PUB - Publikationen an der Universität Bielefeld

#### Purpose Since robots have been introduced in industry, their number in organizations has multiplied. Although appearing in various forms, most have one thing in common: Humans are interacting with them. The purpose of this study is to meta-analytically review research on the effectiveness of Human-Robot Interaction (HRI) at the workplace and give an overview on robot characteristics influencing HRI. #### Methodology For this, we conducted a systematic literature search in psychology and engineering databases (PsycINFO, Web of Science, IEEE Xplore, and ACM Digital Library) and found 2325 unique papers. We included 50 empirical studies that named the investigated technology “robot”, were conducted at a workplace or represent a workplace task, and report effect sizes for effectiveness of HRI (e.g. performance, satisfaction, workload, or trust). #### Results Preliminary 3-level meta-analytic results for the amount of assistance provided by the robot (based on a subgroup 41 effect sizes in 11 studies) show a medium effect on HRI effectiveness (d = .54). τ² values show a strong heterogeneity between effects sizes. Moderator analyses show that only effects on performance and satisfaction are significant (d = .65 and .72). #### Limitations Our meta-analytic effects could be biased because original studies often report statistics only for significant results. #### Implications and Value This study combines theory-driven psychological methodologies with empirical research from the field of robotics to approach the question “What makes a robot a good co-worker?” It is meant to foster interdisciplinary research on HRI and provide insight to researchers and robot designers on what makes HRI successful.