Reanalysis of the Bridge et al. study of suicide following release of 13 Reasons Why
Reanalysis of the Bridge et al. study of suicide following release of 13 Reasons Why
复制标题
DOI：
10.1371/journal.pone.0227545
复制
发表时间：
2020-01-16
期刊：
影响因子：
3.7
通讯作者：
Romer, Daniel
中科院分区：
文献类型：
--
作者：
Romer, Daniel
Bridge等人最近在2017年Netflix系列“13 Reasons Why”发布后，对美国自杀率进行了时间序列分析。“他们的分析发现，该节目在2017年4月发布后的9个月内对10-17岁的男孩产生了强大的影响。我对这一发现提出了两点质疑。首先，在这个故事中，女孩的感染力预计会比男孩更强，其次，他们的分析没有考虑到自杀的强烈长期趋势，特别是在2016年至2017年的男孩中。我使用一个简单的自回归模型重新分析了他们的数据，该模型在去除自相关性和全国自杀趋势后测试了自杀率的变化。我发现，布里奇等人在4月份观察到的男孩人数的增加并不比节目上映前一个月观察到的增加多。这一年的后几个月也没有影响。对于女孩来说，我发现4月份自杀人数略有增加，但并不显著，这是那个月所独有的，可能与该节目的保护性和有害性相结合。总之，我的结论是，很难用每月自杀率的总和来归因于该节目的有害影响。在每周水平上进行更精细的分析可能更有效，但只有在控制了自2008年以来美国自杀率特别强劲的长期变化之后。
Bridge et al. recently presented a time series analysis of suicide rates in the US following the release of the 2017 Netflix series "13 Reasons Why." Their analysis found a powerful effect of the show on boys ages 10-17 for nine months after the show was released in April 2017. I questioned this finding on two grounds. First, contagion would be expected to be stronger for girls than boys for this story, and second their analysis did not take into account strong secular trends in suicide, especially in boys from 2016 to 2017. I reanalyzed their data using a simple auto-regression model that tested for changes in rates after removing auto-correlation and national trends in suicide. I found that the increase for boys observed by Bridge et al. in April was no greater than the increase observed during the prior month before the show was released. There were also no effects in later months of that year. For girls, I found a small but nonsignificant increase in suicide in April that was unique to that month, potentially consistent with a combined protective and harmful effect of the show. In total, I conclude that it is difficult to attribute harmful effects of the show using aggregate rates of monthly suicide rates. More fine-grained analyses at the weekly level may be more valid but only after controlling for secular changes in suicide that have been particularly strong since 2008 in the US.