Statistical Programs |
College of Agriculture | University of Idaho |
Seminar Announcement |
"Applied Statistics in Agriculture" |
Statistics and the
Scientific Method in Ecology
Presented By |
Dr. Brian C. Dennis |
Department of Fish and Wildlife Resources and Division of Statistics University of Idaho |
Tuesday, April 4 3:30 P. M. Ag. Science 323 |
The use of Bayesian statistical analyses has recently been
advocated in ecology, supposedly to aid decision makers and
enhance the pace of progress. Bayesian statistics provides
conclusions in the face of incomplete information. Bayesian
statistics, though, represents a much different approach to
science than the frequentist statistics studied by most
ecologists. The scientific implications of Bayesian statistics
are not well understood. In this presentation, I provide a critical review of the Bayesian approach. I compare, using a simple sampling example, the Bayesian and frequentist analyses. The Bayesian analyses can be "cooked" to produce results consistent with any point of view, because Bayesian analyses quantify prior personal beliefs and mix them with the data. In this, Bayesian statistics is consistent with the postmodern view of science, widely held among nonscientists, in which science is just a system of beliefs that has no particular authority over any other system of beliefs. By contrast, modern empirical science uses the scientific method to identify empirical contradictions in skeptics' beliefs and permit replication and checking of empirical results consistently across belief systems. Frequentist statistics has become an indispensable part of the scientific method. I also offer some critical remarks about statistics education in the ecological sciences. Part of the potential appeal of Bayesian statistics is that many ecologists are confused about frequentist statistics, and statistical concepts in general. I identify the source of the confusion as arising from ecologists' attempts to learn statistics through a series of precalculus "statistical methods" courses taken in graduate school. I prescribe a radical change in the statistical training of ecological scientists which would greatly increase their level of confidence and facility with statistical thinking. |
Return to Seminar Listing. |