Video modeling (VM) has been used as a tool to support skill development in individuals with autism since the late 1990s. While the medium to record and deliver the video model has rapidly advanced over the past 30 years, driven by drastic changes in technology, the defining characteristics behind VM have remained relatively unchanged. Video modeling is defined as the modeling of a target behavior in a recorded format that results in a video representation via an electronic medium (Ayres and Langone 2005; Bellini and Akullian 2007).

One of the seminal studies involving VM and individuals with autism spectrum disorder (ASD) was designed to compare VM to live modeling, and it was designed to teach a variety of skills (Charlop-Christy et al. 2000). Participants included five children with autism ranging in age from 7 to 11 years. Children with differing functioning levels (e.g., different mental ages, language ages, play skills) were purposefully selected to determine whether VM would be effective across severity levels. All of the children reportedly watched television for at least 30–60 min/day. Different target behaviors were chosen for each child depending on his or her need, as determined by assessments he or she received as part of his or her enrollment in an after-school behavior therapy program. Target behaviors included expressive labeling of emotions, independent play, spontaneous greetings, conversational speech, self-help skills, oral comprehension, cooperative play, and social play. Tasks were randomly assigned to the VM or in vivo condition and ranked by trained college students to be of similar levels of difficulty. Adults who were familiar to the participants provided the model in both the video and live conditions. It is important to note that no prompts or tangible rewards were presented to children during the VM condition, whereas in the in vivo condition, prompts to pay attention and social praise for correct responses were provided.

A multiple baseline design across participants was utilized. Results indicated that children acquired skills faster in the VM condition. Children also generalized target behaviors after VM, but did not generalize target behaviors after live modeling (Charlop-Christy et al. 2000). The total time children spent in the VM condition was 170 min, whereas the total time spent in the live modeling condition was 635 min. In other words, the VM condition was more time- and cost-effective than the in vivo condition. The researchers concluded that VM is an effective technique that can support skill development in children with ASD. Since that time, VM has been used to teach a variety of skills to children, adolescents, and adults with autism across multiple contexts.

A meta-analysis analyzing the effectiveness of VM as an intervention tool for children with autism was conducted on 23 studies published between 1987 and 2005 (Bellini an Akullian 2007). A total of 73 participants, ranging in age from 3 to 20 years, were included. The average number of VM sessions conducted was 9.5 with the average duration of each clip being reported as 3 min. Percentage of non-overlapping data points (PND) were analyzed across the three dependent variable categories and revealed that the highest intervention effects were found for functional skills, followed by social communication skills and then behavioral skills (Bellini and Akullian 2007). The researchers concluded that VM is an effective intervention strategy to teach skills to children with ASD, and skills are both maintained and generalized after treatment is concluded (Bellini and Akullian 2007).

Video modeling has been demonstrated to be an effective technique in schools and the community. Research has indicated that educators can apply VM in school settings. Bellini and Akullian (2007) reported that the majority of the studies they reviewed took place in school settings. In school settings, VM has been used to increase academic skills (e.g., Delano 2007; Kinney et al. 2003; Hitchcock et al. 2003), decrease disruptive behaviors (Apple et al. 2005; Buggey 2005), and increase social interactions (e.g., Cihak et al. 2009; Nikopoulos and Keenan 2003; Wert and Neisworth 2003). In general, the focus on community applications of VM has included increasing appropriate behaviors and effective transitions (e.g., Schreibman et al. 2000), but community applications have also focused on daily living skills by helping individuals with ASD choose items to purchase at a store (Alacantara 1994; Haring et al. 1995) and independently purchase items (Mechling et al. 2005).

Video modeling is an effective strategy to promote skill acquisition in children with ASD because they (a) are partial to visual stimuli (e.g., Kinney et al. 2003), (b) can focus more efficiently on restricted fields due to issues with overselectivity (i.e., attending to non-relevant stimuli; e.g., Corbett 2003), (c) often have exceptional memories and are skilled echoers (e.g., Charlop and Milstein 1989), and (d) appear to avoid face-to-face interactions (e.g., Charlop-Christy et al. 2000). In addition, VM has several practical advantages for use as an intervention tool, such as (a) the capacity to present an assortment of examples, (b) concise control over the modeling process, (c) exact repetition and reuse of video clips, and (d) cost and time efficiency of intervention (Corbett 2003).

As mentioned previously, VM has been used to teach a variety of skills to children with autism including imitation, play skills, self-help skills, and social skills (e.g., Ayres and Langone 2005; Cardon 2012; Cardon and Wilcox 2011; D’Ateno et al. 2003). It has been suggested that VM supports the development of observational learning, specifically the cognitive and behavioral changes humans experience as a result of watching or observing others involved in similar actions (Bandura 1977; Corbett 2003). According to Bandura, observational learning, or social learning theory, is critical to the development and survival of human beings as we learn what we should and should not do by observing events that occur around us. For observational learning to be successful, four components must be present: attention, retention, production, and motivation. We pay attention to activities that are modeled and then retain that information to utilize later. The actions that are observed can then be imitated (Bandura 1977). Although every action that is observed does not translate into perfect imitation, the awareness of the action has been created. Further, Bandura posited that the accuracy of imitative acts is reliant on positive reinforcement and continued input from others. The motivational factors that shape imitation in observational learning differentiate Bandura’s theory from Piaget’s in that Piaget attributes the desire to imitate to intrinsic needs as opposed to external factors.

A relatively new theory, proposed by researchers working at the Medical Investigation of Neurodevelopmental Disorders (MIND) Institute at UC Davis, implicates the visual attention differences present in children with autism (Vivanti et al. 2008). Researchers wanted to determine what children with autism look at when an imitative act is demonstrated. They believed that patterns in children’s visual attention could provide insight into how actions are encoded and ultimately which acts are imitated based on those patterns (Vivanti et al. 2008). The study included 18 children with autism and 13 children with typical development. It is important to note that all of the imitative acts presented to the participants were through prerecorded video clips. No live models were used. When controlling for the language level of the participants, researchers discovered that both groups more accurately imitated actions involving objects than those involving gestures. In addition, both groups visually attended to the action region (region where the action was performed) more during tasks that involved object imitation; however, differences were found between the two groups with regard to visual attention to the face region (area of the face of the demonstrator performing the action). Specifically, the children with autism looked at the face region less than half the time the typical children spent looking at the face region. The researchers proposed that the decreased attention to the face region could provide specific insight into the relationship between imitation deficits and social deficits present in children with autism. Future research into the visual attention skills of children with autism and the impact that visual attention has on imitation is necessary to determine the accuracy of the visual attention theory.

In summary, VM is thought to work as an intervention for children with autism because of several specific elements. When used as an intervention tool, VM capitalizes on the visual preferences (e.g., television watching, lining up toys to view them, repeatedly watching objects spin) exhibited in many children with autism (Corbett 2003; Kinney et al. 2003). Further, the screen offers a restricted field of vision and can therefore focus a child’s attention on relevant stimuli while decreasing their tendency to attend to irrelevant stimuli (e.g., Charlop-Christy et al. 2000; Corbett 2003). Screens are highly motivating, and reinforcement to attend to the task is built right in (Corbett 2003). And finally, children with autism attend for longer periods of time to screens as opposed to live presentations of information (Cardon and Azuma 2012; Vivanti et al. 2008).

As is evidenced by the numerous studies, VM is an effective intervention tool that can be used to teach play skills, language skills, self-help skills, social communication skills, functional daily living skills, academic skills, and appropriate behaviors. In addition, a number of reviews and meta-analyses have been conducted that identify the ability of children with ASD to maintain learned skills and to generalize those skills to new, previously undeveloped behaviors (Bellini and Akullian 2007; Ayres and Langone 2005; McCoy and Hermansen 2007; Gelbar et al. 2012). With regard to additional parameters of VM, however, there are several other components that should be considered.

Original research surrounding VM and autism looked at school-age and adolescent children to determine whether it was an effective intervention tool, particularly when targeting self-help skills or social skills (e.g., Charlop and Milstein 1989; Charlop-Christy et al. 2000; Nikopoulos and Keenan 2003). While seminal studies reported that older children with ASD responded positively to VM, more recent research has targeted younger children with ASD. In 2011, Cardon and Wilcox demonstrated that VM to teach object imitation was effective for children with ASD as young as 20 months of age. Further, very young children with ASD responded to VM used to teach object imitation, self-help skills, gestural imitation, and verbal imitation (Cardon 2012, 2013). Given the wide age ranges of children who respond to VM and the rapid response times that have been reported, introducing it as an early technique to support skill development is recommended.

In the home environment, emerging research has demonstrated that VM can be used by caregivers to teach a number of skills. Specifically, caregivers have been able to effectively create their own video models at home in order to teach fine motor skills (e.g., cutting with scissors and correct pencil grip), self-help skills (e.g., bed making), play skills (e.g., pretend play with a doll, puzzle completion), and social skills such as responding with, “No thank you” when offered an unpreferred food item (Cardon 2012). Similarly, VM has been utilized to teach children with autism how to interact with their siblings. Research conducted by Reagon et al. (2006) determined that children with ASD could be taught via VM to interact with their siblings during several different pretend play scenarios (e.g., cowboy dress-ups, playing teacher). It is recommended that VM be considered as a possible intervention technique across settings with a variety of individuals.