Moore C G
J Am Mosq Control Assoc. 1998 Dec;14(4):482-4.
This letter questions the appropriateness of methodology used in a study by Howard and Oliver (J. Am. Mosq. Control Assoc. 13:315-325; 1988). Two independent data sets, collected for different purposes by 2 different groups, were subjected to statistical analysis to determine if the data sets differed. The experimental "design," as described by the authors, is an example of pseudoreplication, which arises when replicates are collected at a scale finer than the one for which conclusions of statistical testing are intended to be drawn. All of the components of a properly designed field experiment (control, replication, randomization, and interspersion) are missing from this study. The authors proceed to draw a series of conclusions from the data presented. Few, if any, of the conclusions can be supported by the evidence presented. The assertions put forward in this paper could have a severe negative impact on efforts to prevent transmission of arboviruses or other pathogens to humans and domestic animals.
这封信对霍华德和奥利弗（《美国蚊虫控制协会杂志》13:315 - 325；1988年）一项研究中所使用方法的恰当性提出质疑。两个独立的数据集由两个不同的小组出于不同目的收集，随后进行了统计分析以确定这两个数据集是否存在差异。作者所描述的实验“设计”是伪重复的一个例子，当重复样本的收集尺度比旨在得出统计检验结论的尺度更精细时就会出现这种情况。这项研究缺少一个设计合理的野外实验的所有要素（对照、重复、随机化和散布）。作者接着从所呈现的数据中得出了一系列结论。所呈现的证据几乎无法支持其中的任何结论。本文所提出的论断可能会对预防虫媒病毒或其他病原体传播给人类和家畜的努力产生严重的负面影响。
