparative GermanHungarian study. Omega 33,
233241.
Frei, A., Schenker, T., Finzen, A., Dittman, V., Kraeuch, K., Hoffmann-Richter,
U., 2003. The Werther effect and assisted suicide. Suicide Life Threat.
Behav. 33, 192200.
Giles, D., 2003. Media psychology. Erlbaum, Mahwa, NJ.
Goethe, J.W., 17741968. Les souffrances du jeune Werther. Prsentation et
traduction par Joseph-Franois Angelloz. Flammarion, Paris.
Gould, M.S., 2001. Suicide and the media. Ann. N. Y. Acad. Sci. 932, 200224.
Gould, M.S., Shaffer, D., 1986. The impact of suicide in televisionmovies. New
Engl. J. Med. 315, 690694.
Gould, M.S., Shaffer, D., Kleinman, M., 1988. The impact of suicide in
television movies: replication and commentary. Suicide Life Threat.
Behav. 18, 9099.
Gross, J.J., Levenson, R.W., 1995. Emotion elicitation using lms. Cogn.
Emotion 9, 87108.
Hassan, R., 1995. Effects of newspaper stories in the incidence of suicide in
Australia: a research note. Aust. NZJ Psychiatry 29, 480483.
Hawton, K., Williams, K., 2001. The connection between media and suicidal
behaviour warrants serious attention. Crisis 22, 137140.
Holmes, E.A., Brewin, C.R., Hennessy, R.G., 2004. Trauma lms, information
processing, and intrusive memory development. J. Exp. Psychol. Gen.
133, 322.
Huston, A.C., Wright, J.C., Alvarez, M., Truglio, R., Fitch, M., Piemyat, S., 1995.
Perceived television reality and children's emotional and cognitive
responses to its social content. J. Appl. Dev. Psychol. 16, 231251.
IMDb, 2010. Available http://www.imdb.com/2010 [On-line].
Jonas, K., 1992. Modeling and suicide: a test of the Werther effect. Br. J. Soc.
Psychol. 31, 295306.
Labelle, R., Daigle, M.S., Pronovost, J., Marcotte, D., 1998. tude psychom-
trique d'une version franaise du Suicide Probability Scale auprs de
trois populations distinctes. Psychol. Psychomtrie 19, 427.
Lang, P.J., Greenwald, M., Bradley, M.M., Hamm, A.O., 1993. Looking at
pictures: evaluative, facial, visceral, and behavioural responses. Psycho-
physiology 30, 261273.
Lynch, T.T., Cheavens, J.S., Morse, J.Q., Rosenthal, M.Z., 2004. A model
predicting suicidal ideation and hopelessness in depressed older adults:
the impact of emotion inhibition and affect intensity. Aging Ment. Health
8, 486497.
Martin, G., 1996. The inuence of television suicide in a normal adolescent
population. Arch. Suicide Res. 2, 103117.
Mehrabian, A., 1994. Manual for the revised Trait Arousability (converse of
the Stimulus Screening) Scale. (Available from Albert Mehrabian, 1130
Alta Mesa Road, Monterey, CA, USA 93940).
Mehrabian, A., 1995. Theory and evidence bearing on a scale of Trait
Arousability. Curr. Psychol. 14, 328.
Mishara, B.L., Tousignant, M., 2004. Comprendre le suicide. Presse de
l'Universit de Montral, Montral.
Murray, J., 1997. The role of dissociation in the development and
maintenance of post-traumatic stress disorder. Unpublished doctoral
dissertation, Oxford University, Oxford, England.
Murray, J., Ehlers, A., Mayou, R.A., 2002. Dissociation and post-traumatic
stress disorder: two prospective studies of road trafc accident
survivors. Br. J. Psychiatry 180, 363368.
Najmi, S., Wegner, D.M., Nock, M.K., 2007. Thought suppression and self-
injurious thoughts and behaviours. Behav. Res. Ther. 45, 19571965.
Niederkrotenthaler, T., Voracek, M., Herberth, A., Till, B., Strauss, M.,
Etzersdorfer, E., Eisenwort, B., Sonneck, G., 2010. Role of media reports
in completed and prevented suicide:Werther v. Papageno effects. Br. J.
Psychiatry 197, 234243.
Niederkrotenthaler, T., Till, B., Kapusta, N., Voracek, M., Dervic, K., Sonneck,
G., 2009. Copycat effects after media reports on suicide: a population-
based ecologic study. Soc. Sci. Med. 69, 10851090.
Nock, M.K., Wedig, M.M., Holmberg, E.B., Hooley, J.M., 2008. The Emotion
Reactivity Scale: development, evaluation, and relation to self-injurious
thoughts and behaviours. Behav. Ther. 39, 107116.
496 L. Pouliot et al. / Journal of Affective Disorders 134 (2011) 488496
Pettit, J.W., Temple, S.R., Norton, P.J., Yaroslavsky, I., Grover, K.E., Morgan, S.T.,
Schatte, D.J., 2009. Thought suppression and suicidal ideation: prelim-
inary evidence in support of a robust association. Depress. Anxiety 26,
758763.
Phillips, D.P., 1974. The inuence of suggestion on suicide: substantive and
theoretical implications of the Werther effect. Am. Sociol. Rev. 39,
340354.
Phillips, D.P., Carstensen, M.S., 1986. Clustering of teenage suicides after
television news stories about suicide. N. Engl. J. Med. 315, 685689.
Pirkis, J., Blood, R.W., 2001a. Suicide and the media. Part I: reportage in
nonctional media. Crisis 22, 146154.
Pirkis, J., Blood, R.W., 2001b. Suicide and the media. Part II: portrayal in
ctional media. Crisis 22, 155162.
Pirkis, J., Burgess, P.M., Francis, C., Blood, W., Jolley, D.J., 2006. The
relationship between media reporting of suicide and actual suicide in
Australia. Soc. Sci. Med. 62, 28742886.
Platt, S., 1987. The aftermath of Angie's overdose: is soap (opera) damaging
to your health? BMJ 294, 954957.
Pouliot, L., Tousignant, M., 2010. Les grappes de suicide: un phenomene
psychosocial reel ou une aberration statistique? Can. Psychol. 51,
120132.
Punamki, R.-L., Komproe, I.H., Qouta, S., Elmasri, M., de Jong, J.T.V.M., 2005.
The role of peritraumatic dissociation and gender in the association
between trauma and mental health on a Palestinian community sample.
Am. J. Psychiatry 162, 545551.
Romer, D., Jamieson, P.E., Jamieson, K.H., 2006. Are news reports of suicide
contagious? A stringent test in six U.S. cities. JoC 56, 253270.
Schmidtke, A., Hfner, H., 1988. The Werther effect after television lms: new
evidence for an old hypothesis. Psychol. Med. 18, 665676.
Schmidtke, A., Schaller, S., 1998. What do we know about media effects on
imitation of suicidal behaviour: state of the art. In: De Leo, D., Diekstra,
R.F.W. (Eds.), Suicide prevention: a holistic approach. Kluwer Academic
Publishers, Netherlands, pp. 121137.
Stack, S., 1996. The effect of the media on suicide: evidence from Japan,
19551985. Suicide Life Threat. Behav. 26, 132142.
Stack, S., 2003. Media coverage as a risk factor in suicide. J. Epidemiol.
Community Health 57, 238240.
Stack, S., 2005. Suicide in the media: a quantitative review of studies based
on nonctional stories. Suicide Life Threat. Behav. 35, 121133.
Thorson, J., berg, P.-A., 2003. Was there a suicide epidemic after Goethe's
Werther? Arch. Suicide Res. 7, 6972.
Till, B., Niederkrotenthaler, T., Herberth, A., Vitouch, P., Sonneck, G., 2010.
Suicide in lms: the impact of suicide portrayals on nonsuicidal viewers'
well-being and the effectiveness of censorship. Suicide Life Threat.
Behav. 40, 319327.
Tousignant, M., Mishara, B.L., Caillaud, A., Fortin, V., St-Laurent, D., 2005. The
impact of media coverage of the suicide of a well-known Quebec
reporter: the case of Gatan Girouard. Soc. Sci. Med. 60, 19191926.
Vallerand, R.J., 1989. Toward a methodology for the trans-cultural validation
of psychological questionnaires: implications for research in the French
language. Can. Psychol. 30, 662680.
Wegner, D.M., Zanakos, S., 1994. Chronic thought suppression. J. Pers. 62, 615641.
Weiss, D.S., Mamar, C.R., 1997. The impact of Event Scale-Revised. In: Wilson,
J.P., Keane, T.M. (Eds.), Assessing Psychological Trauma and Ptsd: A
Practitioner's Handbook. Guilford, New York, pp. 399411.
Westerlund, M., Schaller, S., Schmidtke, A., 2009. The role of mass-media in
suicide prevention. In: Wasserman, D., Wasserman, C. (Eds.), Oxford
textbook of suicidology and suicide prevention: a global perspective.
Oxford University Press, NY, pp. 515523.
Yip, P.S., Fu, K.W., Yang, K.C., Ip, B.Y.T., Chan, C.L.W., Chen, E.Y.H., et al., 2006.
The effects of a celebrity suicide on suicide rates in Hong Kong. J. Affect.
Disord. 93, 245252.
Young, E., Korszun, A., 2010. Sex, trauma, stress hormones and depression.
Mol. Psychiatry 15, 2328.
Zahl, D.L., Hawton, K., 2004. Media inuences on suicidal behaviour: an
interview study of young people. Behav. Cogn. Psychother. 32, 189198.
Zender, R., Olshansky, E., 2009. Women's mental health: depression and
anxiety. Nurs. Clin. North Am. 44, 355364.