Content area
Full text
TAPER, M. L. & S. R. LELE (EDS.), 2004. The Nature of Scientific Evidence: Statistical, Philosophical, and Empirical Considerations. University of Chicago Press, xviii + 567 p., paperback, US$30.00, ISBN 0-226-78957-8.
Statistics, inference, and hypothesis testing are essential ingredients in ecological studies. However, statistics and data analysis are painful for many ecologists and are at the heart of some of the most important controversies in the discipline. The Nature of Scientific Evidence is a an edited volume that includes 16 contributions by ecologists, statisticians, and philosophers of science on how we collect, analyze, and interpret ecological data.
The great strength of this collection is the ongoing conversation between scholars in these different disciplines, even though the participants are not always speaking the same language. The difficulty is that most ecologists (including myself) do not have the formal background necessary in statistics or philosophy to understand all of these papers. The book starts out in a promising way, with a clear introductory chapter by editors Mark L. Taper (an ecologist) and Subhash R. Lele (a statistician) that outlines three major methods of statistical analysis: Fisher's P-value tests, Neyman-Pearson tests, and Bayesian analysis. Each method is illustrated with the analysis of a simple data set on sex ratios in pig litters. However, not one ecologist in 20 (P < 0.05) will know that many statisticians make a distinction between Fisher's Pvalue, which estimate, the probability of data given a hypothesis, and Neyman-Pearson decision rules that distinguish between two alternative hypotheses. Both methods are incorporated into...