Null Hypothesis Significance Testing: Philosophical and Practical Considerations of a Statistical Controversy

This article first outlines the underlying logic of null hypothesis testing and the philosophical and practical problems associated with using it to evaluate special education research. The article then presents 3 alternative metrics—a binomial effect size display, a relative risk ratio, and an odds ratio—that can better aid researchers and practitioners in identifying important treatment effects. Each metric is illustrated using data from recently evaluated special education interventions. The article justifies interpreting a research result as significant when the practical importance of the sample differences is evident and when chance fluctuations due to sampling can be shown to be an unlikely explanation for the differences.