The possibility that unreliability of plasma cortisol level determinations might interfere with the use of the dexamethasone suppression test (DST) in clinical practice was evaluated by examining published data on the reliability of plasma cortisol determinations as well as a comparison of two radioimmunoassay (RIA) methods with a competitive protein binding assay (CPBA). Significant discrepancies between various methods of plasma cortisol level determinations have been reported in most, but not all, studies. The most widely used method, RIA kits, may not produce levels comparable with those produced by the CPBA, used in a recent effort to standardize the DST for use in psychiatry. The use of the DST by physicians should be preceded by an assessment of the validity of the cortisol level determinations in the critical range for the DST by the laboratory they use.