In determining the efficacy of phosphoinositides on ion channels in inside-out patches, we and others typically measure channel activity as a function of the solution concentration of short-chain phosphoinositides such as di-C8-PI(4,5)P2. However, channels are believed to bind phosphoinositides not from solution, but from the phosphoinositide lipids incorporated into the plasma membrane. The extent to which different phosphoinositide lipids partition into membranes is expected to vary widely depending on charge, chain length, and head-group structure, such that the same solution concentration of different phosphoinositides will produce very different concentrations in the plasma membrane. We have previously shown that the polymodal ion channel TRPV1 has a higher apparent affinity for di-C8-PI(4,5)P2 than for diC8-PI(4)P. We re-examined this apparent PI(4,5)P2 selectivity using a model that incorporates an established description of phosphoinositides' partition between solution and the lipid membrane. We found that the data could not distinguish between two very different interpretations: that the channel itself is activated more easily by di-C8-PI(4,5)P2 than di-C8-PI(4)P; or that di-C8-PI(4,5)P2 and di-C8-PI(4)P are equally effective in activating TRPV1, but for a given phosphoinositide lipid solution concentration, there is more di-C8-PI(4,5)P2 in the membrane than there would be di-C8-PI(4)P. We further tested our model on a channel mutant that appears to invert the selectivity for di-C8-PI(4,5)P2 and di-C8-PI(4)P. Our examples have wide ranging implications for the interpretation of dose-response relations of any ligand that binds to lipid membranes, especially when comparing the efficacy of different membrane binding ligands.