Some examples of spurious correlation in the literature

Some examples of spurious correlation in the literature
复制标题

文献中一些虚假相关的例子

DOI:
10.1002/hyp.14348
复制
发表时间:
2021
影响因子:
3.2
通讯作者:
Kirkby M
Kirkby M
中科院分区:
地球科学3区
文献类型:
--
作者:
Kirkby M

文献摘要

参考文献

相似文献

最近发表了几篇文章(Carollo等人,2021年;Di Stefano等人,2018年;Di Stefano等人,2019 a&b;Di Stefano等人,2020年;Nicosia等人,2019年a,b&c;Palmeri等人,2019年),声称显示了从一系列独立的陆上水流和细沟流量测量中获得的测量水力变量之间的有意义的关系。这些论文将实验数据插入到Darcy-Weisbach粗糙度与其他水力变量之间的关系中,其中包括流动雷诺数和弗劳德数。由此产生的异常高的相关系数被认为是一种有意义的一般关系。这里认为,作者无意中推导出了一个关系,而不是一个伪相关的例子(Brett,1996),因为回归的两个变量是强烈相互依赖的,在很大程度上,两者都来自相同的基础水力变量。因此,随机选择的任何合理(或不合理)变量集都将产生同样出色的相关性,不会提供对测量变量之间关系的新见解
Several articles (Carollo et al, 2021; Di Stefano et al, 2018; Di Stefano et al, 2019 a & b; Di Stefano et al, 2020; Nicosia et al, 2019; Nicosia et al, 2020 a, b & c; Palmeri et al, 2019) have been published recently that claim to show a meaningful relationship between the measured hydraulic variables taken from a series of independent overland flow and rill flow measurements. These papers insert the experimental data into a relationship between Darcy-Weisbach roughness and other hydraulic variables, including the flow Reynolds and Froude numbers. The resulting exceptionally high correlation coefficient is claimed to show a meaningful general relationship.It is argued here that the authors have inadvertently derived a relationship that is, instead, an example of spurious correlation (Brett, 1996) in that the two variables regressed are strongly interdependent, both, to a large extent, derived from the same underlying hydraulic variables. As a result, any reasonable (or unreasonable) set of variables chosen at random will generate the same excellent correlation, providing no new insight into the relationships between the measured variables
DOI: --
发表时间: 2020
期刊:
影响因子: --
作者:
C. Di Stefano;A. Nicosia;V. Palmeri;V. Pampalone;V. Ferro
通讯作者: V. Ferro