What should we do about hypothesis testing?

A sizable list of recently published papers criticized the use of hypothesis testing in ecology and wildlife management. I suggest that early neglect of statistical methods in those areas led to excesses, and that the real problem lies in the use of methods designed for controlled experiments in uncontrolled settings. Controlled experiments can be conducted in wildlife research, and the revolutionary use of controlled experiments in medical research indicates their utility and need. Some alternative paradigms (sampling, modeling, population analysis) are less troubled by the hypothesis-testing issues. The recent interest in a model-selection paradigm is warranted but should not be considered a universal solution.