Inference and Generality in Ecology: Current Problems and an Experimental Solution

摘要: Two major points are considered: I) the growth of ecology is constrained and scientific credibility ecologists reduced, because too many generalizations accepted on insufficient evidence, 2) there experimental methods for addressing this problem. I illustrate problems that arise from theories that, although valid in limited domains, have been excessively generalized, examine decreasing disparity populations considered between statistical inferences generalizations. The use random factors designs helps to address domain which valid, population encompassed by much larger than addressed typically fixed factors. particular mixed-model ANOVAs, basic techniques well described familiar most ecologists. Mixed-model provide a strong framework with other assessing generality ecological theories. No single design will be sufficient define broader generalities seek, but these should included more often research programs.