This excellent article by Douglas H. Johnson was published this summer in the Journal of Wildlife Management, 63(3):763-772 and is made available now in electronic version by the Northern Prairie Wildlife Research Center. In the article, Johnson argues that "statistical hypothesis tests add very little value to the products of research." Through a series of simple examples and clear text, Johnson criticizes several often-praised statistical hypothesis methods, and offers alternatives (in the form of estimation and confidence intervals, decision theory, and Bayesian approaches to "hypothesis testing and other statistical practices"). For any student or researcher of ecology, this article makes several fundamental points that should not be overlooked.
Comments