We humans share with other mammals much of our evolutionary history and physiological makeup. Because of these shared physiological characteristics, scientists can use nonhuman mammals to test experimental therapies for various human disorders and diseases. These tests, which are called animal model studies, provide insights into whether a therapy might work in humans without risking human safety. The problem with animal model studies, however, is that in many cases the human disorder or disease of interest involves not only physiological dysfunction, but also psychological attributes that nonhuman animals may not experience (e.g., pleasure, pain, motivation, and addiction). Animal model researchers specify animal behaviors that they assume can serve as proxies for these psychological attributes. Yet these assumptions are substantive theories in their own right, which may be false, but which are taken for granted by the study designs. And if such an assumption is false, then the study results don’t mean what we think they mean and can’t be legitimately extended to humans.