Contrary to the findings of a 2019 study that associated the release of the Netflix series “13 Reasons Why” with an increase in monthly suicide rates among adolescent boys, a reanalysis of the data by the Annenberg Public Policy Center finds no evidence of contagion.
The reanalysis, published today in PLOS ONE, found that after controlling for the dramatic increase in adolescent suicide in recent years, the show’s release had no clear effect.
The reanalysis found that the increase in the suicide rate for boys observed in the original study by Bridge and colleagues for three months after the series’ release was no longer evident after controlling for the ongoing trend in adolescent suicide.
In addition, the increase seen during the first month of the release began the month prior to the release, making it difficult to attribute the rise to the show. The reanalysis found no effects in the subsequent months of that year.
“Our reanalysis casts considerable doubt on the show having an effect on boys,” said author Dan Romer, Ph.D., research director of the Annenberg Public Policy Center (APPC) of the University of Pennsylvania.
The paper, “Reanalysis of the Bridge et al. study of suicide following release of 13 Reasons Why,” was published January 15, 2020, in PLOS ONE.
The original time-series analysis of suicide rates by Bridge et al. found an additional 195 suicide deaths among boys ages 10 to 17 during the nine months following the series’ release, on March 31, 2017, but did not find a similar result among adolescent girls. Romer said he questioned that finding for two reasons:
“First, contagion would be expected to be stronger for girls than boys because this show focuses on the suicide of a high school girl. Second, the prior analysis didn’t take into account strong secular trends in suicide, especially in boys from 2016-2017.”
APPC’s reanalysis did find a modest rise in the suicide rate among adolescent girls during the first month following the release of “13 Reasons Why,” but it was not statistically reliable.
Rising suicide rates for adolescents
Suicide rates for both adolescent males and females have been rising since 2008 and the rise in 2017 was particularly strong. The rate for young males ages 15 to 19 increased by 21 percent between 2016 and 2017. Suicide among female adolescents ages 15 to 19 increased by 7 percent.
There has been disagreement about the causes of the recent rise in adolescent and young adult suicide. Some researchers have suggested that the rise is due to increased use of social media. But a recently published analysis by Romer suggested that the trend may be attributable in part to lingering economic stress from the 2008 financial crisis in combination with increased pressure on teens from their parents to succeed academically.
’13 Reasons Why’ and suicide contagion
When the first season of the Netflix show appeared in 2017, it created widespread concern that its graphic portrayal of a teenage girl’s suicide would lead to imitation among vulnerable young people, especially adolescent girls. This concern prompted researchers to see whether there was evidence of contagion in U.S. suicide rates following the show’s release.
Suicide rates per 100,000 in youth ages 15-19 in the United States, 1981-2017. Source: Centers for Disease Control and Prevention. Image is credited to Daniel Romera.
The first such study, by Bridge et al., found evidence of a jump in suicides among boys ages 10 to 17 in the three months following the show’s release, but no effect for girls. Romer said that analysis relied on a forecasting method to project the likely trend in suicide for 2017. But that forecast failed to anticipate the actual trend.
A second study by different researchers (Niederkrotenthaler et al.) in JAMA Psychiatry had a similar problem, Romer said. That study found an effect in boys and girls 10 to 19 years old in the three months following the series’ release. But it similarly failed to control for the secular trend in suicide, again making it difficult to separate the effects of the show from the trend.
Potential for harm
APPC’s analysis does not rule out the possibility that the show had an adverse effect, especially on girls. A previous study by Romer and colleagues that examined the effects of the second season of “13 Reasons Why” found that viewing the show had both beneficial and detrimental effects on young adult viewers.
The small increase observed in girls in this current reanalysis of the first season data may have been the net result of these opposing effects, Romer said.
In mid-2019, over two years after the release, Netflix edited a graphic suicide scene from the Season 1 finale and asserted that the show encouraged young people “to start conversations about difficult issues such as depression and suicide and get help – often for the first time.”
“Despite that, it does not appear that the show reversed or slowed the ongoing increase in adolescent suicide,” Romer noted. “Even if the series also had a positive effect for some viewers, the producers should recognize the potential for harm to vulnerable audience members.
It should be possible to produce a show that highlights the challenges that young people face without also producing suicide contagion.”
Romer thanked Jeffrey Bridge for providing the data that were not available in his paper, which is in press in the Journal of the American Academy of Child & Adolescent Psychiatry.
The interference of video productions in public health issues occasionally arises in public debate (for example, as reported in press articles, such as in the New Yorker in May 2019: “Netflix and Suicide: The Disturbing Example of 13 Reasons Why”) , prompting scholars to seize the matter in terms of research and prevention. Along with violence , addiction , and public health in general , one of the most paradigmatic, tangible, and perhaps disturbing impacts of cinema and TV shows on people concerns suicidal behaviors [5–7].
Although initially described  and repeatedly demonstrated in the press media [9–11], the Werther Effect (WE) (i.e., the systematic increase of suicide rates following incautious media coverage of suicide news stories ) has been extended to television [13–15].
Based on a systematic review of 11 ecological studies, Pirkis and Blood argued that the association between television broadcasts of suicide and rates of suicidal behaviors is consistent and strong .
Conversely, evidence has been emerging for almost a decade that some reports of suicidal events could be associated with a reduction of suicide rates via the so-called Papageno effect (PE) [17–19].
Although less unequivocal [15,16], evidence confirms that broadcasts of fictional suicides tangibly impact viewers. For instance, audience reception studies based on pseudo-experimental designs suggest that exposure to movies featuring suicide scenes can increase suicidality in already vulnerable individuals [20,21].
Recently, the release of the Netflix show 13 Reasons Why, which depicts the suicidal process of an American adolescent girl, provided an ecological example of the effect of a video suicide narrative at the population level. Indeed, several authors cautioned about the serious—albeit potentially mixed [22–24]—impact of the series on youths’ mental health .
In the months following the broadcast, Niederkrothentaler et al. found that suicide rates among those aged 10–19 years were 13% higher than expected from secular trends . Similarly, Bridge et al.  evidenced that the release of 13 Reasons Why was associated with a significant increase in monthly suicide rates among youth in the US aged 10 to 17 years (incidence rate ratio (IRR), 1.29; 95% CI, 1.09–1.53).
The authors warned about the exposure of children and adolescents to the series. Among the hypotheses proposed to account for suicidal contagion at the individual level, identification may be one integrative and heuristically effective construct [27,28].
Identification relates to a psychosocial process in the interplay between neuroscience, cognitive, and behavioral theories, and developmental conceptualizations. From a phenomenological viewpoint, identification qualifies the experience of temporarily borrowing the perspective of another person (whether real or fictive), while weakening (although not totally “dissolving”) self-awareness .
From this perspective, identification finds a relevant translation in the neuroscience realm under the notion of empathy . According to recent understandings, empathy results from the dynamic integration of two opposite processes:
(1) Embodied simulation of others’ emotions and representations leading to first-/third-person confusion; and (2) mentalization, which enables discriminating one’s own state of mind from that inferred for others [31,32].
The simulation component is heavily involved in the affective arousal triggered by emotionally charged video scenes, while mentalization contributes to tempering emotional contagion .
Interestingly, identification may translate into the behavioral repertoire through social learning. According to Bandura’s theory , the probability of one individual adopting a behavior observed in a character is correlated with the strength of identification with this character.
In suicidal contagion, embodied simulation may be insufficiently balanced by mentalization, leading to such confusion with the character that it leads to an excessive emotional reaction and maladaptive imitation. Conversely, the observation of coping, help-seeking, or help-offering behaviors may lead to protective effects through a similar process.
The strong debates emerging in the scientific community after the release of 13 Reasons Why highlight the need for better insights into the impact of video suicide in terms of contagion. Although recent studies from Niederkronthentaler et al.  and Bridge et al.  provide new epidemiological evidence for the causal influence of fictional suicide on suicide rates, the mechanisms that underlie this macro-individual observation remain poorly known.
Contrary to the press media for which screening tools exist [35,36], both ecological and audience reception studies on video fictional suicide are hindered by the absence of valid instruments to analyze productions.
In this paper, we describe the development and validation process of the Movies and Video: Identification and Emotions in Reaction to Suicide (MoVIES). The MoVIES is a theory-oriented analytic tool designed to quantitatively characterize video scenes depicting a suicidal gesture regarding mechanisms involved in the suicidal contagion process, with a special focus on identification.
The MoVIES opens tracks for future research on contagion related to videos depicting fictional suicide. By informing epidemiological and audience reception studies, it could help in disentangling the relative contributions of the inherent properties of a video stimulus on the one hand, and the vulnerability of the viewer on the other in the causal processes leading to contagion. Importantly, future research using the MoVIES should be conducted keeping in mind important ethical issues.
Unlike journalists for whom scholars have proposed guidelines when reporting suicide, the relevance of editing recommendations for producers, directors, and actors is questionable. The creative process of films or a series is not similar to the objectivity that guides proper information. On the audience side, it should also be clear that instruments such as the MoVIES are not intended to supplant, but rather to promote consultation with suicide prevention experts.
Author Contributions: C.G. acquired, analyzed, and interpreted the data, and wrote the manuscript. C.-E.N. revised the manuscript, supervised the investigation, and provided advice and expertise. M.W. provided advice and performed the statistical analysis. T.B. and F.M. agreed and gave their opinion on the data and text. N.P. contributed to the coordination and formatting of the article. Each author has approved the present version of this text, and agrees to be personally accountable for the author’s own contributions and integrity of any part of the work. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflict of interest.
University of Pennsylvania