School of Engineering and Science, Jacobs University Bremen, Campus Ring 1, D-28759 Bremen, Germany.
Mutat Res. 2010 Mar 29;697(1-2):60-5. doi: 10.1016/j.mrgentox.2010.01.010. Epub 2010 Jan 25.
In a publication that appeared in 2005 (Diem et al., Mutat. Res. 583:178-183) [10] harmful effects (DNA breakage) were reported to occur in rat and human cells after exposure to mobile-phone electromagnetic fields. The extremely low standard deviations in this paper, and in another publication by the same group of authors, prompted Vijayalaxmi to write a critical comment [Mutat. Res. 603 (2006) 104-106] [16]. An investigation by the Medical University of Vienna (Austria) was initiated by a letter by the first author of the present paper, based on the data contained in the reply by the authors [Rüdiger et al., Mutat. Res. 603 (2006) 107-109] [17]. The University published three press releases, stating that "the data were not measured experimentally, but fabricated" and that the Mutation Research paper and another, published by the International Archives of Occupational and Environmental Health (IAOEH) in 2008, should be retracted. So far, neither of these papers has been retracted. Only a Letter of Concern by the Editors of IAOEH, and an Editorial by Mutation Research were published. Here we describe the statistical methods used to identify the evidence of data fabrication. The major point is the small variation in the reported data, which is below the theoretical lower limit derived from multinomial distributions and also lower than those derived from detailed simulations. Another reason for doubt was the highly significant non-equal distribution of last digits, a known hint towards data fabrication. In view of the results of the University's investigation and the evidence presented in this paper, the Diem et al. (2005) [10] publication should be retracted, with or without the authors' agreement.
在 2005 年发表的一篇出版物中（Diem 等人，《突变研究》583:178-183）[10]，报告称在暴露于移动电话电磁场后，大鼠和人细胞中会发生有害影响（DNA 断裂）。该论文和同一组作者的另一篇出版物中，标准差极低，这促使 Vijayalaxmi 发表了一篇批评性评论[《突变研究》603（2006）104-106] [16]。维也纳医科大学（奥地利）根据本论文第一作者的一封信启动了一项调查，该信基于作者回复中包含的数据[Rüdiger 等人，《突变研究》603（2006）107-109] [17]。该大学发布了三份新闻稿，称“这些数据不是通过实验测量的，而是捏造的”，并且《突变研究》论文和 2008 年在《国际职业与环境卫生档案》（IAOEH）上发表的另一篇论文应该撤回。到目前为止，这两篇论文都没有被撤回。只发表了《IAOEH 编辑的关注函》和《突变研究》的一篇社论。在这里，我们描述了用于识别数据捏造证据的统计方法。主要观点是报告数据的变化很小，低于从多项分布推导出的理论下限，也低于从详细模拟推导出的下限。另一个令人怀疑的原因是最后一位数字的高度显著非均等分布，这是数据捏造的已知提示。鉴于该大学的调查结果和本文提出的证据，Diem 等人（2005）[10]的出版物应撤回，无论作者是否同意。
