Video self-modeling has been found to be effective in increasing appropriate behaviors, increasing task fluency, and decreasing inappropriate behaviors. During video self-modeling, a student is filmed completing a task and then mistakes, prompts, and negative behaviors are edited from the video. When the edited video is viewed by the subject student, the student views a perfect model of him or herself successfully completing the given task. Video self-modeling has been used predominately with participants with autism spectrum disorder. This study is a replication of a previous study in which the effectiveness of video self-modeling and video peer modeling was compared (Sherer, Paredes, Kisacky, Ingersoll, & Schreiman, 2001). Sherer et al. evaluated these procedures with high functioning students with autism using a combined multiple baseline across participants and alternating treatment design. This study differs from Sherer et al.'s study in its use of participants who have multiple disabilities and low cognitive functioning. The results show that video self-modeling is effective for some participants while video peer modeling is effective for others. The individual student's preference for one form of video modeling over another form may indicate the method that is best for a particular participant. Implications for further research are included.