Advances in technology are expanding the role of robots in children's lives. In the near future, robots are expected to become learning companions that support early childhood education. However, how children perceive robots is not fully understood. In this talk, I will present findings from my experimental psychological research exploring whether children can learn from robots and how social interactions with robots influence children’s behavior. I will also overview how these findings can contribute to the design of social robots as learning companions and promote the advancement of early childhood education.