Mosston and Ashworth's (2008) reciprocal style of teaching gives learners the opportunity to work in pairs to support each other's learning (one practices a task and the other gives feedback). The effects of pairing learners by companionship (friend and nonacquaintance) on 8-year-old children's motor skill performance and comfort levels were explored in this study. The participants (N = 52) were randomly assigned to a treatment group (n = 40) or a control group (n = 12). Prior to the study, learners in the treatment group were paired by companionship (partners who were friends and partners who were not friends). The same dribbling tasks were taught to the learners in all groups (eight 30-min sessions). A soccer dribbling test and a 7-point semantic differential scale were employed to evaluate the dribbling skill and how comfortable the learners felt giving and receiving feedback, respectively. The results show that learners paired with friends felt more comfortable in giving and receiving feedback than learners paired with nonacquaintances. Also, motor skill development was greater in learners paired with friends than learners paired with nonfriends or learners in the control group. The study supports certain tenets set forth by Mosston and Ashworth (2008) for the reciprocal style.