In the usual welter of uncertainty in gathering and interpreting scientific information for use in environmental decision-making, there will always be considerable probabilities of error. Precautionary principles absolutely require the uncertainty be "loaded" so that errors are more likely to result in decisions that "favours" the environment. "Crying wolf" is thought better, in the long run, than "she'll be right". This makes precaution all about power of tests and probabilities of Type II error. It therefore requires decision-makers to be made to identify concerns so that ecologist can consider realistic effect sizes. This makes it crucial for everyone to be more concerned with understanding variances. Environmental impacts are only definable as interactions, so the trick is to know how to calculate power to detect interactions.
Practical examples will be demonstrated, including issues of power and design when time-courses of impacts are complex and spatial scales of effects are uncertain. Refocussing attention on effect sizes turns out to be of great benefit to environmental decisions, so relevant ecological issues will be considered.
