by Michael D. Anestis, M.S.
Suicide is a remarkably distressing phenomenon. Individuals are often left with a sense of bewilderment and intense sadness at the unexpected death by suicide of a loved one, a friend, or even a cherished celebrity. Needless to say, there is a strong drive on the part of both professionals and the general public to better understand suicidal behavior. The drive for knowledge is an inherently positive thing; however, one significant problem resulting from the need for answers has been the perpetuation of a number of myths regarding suicidal behavior - ideas that appear to offer an explanation but which do not truly reflect the nature of the phenomenon. Amongst these myths are the idea that suicide is an impulsive act engaged in without significant forethought and the idea that suicide reflects cowardice (see our article on Joiner's (2005) interpersonal-psychological theory of suicidal behavior for a description of the flaws in such ideas). In today's article, I would like to address another common misconception of suicidal behavior, the idea that clusters of suicide reflect a contagion.
The 11th leading cause of death in the United States (American Association of Suicidology, 2005), suicide is a significant, albeit statistically rare phenomenon. Although it occurs at a low rate in the general population, its impact is enormous, necessitating a systematic understanding of the risk factors underlying the phenomenon and effective means for intervening in moments of crisis. Despite the rarity of the behavior itself, suicides do, at times, cluster together in time and space. The precise definition of a suicide cluster is the occurrence of two or more completed suicides or suicide attempts bunched non-randomly in time or space (Joiner, 1999). Examples would include multiple suicides in a single high school during a school year or a series of suicides within an inpatient psychiatric facility. Researchers who have examined suicide clusters have specified two particular types: mass clusters and point clusters. Mass clusters do not occur in the same space, but do cluster in time in response to media related phenomenon. An example here would be several individuals nationwide attempting suicide after hearing a newscast detailing the suicide of a celebrity. Point clusters, on the other hand, involve a series of suicides within a small geographic area.
The evidence for mass suicide clusters is inconsistent. For instance, Kessler, Downey, Milavsky, and Stipp (1988) examined adolescent deaths by suicide between 1973 and 1984 and found no relationship between suicide related newscasts and subsequent adolescent suicide rates. To ensure that this did not simply reflect a lack of viewership on the part of adolescents, the authors then examined whether this relationship changed depending upon how many adolescents viewed the broadcasts. This analysis was consistent with the first, demonstrating no relationship between media stories on suicide and actual suicide rates in adolescents. This by no means conclusively proves that mass suicide clusters are nonexistent; however, it does raise questions as to whether the shocking nature of the idea of mass suicide clusters has resulted in a misunderstanding regarding the actual prevalence and nature of such phenomena.
Point clusters, the existence of which is supported by undeniable empirical evidence, typically occur within an institutional setting and inspire an understandably significant degree of concern amongst individuals at or near that institution. One common response to this phenomenon is to assume that the behavior itself is contagious. In the case of physiological illnesses, contagions are readily identifiable biological pathogens that infiltrate the body of one individual, potentially due to proximity to an already infected individual. In the case of behaviors, the precise nature of what would constitute a "contagion" is less clear, although the most readily agreed upon definition is imitation. As such, suicide contagion would be defined by individuals attempting and/or completing suicide in an effort to imitate the behavior of another individual.
The research on this phenomenon is fairly limited, but what has been reported does not support the idea of contagion in point clusters. Joiner (1999) explained that a variety of other factors better explain the occurrence of point clusters. First, the occurrence of negative life events is a significant predictor of suicidal behavior and the death by suicide of a peer would most definitely constitute a negative life event. Secondly, strong support is a buffer against suicidal behavior and, if a peer dies by suicide, social support could quite obviously be diminished. Third, there are a number of individually-based risk factors for suicide, including diagnoses of mental illnesses, that predispose certain individuals to suicidal behavior, particularly when exposed to chronic stress. Fourth, individuals tend to develop relationships assortatively (Grant et al., 2007; Segrin, 2004) - that is, we tend to find friends who are similar to us, both in personality and behavior.
So how do these four points argue against contagion? In the point clusters that have been reported in prior studies, the individuals who completed or attempted suicide have often been friends with one another (Haw, 1995; Robbins & Conroy, 1983). This is not necessary for the point to hold true, but makes the logic potentially easier to follow. In a study on a group of adolescents by Robbins and Conroy (1983), two suicides were followed by five attempts and one hospitalization due to severe suicidal ideation. All of these individuals had been previously admitted to inpatient psychiatric units and had socialized during their hospitalization. As such, a group of similar individuals who were close to one another were simultaneously exposed to severe stress and a reduction in social support. Their psychiatric diagnoses already predisposed them to suicidal behavior in the absence of these other stressors, but the confluence of all four significantly increased the likelihood of this rare and distressing phenomenon, a point cluster of suicides. This is not to say that it was highly predictable that all of these attempts would occur in shared space and time, but rather, it indicates that mechanisms other than contagious imitation were at work.
Stepping aside from anecdotal evidence for a moment - as PBB maintains the belief that empirical data is always superior to anecdotal accounts - Joiner (2003) provided further evidence against suicide contagions in a study he conducted on college roommates. 138 pairs of college roommates were administered a series of questionnaires. As it turns out, roommates who chose to room together were more similar on a measure of suicidality than were roommates who were grouped together randomly. What this finding indicates is that shared vulnerability for suicidal behavior is greater amongst individuals who form assortative relationships. Shared stress amongst vulnerable individuals will result in increased point clusters, not because individuals are simply imitating one another, but rather because individuals in point clusters are not random. In other words, it is not that a random collection of individuals simultaneously become vulnerable to suicide, imitating one another's behavior in response to stress. Instead, individuals already vulnerable to such behavior assortatively relate and respond to shared stress in a similar manner due to their shared vulnerabilities.
So what does the evidence tell us about suicide clusters and contagion? While evidence for mass clusters is unclear, point clusters are a very real and highly troubling phenomenon. Despite concerns regarding contagions, however, the evidence is not consistent with the idea that suicidal behavior is contagious and that imitating the behaviors of others is a primary risk factor. Instead, it appears that clusters occur when vulnerable individuals form friendships with one another and are exposed to shared stress. This distinction is important to make for a variety of reasons. First, it can help to reduce some anxiety on the part of parents and clinicians regarding the fear that suicide epidemics are likely in the face of a single tragedy. Second, it can help to better understand the actual vulnerabilities that drive suicidal behavior and thus increase the effectiveness of interventions aimed at preventing suicidal behavior.
If you or someone you know are experiencing thoughts of suicide, there are a number of resources available. 1-800-273-TALK is a free, 24/7, anonymous hotline staffed by highly trained professionals who want to help you in times of trouble. Additionally, you can call 9-1-1 and inform the respondent of your situation. The authorities will then respond by getting you to a safe place. Finally, please refer to our EST clinics page for a list of clinics that offer empirically supported treatments. This list is not comprehensive, so if you do not see a clinic in your area, this does not mean that no such clinic exists locally.
If you would like to learn more about suicidal behavior, we recommend the following products, all of which are available through our online store:
Books:
- Why People Die by Suicide
- The Interpersonal Theory of Suicide: Guidance for Working With Suicidal Clients
- The Interpersonal Solution to Depression: A Workbook for Changing How You Feel by Changing How You Relate (New Harbinger Self-Help Workbook)
- Treating Suicidal Behavior: An Effective, Time-Limited Approach (Treatment Manuals For Practitioners)
Kindle:
- Suicide Science: Expanding the Boundaries
Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University
Subscribe in a reader
