Inappropriate use of statistical methods leads to unsupported conclusions about risk of plague to an imperiled chipmunk: A critique of Goldberg et al. (2022)

Abstract

Plague is a zoonotic disease caused by the bacterium Yersinia pestis that can cause endangerment or extinction of wildlife populations. A recent study by Goldberg et al. (2022) sought to determine if plague was circulating within the small mammal communities associated with the Peñasco least chipmunk, which has been proposed for listing as endangered on the Endangered Species Act. We identified substantial problems with the statistical approaches used by Goldberg et al., including testing more models than the data can support and failing to recognize models with uninformative parameters. Contrary to their findings, our reexamination of their analyses and results suggests little evidence for plague in this system and that prophylactic measures to prevent plague do not warrant the risks that implementing such a program would have on the chipmunk. We recommend that future studies directly assess the presence of plague in both the associated rodent community and sentinel species. We also recommend that researchers carefully consider the goals of their study and collaborate with statistical/quantitative ecologists to ensure that the study design is appropriate and capable of generating the necessary sample sizes. We strongly emphasize the recommendations made by other researchers about the importance of adequately reporting methods and results in scientific studies. We also urge researchers to evaluate models for uninformative parameters to avoid erroneous conclusions that could potentially lead to wrong management decisions.

Keywords

• Akaike information criterion
• Information theory
• Model selection
• Neotamias minimus atristriatus
• Survival
• Uninformative parameters