Experimental Design and Data Analysis for Biologists

A central part of reporting any scientific work is the presentation of the results, in either tabular or, more commonly, graphical form, and a considerable literature has accumulated about the appropriate ways for displaying quantitative data (e.g. Cleveland 1993, Tufte 1983, 1990, and some recent issues of The American Statistician). Much of this literature focuses on clarity of graphs and it is an issue that has become increasingly important as biologists do multifactorial experiments, often with complex underlying statistical models. We then face the problem of explaining those complex results to an audience that is pressed for time, and deluged by the number of papers published in any given month. In this environment, the presentation of your results becomes almost as important as the work itself, as you must convince a reader that he or she should persist with reading your paper, in the face of the many other demands on their time.

In many cases, the decision whether to read a paper completely is based initially on the title and abstract, which are provided by many of the electronic databases and the web. Having decided to look more closely at the paper, the next decision made is whether to persist with reading it. That decision will be made based in part on how clearly you express your ideas, and there is a long tradition of convincing scientists to write clearly, with several excellent and essential guides (e.g. Pechenik 2001, Strunk & White 1979, Williams 1997).