System-wide trust (i.e., all teammates are trusted equally) is a phenomenon in teams containing non-human members, such as robots or artificial intelligence (AI). If trust is not calibrated according to team members’ individual abilities (i.e., component-specific trust), team performance may be negatively impacted due to under-trust in capable and/or over-trust in relatively incapable teammates. The goal of our study is to examine whether trust calibration and performance in human-robot teams can be improved via interventions that support the ability and willingness to see group members individually. We hypothesize that interventions increasing robots’ perceived human-likeness - by giving them typical human names and voices - will improve trust calibration (i.e., from system-wide to component-specific trust) and performance. Participants will work alongside three robots on a collaborative search task in a virtual reality environment. The robots’ perceived human-likeness will be manipulated between-participants. 52 participants will be collected based on a power analysis assuming a medium-high effect size. Effects of this manipulation will be measured via subjective rating scales (e.g., single-item-trust-rating), as well as objective measures, such as eye tracking (e.g., average count of fixations in the robot teammates’ search quadrants) and performance data (e.g., number of targets found by the team). The proposed project is unique in that it (1) is the first to investigate the construct of system-wide vs. component-specific trust in the human-robot teaming context and (2) uses objective methods with high internal validity to investigate trust calibration in human-robot teams embedded into realistic interactions with high external validity.