Nuggets from a grumpy teacher

Doing Science is like a goldmine: it contains occasional nuggets, but one needs to move a lot of dirt to find them. In ploughing through nearly 300 pages, I learned more about Ivan Valiela than about the scientific method. He presents himself as a professor of the old school, charismatic yet curmudgeonly, possessing strong views, and probably loved and hated by equal numbers of his students and colleagues: a spectre of Albert Einstein crossed with Victor Meldrew.

The book sets itself the monumental goal of introducing "concepts needed inI the process of doing science". An admirable but ambitious aim. Even more ambitious is the intended audience: "anyone from high school students to graduate students, researchers, clinicians, regulators, administrators, and many others." One wonders who the many others could be. With such a huge topic and an impossibly large target audience, it is unsurprising that Valiela fails in his mission, but not without revealing the odd gem.

The first four chapters cover the scientific method: the principles of designing experiments and the processes of gathering information and analysing it. As Valiela is a marine biologist, much of his wisdom seems most applicable to the life sciences. We are told that "analysis of variance, regression, correlation, and frequenciesI are the most frequently encountered in the scientific literature". This might be true in the life sciences, but most physicists, astronomers, chemists and materials scientists will probably be left as nonplussed by this assertion as by the long discussions of small sample sizes, cross-classifications and replicates, not to mention being annoyed by the naive treatment of statistics. The publishers would have done better to have called the book "Doing Life Science".

My admiration for the difficulties of life-science experimentation grew in proportion to my horror at some of Valiela's assertions. "Unreplicated studies cannot... be disregarded" made my scientific flesh creep. "Awareness of the issues does not mean that we must follow every niggardly detail, test every assumption, or follow every statistical demand" left me appalled.

Chapters five to seven address scientific communication. Valiela pursues his vision of communication with the fervent piety of the non-native English speaker who has mastered the language. He hectors on the horrors of jargon, inaccurate word use, dangling pronouns and wordiness, culminating in a table of 40 "words and phrases that can be omitted or replaced with shorter ones". The polemic continues with rules on sentence structure, including noun clusters, dangling participles and weak antecedents. Never again shall I read a scientific paper without trepidation at uncovering such horrors. (Ironically, many of them can be found in the book.) The next three chapters discuss the presentation of results in tabular and graphical form. Valiela whips himself into a frenzy of self-righteousness over symbol sizes and labels, reserving special contempt for the pie chart ("excess fluff"). These overtones spoil the otherwise salutary examples. One wonders if the 36-page gratuitous harangue on "case studies of graphics" - dozens of examples of graphics that Valiela does not like - was really necessary.

The final chapter, "Perceptions and criticisms of science", is enjoyable. Having spent his fury, Valiela turns and offers some common-sense observations on the relationship between science and society, and a number of examples of human fallibility in the scientific domain. Despite the idiosyncrasies and quirkiness, it would be churlish to deny that I learned a number of useful things from Doing Science . But perhaps life scientists would find more nuggets than I did.

Philip Burrows is advanced fellow in the physics department, University of Oxford.