This chapter begins by describing the hypothesis statistics is designed to test. That hypothesis, known as the null hypothesis, states that things do not differ or there is no association between measurements. If that hypothesis is rejected, we conclude that there are differences or associations. Decisions to reject the null hypothesis are based on P-values. The chapter describes the origin and interpretation of P-values. It also discusses errors that can occur in interpretation of the P-value and how to control them. This discussion addresses the classical or frequentist approach to hypothesis testing. The Bayesian approach takes things further, allowing determination of the probability that the null hypothesis is false given frequentist methods have resulted in rejection of the null hypothesis. Both approaches are applied to the situation in which a study includes several hypothesis tests.