One of the most frequent and perhaps frustrating questions the pathologist is called on to answer regarding laboratory tests is, "What is normal?" In recent years the term "normal range" has been replaced by "reference range." "Upper limit of normal" has been replaced by "referent value," "critical value," or "cut-off point." Our literature now reflects increasing contributions from biostatistics, decision theory, and systems analysis. Yet, for most of us, although the names have changed, the problem has not. Understanding Bayes' Theorum will not help the busy internist when he complains about false-positive elevations in alkaline phosphatase activity on hospital admission profiles. The laboratory director must be able to integrate reference range data and predictive value theory in his own environment if he wishes to provide meaningful information to clinicians. It is the objective of this review to focus on the definition and integration of these areas of emerging concern for the laboratorian.