Viewpoint: replication, randomization, and statistics in range research.

Abstract

Appropriate application of significance tests in statistical analyses requires an explicit statement of hypothesis; a clear definition of the population(s) about which inferences are to be made; and a model, a sampling strategy, an analysis, and an interpretation that are consistent with these considerations. In particular, experimental design and analyses must recognize appropriate replication and random selection of experimental units from target population(s). This paper discusses some aspects of these issues in range science research. Textbook examples and examples from range science applications are discussed in parallel in an attempt to clarify issues of randomization and replication in statistical applications.