Abstract

Statistical methods are indispensable for scientific research. In forest sciences, the use of null hypothesis significance tests (NHSTs) has been the rule of thumb to judge hypotheses or associations among variables, in spite of the multiple problems of these techniques and the several criticisms published for many years in other scientific areas. In this review, the origin of current techniques, their most important problems, and some alternatives that are known to most forest researchers are shown. Persistence in using NHSTs, instead of better statistical methods or without adequate complements, could render our work inefficient and risky. Reasons for the permanence of NHSTs in forest sciences are discussed.