Communication barriers contribute low clinical trial participation, which is crucial in advancing breast cancer treatment. Patient-centered communication and considering user characteristics like health literacy may improve communication about clinical trials. Virtual agent interventions can employ these strategies to improve communication. However, it is unclear how the adoption of these strategies impact trust and use of a virtual agent intervention. Therefore, we conducted a study where 103 breast cancer patients interacted with a web-based, patient-centered virtual agent intervention to learn about clinical trials. Although the virtual agent’s communication strategy alone did not influence perceptions of trust, findings revealed that accounting for participants’ health literacy significantly impacted trust and unveiled a more nuanced understanding of patient-centered communication strategies. This work contributes to the design of virtual agent interventions, highlighting the necessity of incorporating user characteristics like health literacy when developing interventions for breast cancer patients for clinical trial engagement.

Figure 1: The single virtual agent (left) and multiple virtual agent (right) interfaces.

Control over an avatar in virtual reality can improve one's perceived sense of agency and embodiment towards their avatar. Yet, the relationship between control on agency and embodiment remains unclear. This work aims to investigate two main ideas: (1) the effectiveness of currently used metrics in measuring agency and embodiment and (2) the relationship between different levels of control on agency, embodiment, and cognitive performance. To do this, we conducted a between-participants user study with three conditions on agency ($\mathrm{n}=57$). Participants embodied an avatar with one of three types of control (i.e., Low - control over head only, Medium - control over head and torso, or High - control over head, torso, and arms) and completed a Stroop test. Our results indicate that the degree of control afforded to participants impacted their embodiment and cognitive performance but, as expected, could not be detected in the self-reported agency scores. Furthermore, our results elucidated further insights into the relationship between control and embodiment, suggesting potential uncanny valley-like effects. Future work should aim to refine agency measures to better capture the effect of differing levels of control and consider other methodologies to measure agency.

Redirected walking can effectively utilize a user’s physical space when traversing larger virtual environments by using virtual self-motion gains for a user’s physical motions. In particular, curvature gain presents unique advantages in redirection but can lead to suboptimal orientations. To prevent this and add additional utility in redirected walking, we formally present strafing gain. Strafing gain seeks to add incremental lateral movements to a user’s position causing the user to walk along a diagonal trajectory while maintaining the original orientation of the user. In a study with 27 participants, we tested 11 values to determine the detection thresholds of strafing gain. The study, which was modeled on prior detection threshold studies, found that strafing gain could successfully redirect participants to walk along a 5.57° diagonal to the right and a 4.68° diagonal to the left. Furthermore, a supplementary study with 10 participants was conducted, verifying that orientation was maintained throughout redirection and validating the obtained detection thresholds. We discuss the implications of these findings and potential ways of improving these quantities in real-world applications.

Virtual humans demonstrate the ability to act as non-judgmental conversational partners, eliciting greater self-disclosure. However, it is unclear what virtual human and conversational characteristics are important when self-disclosing. To address this gap, we conducted a set of qualitative, semi-formal interviews (n = 17) among computer science students to investigate participant mental models of willingness to disclose to virtual humans and characteristics of virtual humans that affect their self-disclosure. Our findings indicate that participants' mental models of virtual humans are largely inconsistent with current literature. This inconsistency appears to eliciting hesitancy and discomfort with virtual humans. Furthermore, trust and listening were identified as two primary characteristics of a virtual human interaction that are valuable towards willingness to disclose. Additionally, these characteristics were also valued in different ways for virtual humans in comparison to real humans. From the interviews, we identify and provide guidelines of designing virtual human interactions and conversations to elicit greater willingness to disclose.