Correlation and Causation in Fisheries and Watershed Management

By Ray Hilborn
School of Aquatic and Fishery Sciences, Box 355020, University of Washington, Seattle, WA 98195. E-mail: [email protected]

Efforts to understand how to manage aquatic ecosystems often rely on correlations between human actions and impacts in the ecosystem. We are often warned that correlation does not imply causation and that the gold standard for identify- ing cause and effect relationships is manipulative experiments. History shows us that correlations are often not causal and that managers should not design policies based on the assumption of causality. However, in the absence of manipulation, correlative evidence may be all that is available. Correlative evidence is strongest when (1) correlation is high, (2) it is found consistently across multiple situations, (3) there are not competing explanations, and (4) the correlation is consist- ent with mechanistic explanations that can be supported by experimental evidence. Where possible, manipulative experi- ments and formal adaptive management should be employed, but in large-scale aquatic ecosystems these opportunities are limited. More commonly, we should emphasize identifying the range of possible causal mechanisms and identify poli- cies that are robust to the alternative mechanisms.

Members click below for the January 2016 Fisheries magazine’s complete issue. Non-members, join here.

This content is for members only. Please login.