The Endocrine Society guideline on Androgen Deficiency in Men emphasized that accurate measurement of testosterone (T) levels is central to the diagnosis of androgen deficiency. Similarly, accurate measurements of testosterone levels are important in the diagnosis of androgen disorders in women and children. However, the accuracy of direct radioimmunoassays for the measurement of total T levels has been questioned, especially in the low range prevalent in women, children, and androgen deficient men. Furthermore, reference limits for total and free T levels generated in a population-based sample of community-dwelling men, women, and children are not available. In the absence of standardized reference limits, the partitioning of total and free T levels into normal, low, or high values is fraught with substantial risk of misclassification. The recommendations for partitioning of individuals into those with low, normal, or high levels should be based on considerations of statistical distribution of total and free T values and the association of outcomes with varying degree of deviations from the reference limits. Ongoing efforts to generate population-based reference ranges for total and free testosterone levels in men and women will provide a framework for the interpretation of serum T levels and enhance the comprehensibility of circulating T values to practicing clinicians. These steps will facilitate the development of rational criteria for the diagnosis of androgen disorders in men, women, and children.