A Philosophy for Effective Monitoring

Monitoring programs do not always provide useful feedback about the effects of human activities on the natural environment. To do so, they must transcend purely descriptive efforts in order to integrate ecological insight, statistical rigor, and questions about policy options. The principles of statistical experimental design provide a framework for this in three ways. Firstly, they make it easier to acknowledge uncertainty, and to recognize that many decisions are experiments that require feedback about their outcomes. Secondly, statistical models enforce the statement of clear-cut hypotheses that structure sampling and analysis. Finally, the language of statistics permits the measurement and optimization of the chance of a false alarm (alpha error), and the chance of missing a real effect (beta error). This permits available resources to be used so as to maximize the information return per effort expended.