A great multiplicity of methods has been used to evaluate the activity of natural antioxidants by using different techniques of inducing and catalyzing oxidation and measuring the end point of oxidation for foods and biological systems. Antioxidant in vitro protocols for foods should be based on analyses at relatively low levels of oxidation under mild conditions and on the formation and decomposition of hydroperoxides. For antioxidant in vivo protocols, widely different methods have been used to test the biological protective activity of phenolic compounds. Unfortunately, many of these protocols have been based on questionable methodology to accurately measure oxidative damage and to assess relevant changes in biological targets. Many studies testing the ex vivo activity of phenolic compounds to inhibit human low-density lilpoprotein (LDL) oxidation have been difficult to evaluate because of the structural complexity of LDL particles and because a multitude of markers of oxidative damage have been used. Although studies with animal models of atherosclerosis have demonstrated the antioxidant effect of phenolic compounds in delaying the progress of this disease, human clinical trials of antioxidants have reported inconsistent and mixed results. Complex mixtures of plant polyphenols have been shown to be absorbed to varying degrees as metabolites in the intestine, but little is known about their interactions, bioavailability, and their in vivo antioxidant activity. Several metabolites identified in human plasma after consuming flavonoids need to be tested for possible nonantioxidant activities. More research and better-designed human studies are required to clarify the complex questions of bioavailability of polyphenols and the factors affecting their in vivo activities. Until we know what relevant in vivo activities to measure, any claims on the biological and health protective effects of natural polyphenolic compounds in our diet are premature.