Morton R H
Department of Statistics, Massey University, Palmerston North, New Zealand.
J Sports Sci. 1993 Oct;11(5):371-5; discussion 377-8. doi: 10.1080/02640419308730008.
A recent paper in this journal has examined the ventilatory compensation for metabolic acidosis with increasing oxygen uptake, making extensive use of mathematical and statistical techniques. The authors conclude that '...respiratory compensation for metabolic acidosis during incremental exercise is a continuous process...'. While this may indeed be so, the following major aspects of the paper render the evidence at best dubious, surely inconclusive and perhaps even worthless: (1) The mathematical and statistical methodology is sloppy, incomplete and, at times, flawed. (2) No discontinuous model is presented for comparative purposes, only three continuous ones. (3) The authors ignore the fact that some threshold models produce smooth continuous data, and as a consequence, a good fit of a smooth continuous model cannot exclude the presence of a threshold. (4) The data in Table 5 show the amplitude parameter A of the preferred model to be not significantly different from zero. This paper draws these matters to the attention of readers of this journal, reviewers of manuscripts and other researchers, with comment and suggestion. This is done in the hope that all those concerned with mathematical and statistical methodologies will become more aware of their proper use.
该期刊最近发表的一篇论文运用了大量数学和统计技术，研究了随着摄氧量增加对代谢性酸中毒的通气补偿情况。作者得出结论：“……递增运动期间对代谢性酸中毒的呼吸补偿是一个持续的过程……”。虽然情况可能确实如此，但该论文存在以下几个主要问题，使得其证据充其量可疑，肯定无法得出结论，甚至可能毫无价值：（1）数学和统计方法草率、不完整，且有时存在缺陷。（2）未提出用于比较的非连续模型，仅给出了三个连续模型。（3）作者忽略了这样一个事实，即一些阈值模型会产生平滑的连续数据，因此，平滑连续模型的良好拟合并不能排除阈值的存在。（4）表5中的数据显示，首选模型的幅度参数A与零没有显著差异。本文提请本期刊的读者、稿件评审人员和其他研究人员关注这些问题，并给出了评论和建议。这样做的目的是希望所有关注数学和统计方法的人能更加意识到其正确用法。
