There is some evidence that alternating physical and observational practice with a partner for the same skill can benefit learning compared to practice alone. What has not been studied is whether a partner's interleaved practice impacts multi-skill learning, when the partner either matches or mismatches their partner's skill. Here we manipulated partners' practice schedules of two golf putting skills. Partners practiced the same ("matched") or different skills in alternation ("mismatched"). Based on previous research where interleaved demonstrations have induced beneficial contextual interference effects, we hypothesized that mismatching a partner on consecutive trials should also promote a similar type of interference in practice, which ultimately aids learning. A third control group was tested, where only one partner practiced while the other observed. All groups practiced for two days, with individual retention tests at the start of day 2 and one week later. Taking turns practicing and observing a partner did not benefit learning compared to the control, pure physical practice group and the matched and mismatched groups did not differ in outcomes. There was, however, evidence that partners were adapting their actions (i.e., compensating for over or undershooting of the target) based on the shots of their partner, in a similar manner to how they were adapting to their own errors. Thus, although partners were influencing each other's performance, it was not ultimately to the benefit (or cost) of overall learning. Partner-mismatching of skills through alternating practice was not sufficient to promote interference in practice and ultimately promote learning.