Ecology needs a causal overhaul.

Area of Science:

• Ecology
• Causal Inference
• Ecological Research

Background:

• Ecology frequently uses causal language but avoids explicit causal inference.
• The field relies heavily on observational data, necessitating robust causal inference methods.
• Current statistical practices in ecology often lack scientific clarity and value.

Purpose of the Study:

• To critique the current state of causal inference in ecology.
• To advocate for the adoption of explicit causal inference methodologies.
• To provide a constructive guide for ecologists to implement causal inference.

Main Methods:

• Critique of common pitfalls in ecological studies (e.g., "Table 2 fallacy", misuse of controls).
• Advocacy for integrating causal diagrams with standard statistical tools like regression models.
• Emphasis on causal inference as a scientific, not purely statistical, problem.

Main Results:

• Current ecological studies often employ scientifically empty statistical procedures.
• Explicit causal inference can be achieved using observational data with rigorous methods.
• Causal inference clarifies what variables to condition on (good controls) versus not (bad controls).

Conclusions:

• Ecology must embrace explicit causal inference to address its core questions.
• Rigorous causal inference, integrated with familiar statistical tools and causal diagrams, can enhance ecological research.
• Adopting causal inference will improve the scientific soundness and impact of ecological studies.