Bioluminescence resonance energy transfer (BRET) is often used to study association of membrane proteins, and in particular oligomerization of G protein-coupled receptors (GPCRs). Oligomerization of class A GPCRs is controversial, in part because the methods used to study this question are not completely understood. Here we reconsider oligomerization of the class A β2 adrenergic receptor (β2AR), and reevaluate BRET titration as a method to study membrane protein association. Using inducible expression of the energy acceptor at multiple levels of donor expression we find that BRET between β2AR protomers is directly proportional to the density of the acceptor up to ~3,000 acceptors μm(-2), and does not depend on the density of the donor or on the acceptor:donor (A:D) stoichiometry. In contrast, BRET between tightly-associating control proteins does not depend on the density of the acceptor, but does depend on the density of the donor and on the A:D ratio. We also find that the standard frameworks used to interpret BRET titration experiments rely on simplifying assumptions that are frequently invalid. These results suggest that β2ARs do not oligomerize in cells, and demonstrate a reliable method of assessing membrane protein association with BRET.