A Primer of Ecological Statistics

In their preface the authors ask, “Why another book on statistics?” Their answer is that they have yet to find a single text that meets two specific needs of ecologists: a general introduction to probability theory including the assumptions of statistical inference and hypothesis testing and a detailed discussion of specific designs and analyses typical in ecology and environmental science. They suggest that their book should supplement more standard texts. Do ecology and environmental science deserve such special treatment? Yes. Many important problems in ecology lead to investigations where assumptions in standard techniques are not met, but students rarely receive further guidance in such cases. Experimental design in ecology is frequently more complex than the customary agricultural or biological examples provided in biometry texts. The authors devote two chapters to experimental design problems in ecology and provide considerable discussion of potential difficulties. The valuable chapter on multivariate analysis should provide enough background for ecology students to gauge suitability of this approach for their research. Chapter 5, “Three Frameworks for Statistical Analysis” is by far the most interesting and potentially the most revolutionary for teaching statistics to ecologists and environmental scientists. The authors pose the simple problem of estimating the number of ant colonies in a field and in a woodland. Hypothetical (unfortunately) data are presented of quadrat counts from both habitats, and then analyzed using a randomization procedure, a parametric analysis, and a Bayesian analysis. The concepts involved are well explained and the comparison in inferences is nicely illustrated. My only regret is that this approach was not developed and extended as a basis for the whole book. There are many advances in dependent data analysis and statistical modeling, which the authors mention in passing, that if included, could provide the ecology student with better analytical techniques and greater possibilities in designing and analyzing studies. The authors note that the majority of students do not have a facility in programming and urge them to get one. Perhaps that kept the authors from extensively connecting ecological research with current developments in statistics. However, if more recent developments are given a thorough exposition, perhaps that will stimulate ecologists to acquire the techniques necessary to use them.