ABSTRACT
Hypothesis testing and confidence intervals play a prominent role in classical statistical texts. In this chapter, the authors describe the fundamental issues with hypothesis testing using a simple coin tossing example that demonstrates the simplicity and clarity of the Bayesian approach compared with the classical approach. They focus on the important problem of testing for hypothetical differences. The authors present the notion of Bayes' factors and model comparison. They deal with confidence intervals. Confidence intervals are a standard technique that is taught and used in classical frequentist statistics for all kinds of data analysis and decision making, few people using them understand what they actually mean. The authors show that how to do it properly using a Bayesian network (BN) approach. They also show how the BN approach provides more powerful and useful analyses when testing hypotheses about differences in populations.
