Citing their students’ low levels of empathy, medical educators have scrambled to implement curricula with the hopes of buffering against the corrosive effects of biomedical and clinical experiences in medical school. The assumption undergirding these studies by social scientists and medical educators alike is that immersion in biomedical education and clinical experience erodes students’ empathic capacities, and that exposure to humanities and social sciences content will amend these losses. But we do not know if this assumption is correct. In this project, we empirically assess this assumption by utilizing a unique data set constructed from student applicant and survey data from the American Medical College Application Service (AMCAS) and the Association of American Medical Colleges (AAMC). We test whether medical school students ( N = 8255) from the United States (U.S.) with different academic backgrounds represented by their college major have different levels of empathy, net of demographic control variables. We report two findings. First, we find that students who majored in humanities or interpretive social sciences disciplines have higher empathy scores than their peers who majored in the positivistic social sciences and STEM (science, technology, engineering, and mathematics) disciplines. Second, we find that the relationship between empathy and time in medical school is more nuanced than we would expect from the existing literature.