Book Review: Biostatistics for Animal Science

The book Biostatistics for Animal Science by Kaps and Lamberson is an excellent textbook or reference book on biometrics for graduate students or researchers in animal science. The book is easy to read and well formatted throughout with numerous easy to follow examples from various disciplines in animal science.

Generally, textbooks on biometrics are deficient in animal science examples and rely heavily upon agronomic examples; however, this book is replete with real-life animal science examples with detailed explanations. Especially helpful is the utilization of SAS© programs and outputs to elucidate and augment the principles and computations involved in each of the examples. SAS© is a powerful statistical computing package that is oft used, and sometimes misused, in data analysis. This text with its philosophy of presenting the principles and theory first and foremost and then using SAS© to illustrate the application of those principles and theories is commendable. Most often, one finds texts on statistical principles and theory or texts that detail SAS statements and procedures, but rarely does one find a text that merges the two as well as Biostatistics for Animal Science has done.

The initial chapters lay the groundwork of the basic principles and theories of statistics and distribution theory that are the underpinnings of the correct interpretation of hypothesis testing and data analysis. The description of these principles and theories are straightforward without becoming mired in statistical jargon that can easily confuse those who are not statisticians. The chapters dedicated to regression analysis and correlation progress in an appealing and natural manner, starting with simple linear regression and ending with nonlinear regression models. The section of Problems with Regression is very informative for graduate students and researchers alike when troubleshooting experimental or field data. For graduate students and researchers with limited experience in data validation or testing for outliers, the subsection on Extreme Observations is particularly useful. Another aspect of regression analysis not often discussed thoroughly, but that this text explains well is that of curvilinear relationships between the dependent and independent variables. This text does an outstanding job of differentiating between a curvilinear (or nonlinear) response surface and a nonlinear model, which is sometimes hard to understand for those not with a statistical background. The chapters detailing analysis of variance are comprehensive and complete. As with regression analysis, the progression for analysis of variance is logical and straightforward. The text begins with one-way analysis of variance with all of its nuances, including a random effects model, and terminates with repeated measures. In between the text systematically discusses and illustrates blocking, change-over designs, which includes a detailed discussion of Latin squares, factorials, nested designs, double blocking, split-plots, and analysis of covariance, which includes a test for heterogeniety with example. As an added bonus, the text in its final chapters covers discrete dependent variables using logit and probit models for binary and binomial variables.

In summary, Biostatistics for Animal Science is an excellent text and should be on every animal scientist's bookshelf.