Abstract The use of video modeling to increase generalization of social play skills in children with autism is discussed. The possible reasons that have made this procedure so favorable among researchers and practitioners are explored. Two studies are described in which video modeling increased the generalization of social play in 6 children, and critical features of procedure are emphasized. Suggestions regarding the potential mechanisms responsible for the effectiveness of this procedure are discussed relative to basic behavioral theory and research. Key Words: Video modeling, generalization, social play, autism, children. Introduction The development of social skills by children with autism is qualitatively and quantitatively different from that observed in other childhood disorders or in typically developing peers as early as the second year of life (e.g., Ruble, 2001; Wetherby, Watt, Morgan, & Shumway, 2007). It has been suggested that social impairment may be a primary deficit that results in the development of secondary deficits such as communicative disorders or self-stimulatory behaviors (e.g., Baker, 2000; Bellini, Akullian, & Hopf, 2007; Rogers, 2000). For example, in verbal children, both the frequency of language use, even in an echolalic form, as well as the development of novel vocabulary have been demonstrated to increase along with increases in social engagement (e.g., Krantz & McClannahan, 1993; Kuhl, Coffey-Corina, Padden, & Dawson, 2005; Nikopoulos & Keenan, 2007; Stahmer, 1995; Thorp, Stahmer, & Schreibman, 1995). Furthermore, inappropriate behaviors have been shown to decrease during periods of active social engagement (Lee & Odom, 1996; Nikopoulos & Keenan, 2006). When children with autism have major difficulties in engaging in social interactions they may also have fewer opportunities to initiate social play (e.g., Beyer & Gammeltof, 2000; Koegel, Koegel, Frea, & Fredeen, 2001; Pierce & Schreibman, 1994) or to learn language (Kaiser, Hester, & McDuffie, 2001). Thus, social and communication skills are often studied as an integrated unit since impairments in the entire social communication domain have been acknowledged as a significant part of the symptom-based genetic research in autistic spectrum disorders (Klin, Jones, Schultz, Volkmar, & Cohen, 2002). Indeed, several studies have demonstrated that engagement in social communicative interactions directly affects other behaviors such as stereotyped or self-injurious behaviors even when these behaviors are not significantly targeted by the teaching program (Green, Gilchrist, Burton, & Cox, 2000; Horner, Carr, Strain, Todd, & Reed, 2002; McGee, Morrier, & Daly, 1999; Rogers, 2000; Summers, Houlding, & Reitzel, 2004). Since social skills are related to the long-term adjustment of and prognosis for both typically and atypically developing children, any program for children with autism must address and promote these valuable skills (Charlop-Christy & Daneshvar, 2003; Hwang & Hughes, 2000). The following sections of this paper will introduce the role of video modeling in promoting social development; describe the reasons for an increasing interest in video modeling among practitioners and researchers; summarize two studies that have examined the ability of video modeling to promote generalization of social skills; and examine elements of video modeling within a behavior analytic framework. Use of Video Modeling to Enhance Social Skills Many studies have been conducted to promote the social skills of children with autism by focusing on increasing their social communicative engagement with others (e.g., adults; typical peers) during play situations. Procedures have included social stories (Kuoch & Mirenda, 2003; Smith, 2001), peer-mediated intervention (DiSalvo & Oswald, 2002; McGrath, Bosch, Sullivan, & Fuqua, 2003), pivotal response intervention (Koegel, Koegel, Shoshan, & McNerney, 1999), self-management package (Newman, Reineche, & Meinberg, 2000), script-fading procedure (Krantz & McClannahan, 1993; Sarokoff, Taylor, & Poulson, 2001), and tactile prompts (Shabani et al. …