Observation of a model has been shown to facilitate motor learning, yet the optimal frequency of modeling combined with physical practice has not been well studied. Under an experimenter-controlled learning environment, an alternating schedule of one physical practice trial followed by one observational practice trial (100% frequency) was shown to be the most effective, with a 10% frequency providing no learning gains. When participants self-controlled the scheduling of observational practice, however, participants selected a 10% frequency schedule but still yielded the same observational learning benefits as those in the experimenter-imposed 100% group. Due to these conflicting results, the aim here was to explore whether higher self-controlled observation frequencies would generate greater learning. Forty-eight participants were tasked with learning the pirouette en dehors while assigned to one of four groups with differing constrained self-controlled observation frequencies: (1) 25%, (2) 50%, (3) 75%, or (4) no constraint imposed. Participants received 60 practice trials divided into four blocks of 15. Physical performance assessments were completed at pre-test, after acquisition blocks 1, 2, 3, and 4, and at a 24-hour post-test. Participants' performance increased throughout acquisition (p<.001) and they also received higher scores at post-test when compared to pre-test (p<.001). There were, however, no significant differences across the groups. As such, it can be concluded that similar motor learning outcomes were obtained under all of the self-controlled observation frequencies examined here.