In this experiment, we examined whether self-observation, via video replay, coupled with the viewing of a skilled model was better for motor skill learning than the use of self-observation alone. Twenty-one female gymnasts participated in a within design experiment in which two gymnastics skills were learned. One skill was practiced in conjunction with the self-observation/skilled model pairing and the other with only self-observation. The experiment unfolded over five sessions in which pre-test, baseline, acquisition, retention, and post-test scores were obtained. Analysis of the physical performance scores revealed a significant Condition ×Session interaction in which it was shown that there were no differences between the intervention conditions at baseline and early in acquisition; but, later in acquisition, those skills practiced with the self-observation/skilled model pairing were executed significantly better than those with only self-observation. Also, an error identification test showed that participants had significantly higher response sensitivity scores for those skills learned with the paired intervention compared to self-observation alone. These results suggest that pairing self-observation with a skilled model is better in a gymnastic setting than self-observation alone.