Fallacy of per-weight and per-surface area standards, and their relation to spurious correlation.
Fallacy of per-weight and per-surface area standards, and their relation to spurious correlation.
复制标题
按重量和按表面积标准的谬误及其与虚假相关性的关系。
DOI：
10.1152/jappl.1949.2.1.1
复制
发表时间：
1949
影响因子：
3.3
通讯作者：
J. Tanner
中科院分区：
文献类型：
--
作者：
J. Tanner
在生理学和临床医学中，一些测量结果（例如耗氧量和心输出量）通常表示为单位重量或单位表面积比率。事实上，这两个变量的正常标准都是在此基础上构建的；随着越来越多的类似变量在人体上进行测量，以这种方式制定的标准和报告的结果的数量预计会增加。因此，指出这些标准在理论上是错误的，并且在实践中（除了下面讨论的非常特殊的情况下）具有误导性，似乎是有用的。所涉及的谬误可以被视为统计学家众所周知的指数虚假相关性的一个特例。作者在制定人类心输出量的一些新标准时首次意识到这种情况（I）。对文献的粗略回顾表明，使用这些比率的后果根本没有被广泛认识到。一些例子立即浮出水面，其中研究人员得出了未经数据证实的积极结论，被结果中看似无法解释的现象所困惑，提出了一种效果较差且偏差较大的正常标准，而不是更有效且偏差较小的正常标准，将两种生理功能之间的相关性从非常高降低到中等值，并发明了一种新的临床综合症。所有这些事件全部或主要是由于比率使用不慎造成的。本文的后半部分给出了每个例子，以表明接下来的理论讨论具有非常实际的结果。这些例子涉及心输出量、成人和儿童的基础代谢标准、体型和耗氧量的关系以及男性血浆容量标准的选择；在狗中，统计肾血浆流量和肾小球滤过率。
Both IN PHYSIOLOGY and clinical medicine, the results of some measurements, for example, of oxygen consumption and cardiac output, are commonly expressed as per-weight or per-surface area ratios. Normal standards for both these variables, in fact, have been constructed on this basis; and as more similar variables come to be measured on the human being, the number of standards constructed and of results reported in this way, may be expected to increase. It seems useful to point out, therefore, that such standards are theoretically fallacious, and in practice (except under very special circumstances discussed below) misleading. The fallacy involved may be considered as a special case of that well known to statisticians as the spurious correlation of indices. The writer first became aware of this situation when constructing some new standards for cardiac output in man (I). A cursory review of the literature revealed that the consequences of using these ratios were not at all widely realized. Examples immediately came to light where investigators had drawn positive conclusions not justified by their data, had been confused by a seemingly uninterpretable phenomenon in their results, had proposed a less effective and more biased normal standard in preference to a more effective and less biased one, had reduced a correlation between two physiological functions from a very high to a medium value, and had invented a new clinical syndrome. All these events were wholly or in major part due to the incautious use of ratios. Examples of each are given in the second half of this paper, by way of making the point that the theoretical discussion which now follows has a severely practical upshot. The examples concern cardiac output, basal metabolism standards in adults and children, the relation of body build and oxygen consumption, and the choice of plasma volume standards in man; and in the dog, the statistics of renal plasma flow and glomerular filtration rate.