This page was last modified 1 March 2004. It is maintained by Mick Keough
Statistical analysis is at the core of most modern biology, and many biological hypotheses, even deceptively simple ones, are matched by complex statistical models. Prior to the development of modern desktop computers, determining whether the data fit these complex models was the province of professional statisticians. Many biologists instead opted for simpler models whose structure had been simplified quite arbitrarily. Now, with immensely powerful statistical software available to most of us, these complex models can be fitted, creating a new set of demands ∓ problems for biologists.
We need to:
The analysis may be done by professional statisticians, rather than statistically trained biologists, especially in large research groups or multidisciplinary teams. In these situations, we need to be able to speak a common language -
This book aims to place biologists in a better position to do these things. It arose from our involvment in designing and analyzing our own data, but also providing advice to students and colleagues, and teaching classes in design and analysis. As part of these activities, we became aware, first of our limitations, prompting us to read more widely in the primary statistical literature, and second, and more importantly, of the complexity of the statistical models underlying much biological research. In particular, we continually encountered experimental designs that were not described comprehensively in many of our favourite texts. This book describes many of the common designs used in biological research, and we present the statistical models underlying those designs, with enough information to highlight their benefits and pitfalls.
Our emphasis here is on dealing with biological data - how to design sampling programs that represent the best use of our resources, how to avoid mistakes that make analyzing our data difficult, and how to analyse the data when they are collected. We emphasise the problems associated with real world biological situations.
Our approach is to encourage readers to understand the models underlying the most common experimental designs. We describe the models that are appropriate for various kinds of biological data - continuous and categorical response variables, continuous and categorical predictor or independent variables. Our emphasis is on general linear models, and we begin with the simplest situations - single, continuous variables - describing those models in detail. We use these models as building blocks to understanding a wide range of other kinds of data - all of the common statistical analyses, rather then being distinctly different kinds of analyses, are variations on a common theme of statistical modeling - constructing a model for the data and then determining whether observed data fit this particular model. Our aim is to show how a broad understanding of the models allows us to deal with a wide range of more complex situations. We have illustrated this approach of fitting models primarily with parametric statistics. Most biological data is still analyzed with linear models that assume underlying normal distributions. However, we introduce readers to a range of more general approaches, and stress that, once you understand the general modeling approach for normally-distributed data, you can use that information to begin modeling data with nonlinear relationships, variables that follow other statistical distributions, etc.
One of our strongest beliefs is that we understand statistical principles much better when we see how they are applied to situations in our own discipline. Examples let us make the link between statistical models and formal statistical terms (blocks, plots, etc.) or papers written in other disciplines, and the biological situations that we are dealing with. For example, how is our analysis and interpretation of an experiment repeated several times helped by reading a literature about blocks of agricultural land? How does literature developed for psychological research let us deal with measuring changes in physiological responses of plants?
Throughout this book, we illustrate all of the statistical techniques with examples from the current biological literature. We describe why (we think) the authors chose to do an experiment in a particular way, and how to analyse the data, including assessing assumptions and interpreting statistical output. These examples appear as boxes through each chapter, and we are delighted that authors of most of these studies have made their raw data available to us. We provide those raw data files on a web site, allowing readers to run these analyses using their particular software package. The other value of published examples is that we can see how particular analyses can be described and reported. When fitting complex statistical models, it is easy to allow the biology to be submerged by a mass of statistical output. We hope that the examples, together with our own thoughts on this subject, presented in the final chapter, will help prevent this happening.
From the preface to the book