THE STATISTICAL SLEUTH is a textbook for a second course in statistics, centered around case studies of real data problems to emphasize the process of data analysis and the communication of statistical conclusions.
AUDIENCE The Sleuth was written to train graduate students in disciplines other than Statistics to correctly draw and communicate statistical conclusions for their Master's and Doctoral theses, and for their eventual careers as scientists. It has also been used for advanced undergraduate courses and for graduate courses in behavioral science, biology, biostatistics, ecology, education, health sciences, and environmental sciences.
TOPICS The book begins with the two-sample problem to review procedures for statistical linference, and to introduce general ideas about the process of data analysis--particularly about robustness and resistance--and the communication of conclusions, including the connection between study design and the wording of conclusions; then proceeds to topics of regression, analysis of variance, generalized linear models, multivariate analysis, and study design.
WHAT'S NEW New topics in the third edition include reasoning fallacies associated with hypothesis tests, control of false discovery rates for genomics and data mining applications involving very large families of hypothesis tests, Monte Carlo simulation in data analysis, generalized estimating equations for generalized linear models with correlated responses, and negative binomial regression.