Copycat deaths are likely in the wake of the suicide of Robin Williams. More restrained reporting could help
When US actor Robin Williams killed himself, a flood of media coverage followed. Some of it has been criticised for glamorising his death in a way that might prompt vulnerable readers to follow suit.
Are the critics right? Research stretching back decades suggests that such fears are well founded. Empirical work on the possible role of the media on copycat or imitative suicide began with a seminal article by University of California professor David Phillips in 1974. He checked the impact of 34 widely publicised suicide stories that appeared on the front page of The New York Times.
The coverage spanned the years 1948 to 1967. On average, there were 22.03 additional suicides during the month following a story. Of the 34 cases, celebrities tended to trigger a larger copycat effect. For example, the suicide of former US secretary of defence James Forrestal on 10 March 1949, was associated with 55 additional suicides. That of US senator Lester Hunt on 20 June 1954, saw an increase of 89 suicides. The largest rise by far, however, involved the suicide of Marilyn Monroe on 5 August 1962. In the month after reports of her death there was an increase of 197 suicides.
Since this pioneering work there have been more than 100 empirical investigations of copycat suicide. A review of 419 findings from the first 55 investigations showed that only 35.8 per cent documented an increase in suicide after media coverage. Given that most evidence is not consistent with a copycat effect, a search for the conditions under which a story may elicit imitative suicides has been a key theme in this work.
The most important factor distinguishing studies that report a copycat effect from the ones that do not is whether or not a celebrity is involved. In particular, copycat effects are most likely to be reported in work focused on two distinct types of celebrities: those in politics and entertainment. The analysis of those 419 findings found that studies based on either or both of these subtypes were 5.27 times more likely to report an increase in suicides following coverage.
The latest evidence comes from South Korea in a journal paper out this month (Suicide & Life Threatening Behavior, p 457). An apparent record increase in suicide followed the widely reported suicide of a popular actress. Jin-shil Choi's death on 2 October 2008 was followed by 429 additional suicides. Choi was a "national actress" analogous to Julia Roberts in the US.
Given her fame – she starred in 18 films – there was much public interest and massive amounts of news coverage. In a typical week in South Korea there are 50-100 news items about suicide. During the month after Choi's death this rose to more than 1,600. The reporting tended to be inconsistent with the recommendations of the World Health Organization for media coverage of suicide.
Those guidelines urge particular caution when reporting celebrity suicides and other key points such as avoiding detailed description of the method used and sensationalising or normalising suicide. Only 10.9 per cent of news reports mentioned a suicide prevention hotline number or other source of help for distressed readers. About 40 per cent described the method of suicide. The researchers concluded "that massive and noncompliant media coverage of a celebrity suicide can cause a large-scale copycat effect".
Choi's death was associated with numerous strains which may have echoed with some already suicidal persons.
These problems included her victimisation in domestic violence, her divorce, and vicious rumours on the internet attributing the suicide of a fellow actor to Choi.
As in the case of many previous investigations, the increase in suicide after Choi's death followed age and gender patterns. For females the increase was 116 per cent. For the young the increase was 110 per cent.
However, for males under 29 there was no significant increase. The largest increase was for young females.
Females under 29 had an increase in suicide of 200 per cent.
The impact of Choi's death was strongest in the week immediately following her suicide. Some 85.9 per cent of the media reports occurred in this time frame. Research shows the copycat impact is strongest for a week to 10 days after the suicide.
National, reliable data on suicide trends in the US during the month of substantial coverage of Robin Williams's death, will not be available for at least three years. However, some hypotheses can be generated on its impact. It can be assumed that any increase in suicide is most likely to be among already suicidal persons of the same age and gender. For example, Hollywood star Charles Boyer's suicide at the age of 78 was followed by an increase in suicide among elderly males. The one risk factor that has the most evidence behind it involves the amount of media coverage of a suicide. The greater the coverage, the greater the likelihood of an increase in suicide in the nation.
The more Williams's suicide is discussed, if all else is equal, the greater the odds of a copycat effect. It is, however, doubtful that the impact will be as great as that of Monroe or Choi. They killed themselves at the peaks of their careers and popularity. In addition, the review of 419 findings in 55 studies determined that research that focuses on female suicide rates was 4.89 times more likely to find a copycat effect than other research. Williams's gender could conceivably prevent a record number of copycat deaths.
Concerns raised about some of the coverage of Williams shows the WHO's recommendations are still being ignored. The guidelines are in place for good reasons. Following them ought to be a priority in such high-profile cases.
Steven Stack is a professor in the departments of psychiatry and criminology at Wayne State University in Detroit, and a recipient of the Dublin Award from the American Association of Suicidology for his research