Agricultural Experimentation: Design and Analysis
Design & analysis
Bibliographic details
Additional authors
Hills, Frederick Jackson
Summary
Few agricultural research workers have the time to master the details of abstract and sophisticated mathematics, yet they would like to gain a general understanding of the logic and reasoning involved in the designing and conducting of experiments. In short, they would like to learn engouh of the basic principles of statistics to be able to design experiments properly and to draw valid conclusions from the results. This book is written to satisfy these needs. Mathematics beyond simple arithmetic has been kept to a minimum. Many of the mathematical relations are presented simply as facts without formal proof. Yet, every effort has been made to make the discussion mathematically correct and to avoid the dangers of oversimplification. Separate chapters are devoted to each of seven experimental designs that probably constitute over 90 percent of the designs used in agricultural research. Four chapters are devoted to coorelation and regression (linear, curvilinear, and multiple). Regression is presented in the context of the analysis of variance as well as a technique used in survey-type research. A special feature is a chpater on the use of shortcut methods for handling regression when the experimental treatments or observations are equally spaced. For each technuque emphasis is on a detailed step-by-step procedure for computing the essential statistics. The spiral binding has been chosen so the book will lay open to enable users to follow a procedure in relation to their own work. Yet, this is much more than a cookbook. In each case, the logic and reasoning behind the analysis is explained. An entire chapter is devoted to the assumptions underlying the analysis of variance and the ways of handling data that do not satisfy these assumptions. Special emphasis is given to the subject of mean separation (determining which of several means are significantly different). It is evident from the current agricultural research literature that there is a great deal of misunderstanding about this subject. As a result, important conclusions justified by the data are often overlooked. This is especially true with regard to the method known as the functional analysis of variance or the method of orthogonal coefficients. This extremely simple, yet powerful technique is unfamiliar to a large proportion of agricultural research workers. This and other methods of mean separation are fully discussed.