Quantitative exponential modelling of copycat suicides: association with mass media effect in South Korea

Abstract

Background. There is ample evidence media reporting of celebrity suicides increases copycat suicides. This study had three aims: (a) to quantitatively examine copycat suicides with exponential modelling that predicts the copycat suicide effect of South Korean celebrity suicides; (b) to investigate the association between media effect and subsequent suicides following celebrity deaths; and (c) to investigate the extent in which media influences the increase and rate of decline of copycat suicides following a celebrity suicide.

Methods. All suicides during 1991-2010 in South Korea were included in this study utilising a nationwide database. Fifteen celebrities were selected based on the frequency of media reports following 1 week after their suicide. The media effect was obtained through the Korean Integrated Newspaper Database System. Exponential curve fits and correlation analyses investigated the quantitative effect of copycat suicides.

Results. After controlling for baseline number of average suicides, there was a marked increase in the number of suicides following each celebrity suicides, which followed an exponential model. There was a significant correlation between the total number of copycat suicides and number of media following the celebrity suicide (r = 0.74, p < 0.01). There were weak-to-moderate correlations between the amplitude of increase in suicides (r = 0.45, p = 0.09) and rate of decline (r = 0.38, p = 0.16) with the total number of media coverage.

Conclusions. Copycat suicides following celebrity suicides follow exponential modelling. Additionally, there is a strong media effect between the number of media reports following the days after celebrity suicides and subsequent copycat suicides. This may also be associated with the amplitude and rate of decline of copycat suicides. This suggests that improving media reporting and implementing preventative interventions for vulnerable populations may be important.