| Review: |
This book aims to introduce statistical methodology: estimation, hypothesis testing and classification, to a wide range of readers, by sampling the data in hand. The resampling methods - permutations, cross-validation, and the bootstrap - require no mathematics beyond high-school algebra, and are easy to apply. For those new to using statistics in their work, the first five chapters of the book cover descriptive statistics, cause and effect, sampling, hypothesis testing, and estimation. Chapter 6 looks at the formal properties (the significance, level, power and robustness) of the tests of hypotheses that have been discussed. The next four chapters cover: testing hypotheses that concern categorical data and ordinary data; experimental design and analysis to investigate multiple factors simultaneously; multiple variables and multiple hypotheses; classification and discrimination in assigning categories to items or individuals; and survival analysis and reliability. The last chapter provides an expert system which will aid the reader in deciding which statistic to use. |