Using Classification and Regress Trees

By:
Xin Ma, University of Kentucky

Classification and regression frees (CART) is a relatively new statistical technique. Theoretical and particularly applied books on this topic are few. This book introduces CART analysis to faculties, researchers, data analysts, and graduate students in medicine (including nursing and public health), economics (including finance), behavioral sciences, and social sciences where CART is most promising as a powerful analytical tool. The goal is to make readers capable of (a) understanding the statistical (conceptual) framework underlying CART analysis, (b) performing CART analysis with the assistance of statistical software packages, (c) interpreting and reporting results from CART analysis, and (d) reading and critiquing research papers using CART analysis. Specifically, this book is applied statistics, intended to train readers (a) to become knowledgeable “consumers” of academic studies based on CART (i.e., being able to read, understand, and critique studies using CART) and (b) to become knowledgeable users of CART (i.e., knowing when, why, and how to apply CART, being able to perform a sound analysis with the assistance of statistical software packages, and being able to interpret and report results of the analysis). This book will emphasize its applied nature by using easy language and illustrative graphs. Examples selected from existing research studies will be offered to show (a) how to formulate research questions for CART analysis, (b) how to describe statistical procedures associated with CART analysis, and (c) how to present research findings from CART analysis. This book hopes to use these writing approaches to attract readers with minimum statistical background.