While the use of virtual characters in medical education is becoming more and more commonplace, an understanding of the role they can play in empathetic communication skills training is still lacking. This paper presents a study aimed at building this understanding by determining if students can respond to a virtual patient's statement of concern with an empathetic response. A user study was conducted at the [blinded] College of Medicine in which early stage medical students interacted with virtual patients in one session and real humans trained to portray real patients (i.e., standardized patients) in a separate session about a week apart. During the interactions, the virtual and 'real' patients presented the students with empathetic opportunities which were later rated by outside observers. The results of pairwise comparisons indicate that empathetic responses made to virtual patients were rated as significantly more empathetic than responses made to standardized patients. Even though virtual patients may be perceived as artificial, the educational benefit of employing them for training medical students' empathetic communications skills is that virtual patients offer a low pressure interaction which allows students to reflect on their responses.