Developing a Virtual Coach for Chronic Patients: A User Study on the Impact of Similarity, Familiarity and Realism

A Two-Study Approach to Explore the Effect of User Characteristics on Users’ Perception and Evaluation of a Virtual Assistant’s Appearance

“…However, there is prior research that supports findings of Study 1, since Van Wissen et al [22] showed that participants preferred a realistic-looking agent over a cartoon-stylized agent to be their virtual nurse. The most reasonable explanation for these results seems to be the type of application, since Ring et al [5] and also van Wissen et al [22] showed that realistic agents are more appropriate for medical tasks. In line with this, Robertson et al [37] found that people even stated that they get angry about the use of cartoon-stylized agents in medical applications, since it somehow stultifies a very serious task.…”

“…Previous studies mostly showed that lower degrees of realism evoke a more positive perception of the agent [10,13,36]. However, there is prior research that supports findings of Study 1, since Van Wissen et al [22] showed that participants preferred a realistic-looking agent over a cartoon-stylized agent to be their virtual nurse. The most reasonable explanation for these results seems to be the type of application, since Ring et al [5] and also van Wissen et al [22] showed that realistic agents are more appropriate for medical tasks.…”

“…A realistic character seems to be more suitable for medical tasks, while the cartoon-like agents were preferred in social tasks [5]. In contrast to this, in the context of health applications, a more realistic-looking agent was evaluated more positively with regard to likability, competence, and trustworthiness compared to a stylized agent [22]. These findings might be caused by the specific application, since Ring and colleagues [5] showed that users prefer a realistic character for medical tasks.…”

“…In summary, prior research [5,11,13] showed that a cartoon stylization of a virtual agent evokes the perception of friendliness and that this is true for both the shade (or rendering style) of the agent and the proportions of the agent [5]. At the same time little evidence exists to demonstrate that for medical and health-related tasks a more realistic appearance might be preferred [22].…”

“…Prior studies mostly used humanoid agents [5,13,22], while Sträfling et al [11] showed that a cartoon-like rabbit evoked more friendliness. Therefore, the beneficial effect of a cartoon stylization seems to not necessarily be limited to humanoid agents.…”