Statistical Methods: A Geometric Primer

Are you an applied researcher or a statistics student? If so, have you ever wondered if there is any rhyme or reason behind the apparent jumble of ad hoc statistical “recipes” that are most commonly used by researchers? If you have, then this book is for you! Within its pages, the rhyme and reason are explained in a surprisingly simple, visual and appealing manner. The statistical methods referred to are those based on the “Normal” distribution. They include the paired and independent samples t tests, analysis of variance, and regression. The rhyme is explained by analysing four case studies, one for each of these topics. The key to the rhyme is to use elementary vector geometry. Then the crucial ideas are easily explained by drawing right-angled triangles and applying Pythagoras’ Theorem. “Keep it simple” has been our aim. The geometric ideas are first introduced and applied to data sets in two and three dimensions, before we gently introduce higher dimensional examples. Throughout, we use real data sets taken from real research studies, and take care to explain the interpretation of the statistical analyses. In this way, the book integrates the algebraic, numeric, geometric and practical aspects of these basic statistical methods. We hope that many students as well as practicing researchers and statisticians will enjoy this novel visual approach to these ideas.

Dave Saville is a consulting statistician working in agricultural research; Graham Wood is a university professor involved in teaching statistical methods. They met and became friends when studying mathematics at the University of Otago, in Dunedin, New Zealand. Both have since worked for over forty years in their current fields.