Designed for a graduate course in applied statistics, Nonparametric Methods in Statistics with SAS Applications teaches students how to apply nonparametric techniques to statistical data.
It starts with the tests of hypotheses and moves on to regression modeling, time-to-event analysis, density estimation, and resampling methods. The text begins with classical nonparametric hypotheses testing, including the sign, Wilcoxon sign-rank and rank-sum, Ansari-Bradley, Kolmogorov-Smirnov, Friedman rank, Kruskal-Wallis H, Spearman rank correlation coefficient, and Fisher exact tests.
It then discusses smoothing techniques (loess and thin-plate splines) for classical nonparametric regression as well as binary logistic and Poisson models.
The author also describes time-to-event nonparametric estimation methods, such as the Kaplan-Meier survival curve and Cox proportional hazards model, and presents histogram and kernel density estimation methods.
The book concludes with the basics of jackknife and bootstrap interval estimation. Drawing on data sets from the author's many consulting projects, this classroom-tested book includes various examples from psychology, education, clinical trials, and other areas.
It also presents a set of exercises at the end of each chapter.
All examples and exercises require the use of SAS 9.3 software.
Complete SAS codes for all examples are given in the text.
Large data sets for the exercises are available on the author's website.