In this paper, we examine whether patient narratives alter the impact of stereotyping on choice of primary care clinicians: in this case, the common presumption that female doctors will be more attentive to empathic relationships with patients. 1052 individuals were selected from a nationally representative Internet panel to participate in a survey experiment. Participants were given performance data about 12 fictitious primary care physicians, including a randomized set of narrative feedback from patients. We compared the choice of clinician made by participants who value bedside manner and were exposed to narratives in the experiment, compared to those valuing bedside manner who had not had this exposure. We estimated multivariate logistic regressions to assess whether exposure to patient comments that "disrupt" stereotypes influenced choice of physicians. Participants who saw patient comments and had previously reported caring about bedside manner had a 67% higher odds of choosing a female physician than those participants that did not see a patient comments, controlling for the content of the narratives themselves. When participants were exposed to patient comments that disrupt gendered stereotypes, they had a 40% lower odds of choosing a female physician. Simple exposure to patient narratives that do not clearly disrupt gendered stereotypes increased the likelihood of choosing a female clinician by priming attention to relational aspects of care. However, when the content of a sufficient proportion of patient comments runs counter stereotypes, even a minority of narratives is sufficient to disrupt gendered-expectations and alter choices.