Moving Beyond p < 0.05 in Ecotoxicology: A Guide for Practitioners

Area of Science:

• Ecotoxicology
• Statistical Inference
• Environmental Science

Background:

• Null hypothesis significance testing (NHST) has been the primary statistical method in ecotoxicology.
• NHST relies on p-values to determine statistical significance, but has faced long-standing criticisms.
• Incorrect reporting of p-values persists in ecotoxicology literature, as evidenced by reviews of Environmental Toxicology & Chemistry (ET&C).

Purpose of the Study:

• To highlight the limitations of NHST and p-values in ecotoxicology.
• To advocate for the adoption of alternative statistical inference methods.
• To explore the future of statistical practices in ecotoxicology beyond p-values.

Main Methods:

• Review of statistical reporting practices in ecotoxicology journals.
• Discussion of the criticisms and limitations of NHST.
• Description of alternative statistical methods relevant to ecotoxicology.

Main Results:

• A review of the 2019 ET&C volume revealed continued incorrect reporting of p-values.
• The American Statistical Association (ASA) has recommended against the use of "statistical significance" in 2019.
• There is a clear need for ecotoxicologists to adopt alternative statistical approaches.

Conclusions:

• Ecotoxicology must move beyond traditional NHST and p-values.
• Alternative methods such as confidence intervals, regression analysis, and Bayes factors offer more reliable inferences.
• Adopting new statistical tools will advance the field of ecotoxicology.