Overcoming the pitfalls of categorizing continuous variables in ecology, evolution and behaviour.

Area of Science:

• Ecology
• Evolutionary Biology
• Behavioral Science

Background:

• Continuous variables (e.g., body mass) are often analyzed as categories (e.g., large vs. small) in biological research.
• This practice can decrease statistical power and alter research interpretations.

Purpose of the Study:

• To quantify the prevalence of categorizing continuous variables in ecology, evolution, and behavior journals.
• To demonstrate the negative impacts of this categorization on statistical power and interpretation.
• To provide recommendations for maintaining continuous variable analysis.

Main Methods:

• A mini-review of 72 publications across six prominent journals.
• Summarization of frequently categorized metrics.
• Simulation of a dataset using common variables to illustrate categorization drawbacks.

Main Results:

• Categorization of continuous variables was found in 31% of reviewed publications.
• Analysis showed that categorization reduces statistical power and can change scientific conclusions.
• Continuous variables can and should be collected and analyzed as such.

Conclusions:

• Keeping variables continuous throughout the scientific process is crucial for increasing statistical power.
• This approach facilitates larger synthesis studies and more trustworthy conclusions.
• Adopting continuous variable analysis is essential for accurate predictions regarding climate change impacts.