When seeking a primary care physician, individuals often select physicians that shares their gender identity. We utilized a novel pre-registered experimental design to investigate whether these gender concordance preferences are due to homophily or deeper behavioral differences that systematically covary along gender. We collected 500 physician profiles from a health insurance website. Then, we edited these profiles such that each physician profile could be presented as either male or female, regardless of the physician’s true gender. We ran a 2 (presented male vs. presented female) x 2 (true male vs. true female) experiment and found that physician selection was predicted by patient-physician gender concordance along both presented gender and true gender, suggesting that both homophily and behavioral gender differences play a role. These effects are driven by female dyads such that female participants were more likely to select profiles that were presented as female as well as profiles that were written by female physicians, regardless of the presented gender. Additionally, we found that patient empowerment partially explained the effect of true gender on physician selection, suggesting that profiles written by female physicians were more likely to be selected by female participants because they discuss patient empowerment more than male physicians.
