The New Statistics With R – An Introduction for Biologists
Oxford University Press, £29.99
Andy Hector has 20 years' experience in teaching statistics. So what does he mean by 'the new statistics' and what is 'R'? The title is not misleading: 'R' is free software for statistical computing, but is an unknown quantity to many biologists.
This little book provides a working guide to using some of the free statistical packages available on the internet within the R envelope and discusses how best to compare data sets using outputs from statistical routines within R. Because I am one of the 'broadband disadvantaged rural minorities', I had avoided using the R statistics package at home, but assessment of the book provided a stimulus to evaluate some of the statistical routines described.
However, the purpose of the book is not just to introduce the user to R. It aims to encourage biologists to reconsider how they determine and describe differences in observations following different treatments. I endorse the author's recommendations to describe variations between data sets in terms of confidence intervals rather than mindless reporting of probability significance levels.
The style of The New Statistics With R is typically didactic and provides worked examples of the analysis of published data. The examples are predicated on the use of linear models including analysis of variance, linear regression, covariance, maximum likelihood and generalised linear models, and mixed-effect models.
The book is suitable for undergraduate and graduate students, researchers and practitioners in biological sciences. I found it refreshing and worthy of wide use.