Biostatistical Design and Analysis Using R

 R is a powerful and flexible statistical and graphical environment that is freely distributed under the GNU Public Licencea for all major computing platforms (Windows, MacOSX and Linux). This open source licence along with a relatively simple scripting syntax has promoted diverse and rapid evolution and contribution. As the broader scientific community continues to gain greater instruction and exposure to the overall project, the popularity of R as a teaching and research tool continues to accelerate.

 It is now widely acknowledged that R proficiency as a scientific skill set is becoming increasingly more desirable and useful throughout the scientific community. However, as with most open source developments, the emphasis of the R project remains on the expansive development of tools and features. Applied documentation still remains somewhat sparse and somewhat incomprehensible to the average biologist. Whilst there are a number of excellent texts on R emerging, the bulk of these texts are devoted to the R language itself. Any featured examples therein are used primarily for the purpose of illustrating the suite of commonly used R features and procedures, rather than to illustrate how R can be used to perform common biostatistical analyses.

 Coinciding with the increasing interest in R as both a learning and research tool for biostatistics, has been the success of a relatively new major biostatistics textbook (Quinn and Keough, 2002). This text provides detailed coverage of most of the major statistical concepts and tests that biologists are likely to encounter with an emphasis on the practical implementation of these concepts with real biological data. Undoubtedly, a large part of the appeal of this book is attributable to the extensive use of real biological examples to augment and reinforce the text. Furthermore, by concentrating on the information biologists need to implement their research, and avoiding the overuse of complex mathematical descriptions, the authors have appealed to those biologists who don’t require (or desire) a knowledge of performing or programming entire analyses from scratch. Such biologists tend to use statistical software that is already available and specifically desire information that will help them achieve reliable statistical and biological outcomes. Quinn and Keough (2002) also advocate a number of alternative texts that provide more detailed coverage of specific topics and that also adopt this real example approach.