An Introduction to Evidence Based Statistics

This book provides an accessible exposition of the likelihood approach to undergraduate and postgraduate students in biomedical, biological and social sciences. The book contains 9 chapters, beginning with an introduction on the limitations of P values and a description of how the likelihood approach better represents statistical evidence. Chapter 1 covers the definition of effect size, its different measures, and importance. Next, Chapter 2 examines the likelihood approach, covering how the likelihood ratio is driven by increased data either to support or contradict one or the other hypothesis. Related and independent samples with t are covered in Chapter 3, while ANOVA is the theme of chapter 4. This chapter uses the Glover & Dixon approach to compare two competing models. Chapter 5 discusses correlation and regression, while Chapter 6 examines the Armitage stopping rule. The author examines equivalence tests in Chapter 7, using likelihood and confidence intervals to determine the absence of an effect using Lakens’ approach with modification of Cohen’s d. The likelihood ratio for one-way and two-way categorical data is covered in Chapter 8. Finally, Chapter 9 concludes the text with other useful techniques, such as the minimum Bayes factor and the false positive risk.