Video self-modeling (VSM) uses a method called "feedforward" to provide children the opportunity to view themselves as they perform in a more advanced or appropriate manner than they do presently. Typically, this is accomplished through the careful editing of videos. Studies on VSM and social skills with children on the autism spectrum ages 3 to 4 years have produced mixed results. In this study, a single-subject multiple-baseline design across four children (three on the autism spectrum and one with Down syndrome) with a mean age of 4 years 2 months was used to determine whether VSM would facilitate social initiations. This study was meant to be a replication of a study published in 2012. Positive changes were seen for all four children. The relationship between age and VSM efficacy is discussed along with other factors that may influence VSM outcomes with young children.