This paper discusses the fact that reliability is about scores and not tests and how reliability limits effect sizes. The paper also explores the classical reliability coefficients of stability, equivalence, and internal consistency. Stability is concerned with how stable test scores will be over time, while equivalence addresses the relationship between various forms of a test. Coefficient alpha is one of the common ways to describe the internal consistency of a test. The features that influence coefficient alpha are described, and the new "reliability generalization" meta-analytic technique is also summarized. (Contains 3 figures and 13 references.) (SLD)