The Limits of Statistical Significance – Can we infer environmental significance from stats?
Presenter: Ilich Lama
We are bound by the limits of the sophisticated statistical methods we have developed to make sense out of our measurements and observations as we determine what is significant in life and nature. Statistical significance is a concept often used by practitioners to assess the existence of an environmental effect. Once an effect is deemed statistically significant, researchers tend to dismiss the possibility that it may be trivial. Conversely, when a statistically significant effect is not found, the tendency is to assume that the absence of statistical evidence is evidence of the absence of a meaningful effect. Both outcomes distort the scientific process and could result in irrelevant findings or devastating environmental consequences. In this presentation, we will provide the audience with an introduction to statistical significance, identify some of the critical shortcomings of conventional hypothesis testing, and briefly discuss how these concepts connect to the problem of detecting environmental changes that really matter, not merely those statistically noticeable.