Video modeling has been used in the field of Applied Behavior Analysis (ABA) to teach skills such as conversational speech (Charlop & Milstein, 1989), perspective taking (Charlop-Christy & Daneshvar, 2003), and complex play sequences (D'Ateno et al, 2003). Some research evaluating the models has been conducted to include an investigation of effectiveness when siblings as the identified model (Taylor et al, 1999).
In a video modeling procedure, the learner watches the video (often looped 2 or 3 times), then is given the materials to play/work with. Using a carefully created task analysis the teacher would score the steps the individual completed correctly and which ones they completed incorrectly. As a general rule, if the learner has steps 1, 2, and 3, begin teaching step 4. If a learner displays more splinter skills such as steps 1, 2, 5, 7, and 10, then begin by teaching the total task. Deciding whether or not to include language in the video depends on the skill you are targeting and the skills the learner has. For social skills, and commenting during play, many times my teaching videos do include comments.
Learn more (read the relevant research) about VIDEO MODELING.