The correct interpretation of data is fundamental to the study of G-protein-coupled receptor pharmacology. Often, new assay technologies are assimilated into the drug discovery environment without full consideration of the data generated. In this study, the authors look at µ-opioid receptor agonists in three different assays: (1) [(35)S]GTPγS binding, (2) inhibition of forskolin-stimulated cAMP production, and (3) β-arrestin recruitment. Agonist-concentration effect curves were performed before and after treatment with the irreversible antagonist β-funaltrexamine, and where appropriate, these data were fitted to the operational model of agonism. The Z' value was highest in the β-arrestin assay, followed by the [(35)S]GTPγS and cAMP assays. The cAMP data fitted well to the operational model, as did the [(35)S]GTPγS data, but the [(35)S]GTPγS assay led to an apparent overestimation of K(A) values. However, in the β-arrestin assay, data did not fit the operational model, as treatment with β-funaltrexamine reduced the Emax proportionally to receptor number, with no change in EC(50). In addition, the EC(50) values generated correlated well with affinity values. In conclusion, the β-arrestin recruitment assay does not fit with traditional pharmacological theory but is of great utility as the EC(50) value generated is a good approximation of affinity.