The Use of Self-Modeling to Promote Social Interactions Among Young Children

Abstract

Video self-modeling (VSM) has been used to teach social skills to children with autism older than 4 years of age. Attempts to use VSM with younger children with disabilities have produced less than favorable results; however, it is unclear whether VSM could be used to promote social initiations by typically developing children. Thirty minutes of staged filming, in which the four typically developing participants were prompted to interact with a peer with autism on the playground and inside during center time, was edited into 2.5- to 3-min clips. Each clip took less than 2 hr to edit. Data were collected on frequency of solitary play, initiations, parallel play, and engaged play and were then analyzed in a multiple-baseline-across-participants single-case design. Visual analysis led to the conclusion that VSM did not affect the typically developing children’s behavior. Limitations of the study and cautions for using VSM with very young children are discussed.

Keywords

autism video modeling social interaction young children

Social skill deficits comprise one of three main issues symptomatic of autism spectrum disorders (i.e., autistic disorder, Asperger’s disorder, pervasive developmental disorder–not otherwise specified) in Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR; American Psychiatric Association, 2000). Thus, there has been a concerted effort in research and practice to find effective methods to promote appropriate social behaviors and to extinguish maladaptive behaviors (Bellini, Peters, Benner, & Hopf, 2007). A good deal of this research has been conducted with preschoolers. This is at least partly due to findings that support the need for intervention as early as possible with this population (Woods & Wetherby, 2003). The intervention with the best research support with children with autism is intensive forms of applied behavior analysis (ABA) in which physical or verbal prompting, positive reinforcement (verbal or tangibles), and fading or shaping are used to train task-analyzed behaviors or to extinguish undesirable behaviors (Rosenwasser & Axelrod, 2001). The best known of these intense ABA strategies is the Lovaas method (Lovaas, 1987), in which a 40-hr-per-week regimen is recommended.

Over the last two decades, a slowly growing database of research has been accumulating regarding a less intense and less time-consuming method of instruction that appears to be successful for children on the autism spectrum: video modeling (e.g., Bellini & Akullian, 2007; Buggey, Hoomes, Sherberger, & Williams, 2011). There are three main forms of video self-modeling (VSM) that have been used with children with autism: peer, point-of view, and self. Peer modeling utilizes classmates or other children of approximately the same age who are filmed demonstrating a behavior (Ganz, Earles-Vollrath, & Cook, 2011; Nikopoulos & Keenan, 2003). Peer modeling is the most researched of the three methods and subsequently the terms peer modeling and video modeling are sometimes used interchangeably. There are several online sites offering video modeling services (Buggey, 2009); however, all use peers as models.

Point-of-view modeling is the newest form of video modeling, with the least research (e.g., Hine & Wolery, 2006; Schreibman, Whalen, & Stahmer, 2000). This method requires the camera to be held at what would be the child’s eye level as an adult or peer walks through the steps to a task. The resulting movie is intended to allow viewers to observe the task being completed as it would appear from their own eyes.

The other form of video modeling is VSM (Buggey, 2009; Dowrick & Raeburn, 1995). VSM typically requires the editing of raw footage to make it appear a child is performing advanced skills or appropriate behaviors. For example, individual words can be imitated by a child. When filmed and uploaded to video editing software, these words can be extracted and combined into short sentences. Children capable of role-playing can act out social situations that present challenges to them. If the resulting clip is all positive, it can be used in its entirety without editing. A feature of all forms of video modeling is that only positive images are used.

A shared advantage of all forms of video modeling relative to children with autism is that they use a medium that is a good fit with a population that tends to consist of visual learners (Bauman, 1999); however, VSM has the added advantage of providing children the opportunity to see themselves succeeding. Bandura’s (2001) research in social learning emphasized the relationship of self-efficacy to success. Seeing oneself succeed can be a powerful tool in building self-efficacy, if the behaviors depicted are developmentally appropriate. Bandura also found that the best models are those most similar to the viewer in all attributes, including performance. Having self as the model maximizes this aspect of modeling.

Virtually all of the research on self-modeling with young children has focused on children with disabilities, and a majority of these have concentrated on children with autism. At least 14 self-modeling studies have been carried out with children with autism (Gelbar, Anderson, McCarthy, & Buggey, 2012). All have reported positive outcomes except a study conducted by Buggey (2012), which included four children who were 3 years old. In another study using VSM with preschoolers with autism (Buggey et al., 2011), three 4-year-old children and one 3-year-old child participated. The 4-year-old children made good gains in social initiations, but the younger child did not. The only other study utilizing VSM in which no results were seen was conducted by Clark et al. (1993). These researchers worked with six children ranging in age from 3 to 5 years who were diagnosed with oppositional-defiant disorder.

Other researchers, such as Bellini, Akullian, and Hopf (2007) and Wert and Neisworth (2003), have reported positive outcomes with preschoolers between the ages of 4 and 5 years. The four children in the Buggey (2012) study were 3 years old, making them the youngest participants in any study on video modeling, although Litras, Moore, and Anderson (2010) reported success training social skills with a 3-year-old with autism. These mixed results with young children raise the possibility of a lower age limit at which point VSM, and possibly all forms of video modeling, are no longer effective. Developmental factors that constitute the prerequisite skills necessary for success with video modeling have received little attention in the research and the use of VSM with typically developing preschoolers has not been reported.

Peer-mediated interventions have been used effectively to promote social skills of preschoolers with autism (e.g., DiSalvo & Oswald, 2002; Strain, Kerr, & Ragland, 1979; Strain & Kohler, 1998) and have become one of the key elements of the Learning Experiences Alternative Program (LEAP; Strain, Kohler, & Goldstein, 1996). The typical peer-mediated type of intervention entails the training and use of typically developing peers to act as teaching assistants or tutors targeting the social and communication deficits of peers with autism. The training often entails adult (and potentially peer) modeling, instruction about the disability, and prompting aimed at shaping the peers’ behaviors. Although peer-mediated strategies with preschoolers have received considerable attention by researchers over the last 30 years, the extension of these methods into the early intervention years has not.

Considering the research done to this point, we know very little about whether peers can be trained to act as mediators or tutors for children with autism when the children to be trained are 3 years or younger. Likewise, we know very little about the ages or developmental perquisites for successful use of video modeling, specifically, self-modeling. Are the problematic results seen in children with autism below 4 years of age a result of age or disability-related issues? Thus, the research question in the present study was as follows:

Would typically developing children between 2 and 3 years of age increase their interactions (i.e., initiations, parallel play, and engaged play) with children with autism in free-play settings when they were shown videos of themselves modeling these interactions?

Method

Participants

Four typically developing children who were enrolled in a classroom serving children between 2 and 3 years of age were chosen for the study, along with two children diagnosed with autism. The classroom was being developed as a model early intervention site for children with autism and was part of a larger institute that served children birth to 5 years in a moderate-sized southeastern U.S. city. Most of the classrooms at the institute had a 50/50 ratio of children without and with disabilities; however, only two children with disabilities were served in this class, both with autism. Other classes were often on the playground at the same time, so the children were exposed to children with a range of disabilities. The four peers were chosen after observation determined no hesitation when prompted to hold hands with one of the children with autism (during transitions) and some degree of perseverance if the child with autism initially rejected the offer to hold hands (trying to recapture a withdrawn hand or asking the teacher for assistance). The children had been attending class together for 6 months and were familiar with one another.

Fifteen children were enrolled in this classroom, which had one special education teacher, two paraprofessionals with baccalaureate degrees, and a volunteer grandmother as the staff. Six of the children attended 5 days a week and the other nine attended between 2 and 4 days. Two of the four typically developing peers in the study attended full-time, whereas two attended part-time. All had well-developed language, motor, and social skills.

Children with autism

The two children with autism were similar in several ways. Both had pleasant dispositions, were almost always smiling, and spent most of their free time engaged in solitary play. The one exception to the solitary play was with the computer, at which time they would sit with each other or peers without prompting. They also were close developmentally as evidenced by the assessment information that follows. Neither displayed violent or self-injurious behavior nor did they exhibit noticeable repetitive behaviors other than running back and forth on the playground. The children did not respond to verbal initiations, and rarely complied with requests or demands.

On the playground, the two children with autism roamed on their own. When other children moved into proximity, the children with autism tended to move away. Again, there was one exception to this. There was a tire swing with room for three, and the children with autism would ride on this following physical or verbal prompts by classroom staff. The children expressed pleasure while on the swing as evidenced by smiling and laughing, but they would not make eye contact with others. Although the children with autism were not seen interacting with others, they did play on a variety of equipment, including slides, climbing apparatuses, and ramps, and in the sand area. The time at each area was relatively short, and they spent much of the playground time on the move, often circling to the outside of the play area. Both children were reported to have a minimum of expressive language, but no word or sentence utterances were heard during observations.

One of the children with autism, Larry, was 2 years 8 months old at the outset of the study. His score on the Childhood Autism Rating Scale (CARS) was 62, placing him in the severe range of autism. His result on the Vineland Adaptive Behavior Scales (VABS) was at the first percentile, with specific weaknesses in socialization and communication. The second child, Jerry, was 2 years 6 months old. He scored in the second percentile of the VABS and in the first percentile of the Mullen Scales for Early Learning. His CARS score was 58, which fell into the severe autism range. Both children had been assessed by trained staff (who held master’s degrees in education) from the Tennessee Early Intervention system, with input from the parents to complete the CARS. A diagnosis of autistic disorder was given by their pediatricians.

There were no planned activities specifically targeting social interaction between children with and without disabilities, although teachers would occasionally prompt or physically direct children with autism to the sand or swing area on the playground or to specific centers within the classroom. Prompts also were given to children without disabilities to interact with the children with autism. This was especially true during transitions when children would partner and hold hands.

Typically developing participants

Mattie was 3 years old at the outset of the study and attended the class Monday through Thursday. She was very verbal and social and played with all children in the classroom except the two with autism. She did not hesitate when asked to assist one of the children with autism with a task or to hold hands as partners during transitions. (None of the children in the class showed any aversion to the children with autism except on rare occasions, such as a collision or obstruction on the part of the children with autism: a normal reaction to any child in that situation.) Mattie often played in a group of two to four other classmates on the playground. This also held true in the classroom during center time when children could chose their activities.

Jon was 2 years 11 months old at the beginning of the study. He was very active physically, doing much climbing and running on the playground. He had a “best friend” in the class, but interacted with all of the other children.

Kathi was 2 years 8 months old when the study began. Compared with the other children, she was a little more reserved in her socializations and talked very quietly. On the playground she was more likely to play alone and sometimes did not participate with the other girls when they were in a group, although this was not always the case.

Lily was 2 years 6 months old. She interacted well with the other girls, although she was somewhat shy with strangers. She did not respond to researchers or staff when asked to verbalize during the making of the video used in intervention, although she responded well to all other requests during filming regarding interactions with the child with autism.

Settings

Observation and filming were carried out in two locations, playground and classroom. The classroom was a large, well-lit room with large windows, and it contained wheeled, 2 foot, movable storage shelves that were used to divide the room into centers. The centers included blocks, computer, literacy, dress-up, and science (including animals). Occasionally a water-play area was included. There was a large circular rug that was used for large-group circle time. The lead teacher had an office with desk and computer and this is where videos were shown to children. There also were two bathrooms/change areas.

The playground area was approximately 10,000 square feet with very modern equipment being beta tested at the site. Three large structures with elements for climbing, sliding, and manipulation served as the centerpieces, with swing and sand areas adjoining. A sidewalk circled through the playground beginning and ending at the exit door from the learning center. A variety of riding equipment was available for use.

Children spent 20 to 30 min on the playground midmorning. This was followed by a transition back to the classroom where the children would have a bathroom break and then would go to a short circle time activity. Typically, an adult sat with each child with autism during circle time, but the two children were allowed to roam if they were not attending. Children would then be permitted to attend a center of choice, with the adults overseeing one or two special activities like water-play or painting. The children with autism often were redirected to these activities and would sometimes participate, especially with drawing or painting. Otherwise, they would move about the room, rarely spending more than a few minutes playing with or manipulating material before moving on.

Design

The present research was conducted using a single-case, multiple-baseline design across four students. Data were analyzed using visual inspection of graphs to determine the presence of a functional relation between social initiation and VSM.

Dependent Variables

For the purpose of this study, three forms of social interaction were targeted as the dependent variables: (a) initiation, (b) parallel play, and (c) engaged play. Social initiation was defined in terms of purposeful physical approaches and vocalizations. Physical approach was defined as moving into proximity of the child with autism such that it resulted in some form of engagement, such as eye contact for more than 5 s or appropriate touching (pushing and hitting were not counted). Parallel play was defined as participating in the same activity in close proximity to the child with autism for a minimum of 30 s without verbalization or exchanges of play equipment. Engaged play was the same as parallel play except that more than one verbalization or cooperative exchange of materials had to occur during the play activity. Computer activity was considered separately. Most of the time, the typically developing peers and children with autism were focused on the screen, often not even glancing at whoever was next to them. Even though moving into proximity and a form of parallel play may have been occurring, it was decided to exclude computer time from the data analyses. It was rarely clear whether there was engagement beyond the computer screen.

Observation and data collection

Two observers were used in this study: the primary author, who is a university professor, and the second author, who was a master’s degree student in the psychology department at a university. Observers carried clipboards and stopwatches and moved into proximity of the child being observed so that verbalizations could be heard. Both observers had been present on the playground almost daily during the 4 months prior to the study, so the children were used to the observers’ presence. When two observers were watching the same child on the playground for interobserver agreement (IOA) purposes, they positioned themselves opposite of each other to the point allowed by the playground equipment. Classroom observations were attempted through an observation room window; however, the children often disappeared behind dividers and could not be seen when they were low to the floor. Thus, the observers moved into the classroom and had to position themselves so that they could see the children with autism. They moved if the child went out of view.

Data were collected using event recording. The observers conducted daily 15-min observations during playground and center times. Because any interaction initiated by the typically developing peers would require the participation of one or both of the children with autism, the latter children became the target of observation. In this way, all interactions with the children with autism could be recorded, including those of the four typically developing peers who were watching videos. An observer focused on one of the children on the playground and then switched to the other child in the classroom during center time. The observers rotated which child they would view first so that each child was observed in both locations an equivalent number of times by each researcher. Thus, each child had two data points per day. This number was reduced to one when IOA was being measured and both observers focused on one child outside and on the other inside. This, too, was rotated so that children were observed an equal number of times inside and outside.

A data collection form was created that allowed observers to record the type of event (physical or verbal initiation, parallel play, engaged play, teacher-prompted event), time of occurrence, and context of the interaction. The most important aspect of the context for this study was who was involved; however, antecedents and location also were recorded. The form had sections for the focus child and peers based on who initiated. This was done so that it would be easier to extract data to determine whether the children with autism began initiating at some point during the study and with whom. In addition, there was a category of solitary play defined as activity with play material or engagement in physical activity without peer involvement. This was added to determine if time engaged in solitary play decreased should activity with peers increase. The data collection form is shown in Figure 1.

IOA

A 90% agreement rate across observers was set as the criterion for acceptable IOA. The observers had observed other children in the same outdoor environment for 6 months and had tracked similar behaviors using a similar form. They attained a 94% IOA over that time. In addition, for the present study, observations on the playground and in the classroom were conducted for 1 month prior to beginning the baseline phase of the study. Time was spent observing typical peers as well as the children with autism because observations of the latter often resulted in no occurrences. It was found that there was excellent agreement on event occurrence; however, there was some disagreement with categorization of behaviors. Thus, this pre-baseline period involved much discussion about how to classify borderline behaviors when there was disagreement. Baseline data collection began once criterion for IOA was reached. An additional 25% of observations were dedicated to IOA throughout the study. Identical data collection procedures were used in the baseline, intervention, and maintenance phases of the study. Agreement of 99% occurred during the study, with only one disagreement in which an observer failed to see an interaction recorded by the other.

Independent Variable

The independent variable in this study was a 2.5- to 3.0-min video featuring each participant socially interacting with his or her peers with autism. This included footage of the child initiating interactions and playing together with the children with autism.

Footage of playing together was obtained by keeping the typically developing peer and one of the children with autism in the classroom (when vacated) or on the playground after the class had gone in. At that time, the typically developing peer was prompted to interact with the child with autism. The researchers followed the lead of the child with autism as he moved among centers or among equipment on the playground. The typically developing children were prompted to call out to the other child, to help with an activity such as a puzzle, to slide simultaneously on a double slide, to bring him material, and so forth. Kathi’s voice was not loud enough to be picked up clearly by the camcorder, and Lily refused to verbalize; however, both of these children followed directions to physically initiate by tapping on the shoulder, doing the “come here” gesture with the index finger, and taking a hand and leading.

The footage collected was then edited to make it seem that interactions were seamless and occurred frequently. At least one example of the targeted behaviors was included in all videos. Anything less than an exemplar of the target behaviors was removed. For example, a short clip showing the typically developing child bringing a puzzle and handing it to the child with autism was combined with a later clip showing the two working together on the puzzle (the peer having been prompted to help). Thus, the combined clips would illustrate a physical initiation and engaged play. Any initiation—whether verbal (e.g., “Play with me”) or gestural—was combined with a clip of engaged play to illustrate the cause–effect relationship even though this never occurred in reality. The edited clips became the body of the video, and two other short segments were added. The first was an introduction in which a narration labeling the behavior (“Here’s Kathi playing nicely with Larry”) was overlaid on a 5-s freeze frame of the two children playing. At the end of the video, another illustrative and flattering freeze frame was added, along with a mild reinforcement and labeling of behavior (“Nice job playing with Larry,” followed by the sound of clapping). These additions were used to try to ensure that the children focused on the salient feature of the video.

The videos were created on a MacBook Pro using iMovieHD software. Footage was recorded on a Sony MiniDVD DCR-TRV22 Handycam and then downloaded to the computer. The software automatically broke the footage into shorter clips. These clips could be scanned quickly by dragging and clicking a cursor, and scenes of the children in close proximity were examined for appropriateness. Unwanted footage was deleted, leaving the best examples of the behaviors, which were then combined for the final movie. The movies were saved in Quicktime format and put on a DVD. The entire process took less than 2 hr per video. All of the videos were very similar in content and format, although activities on each varied based on children’s play preferences. Kathi and Jon observed videos featuring Larry while Mattie and Lily were shown videos of play interactions with Jerry.

Two weeks after baseline data were collected for all four typically developing children, the first video was made for Mattie. Filming occurred on successive days, one session in the classroom and the other on the playground. Each filming event lasted about 15 min. The 30 min of video was edited to show 2.5 to 3 min of apparent interactions across both the playground and classroom settings. Three additional days of baseline data were collected to ensure that the filming had not affected behaviors. Children were shown the videos daily over a 1-week period. Thus, the two children who attended every day watched the video five times. Kathi, who attended 4 days, watched four times, as did Lily, although because she only attended on Monday, Wednesday, and Friday, the video was viewed on the subsequent Monday as well. The minimum of four viewings was deemed sufficient because it has been reported that changes tend to occur soon after children see their videos (Buggey, 2005; Dowrick & Raeburn, 1995). In addition, because a treatment phase should be extended until changes in behavior are seen (Kennedy, 2005), it was decided that 4 days would be sufficient time to determine whether change was occurring. This basic procedure for filming and viewing was continued across all children. Once Mattie had completed the 5 days of viewing, her video was withdrawn and Jon began watching his. The viewing and withdrawal of videos continued in this manner with Kathi and Lily.

Children viewed the video with the teacher in her small office just off of the classroom. The door was left open so that at least one of the researchers could watch the viewing. This was important not only for treatment fidelity but also because it allowed the researcher to record the child’s reaction to the video. Children this young had not participated in a VSM study, and the researchers wanted to see how well the children attended to the movie. All of the children appeared to attend well, although after Mattie watched her video for the first time she made the comment “nice flowers.” Upon looking at the computer screen, one could see that the background wallpaper was flowers and was visible outside the edges of the movie viewer. Thus, the direction of the children’s head could be used to gauge attention, but it cannot be stated for certain that attention was actually on the video. Jon was distractible during his first viewing, looking at the objects in the teacher’s office, where children were not usually allowed. Thereafter, he seemed to attend well, as did the other two children. The teacher was consistent with the showing of the video and followed instruction to make no comments to the child other than responding succinctly to queries.

The viewing occurred between playground and center times during a transition in which children washed and took care of bathroom needs. There is evidence to suggest that using video priming, in which the video model is shown just before the desired behavior is to occur, can be effective (e.g., Schreibman et al., 2000). However, positive results with VSM have occurred when the model and desired behavior are not close temporally. Thus, viewing of videos occurs when it causes the least distractions to the child or at a time of convenience to the teacher (Buggey, 2009). Children in schools are often shown their videos when they arrive and before all children are present (e.g., Bellini, Akullian, et al., 2007; Buggey, 2005). Thus, the children in this study were shown the video when they were not engaged in a specific activity and were taking turns with bathroom needs. The timing offered limited disruption and was convenient.

The study was approved by the Institutional Review Board at the authors’ university. Because filming was involved, great care was taken to ensure that parents were well informed of the self-modeling procedure and the disposition of the videos once the study ended. Parents of the typically developing peers were given the choice of having all footage destroyed, receiving a copy of their child’s video while destroying all other footage, or allowing educational use of the video for training purposes. The same options were presented to the parents of the children with autism. All participating families knew in advance that if any parent wanted all traces of video destroyed, this choice would take precedent.

Treatment Fidelity

A review of journal entries and discussion between the two observers revealed that the teacher at no time showed the video in a manner that fell outside of her instruction. It was shown at the same time each day and, with the exception of saying “good job” and occasionally verbally redirecting the child to the screen (all occurring during the first viewing by the child), without comment.

Results

Frequency of play interactions among the four typically developing children and two children with autism did not change over the course of the study. The results per data collection session are shown on the graph in Figure 2. The graph represents cumulative data in which the number of social initiations, parallel play, and engaged play (which never occurred) were combined. Jon had two occurrences during maintenance that were higher than any during his baseline or intervention; however, this change was very unlikely to have been caused by the intervention. Because all interactions with the children with autism were recorded, it was possible to determine whether there were changes in behavior of any other peers. During the study, only five other interactions occurred, initiated by three different peers. Three of these occurred during baseline, two during intervention phases, and one when all children were in maintenance.

There did not appear to be any relationship relative to who appeared in the video with the typically developing children. Mattie’s two interactions were with the same child; however, Jon had three interactions with each child during maintenance. The three interactions between Jon and Larry on the same day in Jon’s maintenance data involved a minor conflict with possession of a ball. These occurred indoors. The approaches by Jon did not involve negative behaviors, so they were recorded as initiations.

The location, indoors or outdoors, did not seem to be an influencing factor. Jon interacted four times in the classroom and twice outdoors. Mattie had one interaction in each setting and Lilly’s pattern was three inside and two outside. The three inside were during painting activities (that had not occurred during baseline), where they were next to each other. The videos were shown just after playground time and before inside activities. The effects seen during classroom time did not tend to indicate that a priming effect was occurring by showing the video prior to inside activities.

Social Validity

The teacher and two paraprofessionals were asked their opinions regarding the VSM procedure, its application in the classrooms, its effects on the students, and the feasibility of using VSM in the future. Information was procured through informal interviews beginning 1 week into the first intervention phase and continuing once every other week for the remainder of the study. More formal assessment of social validity was not undertaken because the researchers did not want the teachers to associate extra paperwork with the method and it was felt that more accurate information could be gained from conversations.

The teacher and paraprofessionals were universally supportive of the method. The teacher was working on her master’s degree and expressed a desire to use VSM for her thesis, focusing on simpler, life-skills behaviors. There were questions about the feasibility of filming during classroom time, and it was suggested that support staff could do the filming based on behaviors that they selected.

Discussion

Because of the problems seen in the literature associated with VSM use with children with disabilities under 4 years of age (Buggey, 2012; Clark et al., 1993), this study was undertaken to see whether typically developing children within this age group would respond any better. It was hoped that insight into the developmental skills necessary for successful video modeling could be ascertained. It is logical to assume that there is an age threshold at which self-modeling becomes effective that is probably related to social and cognitive development and the type of behavior that is targeted. Self-recognition typically develops before 2 years of age; however, there is evidence that self-awareness and the ability to recognize that play is not always real occurs between 2 and 3 years of age (Lewis & Ramsay, 2004). These and other developmental skills may be prerequisites for successful video modeling.

The lack of results here adds to the small, but growing, database of self-modeling use with children below 4 years of age. This constitutes the first examination of VSM with children below 3 years of age. At first glance, it would appear that there is an age threshold for self-modeling that occurs at 4 years of age; however, the studies without participant gains have all involved social skills. It is possible that results are behavior specific and it would be premature to generalize the findings beyond the type of social skills addressed in these studies. This is certainly an area in need of further study, but it is safe to say that self-modeling use with children below 4 years of age in classrooms, homes, and clinics needs to proceed with caution, and parents need to be informed of possible limitations of VSM.

This study also represents the first time that typically developing children served as participants and thus may indicate that there indeed are developmental criteria that must be mastered before self-modeling can be used for social skills training. The children with autism in our study exhibited no physical indications of their disability, and the behavior manifestations were limited to social isolation and lack of verbalization. Any approaches carried out by typically developing children when the children first met were probably met with the same avoidance (averting eye gaze, turning around, or moving away) as witnessed during the study. This probably was interpreted as rejection that may have served to extinguish or inhibit further initiations. There is a danger that if very young children learn to avoid peers with autism because of initiations not being reinforced or even acknowledged, the avoidance will continue in later years. This problem might be mitigated by good adult modeling by preschool staff and therapists, but would probably still remain a challenge. More needs to be learned about development of skills needed for empathy, consideration, tolerance, and an ability to understand an “invisible” type of disability. Understanding how and when these concepts develop will allow professionals to identify when to introduce peer-based interventions, including forms of video modeling. It also would behoove researchers to investigate very young children’s attention behaviors, particularly the ability to recognize salient features of a video.

It is hoped that researchers will continue to study the “frontiers” of self-modeling and not be dissuaded by the possibility of seeing no changes in behavior. Finding that a treatment is not effective may disappoint those persons associated with study participants, but it may provide more valuable information for practitioners in the long term.

Limitations

Single-case designs have the inherent problem of low sample sizes, even though three children are deemed adequate for establishing control (Kennedy, 2005). In addition, the targeted dependent variables are typically narrow in focus; thus, it is often difficult to make generalizations across behaviors. In terms of the present study, the only claim that can be made is that four children who were less than 3 years old did not increase their play interactions with children with autism as a result of watching VSM videos.

There was a threat of contamination in this study. Although students watched their tapes in private, the filming was often done with the knowledge of other students. Moreover, the presence of the researchers in proximity during all observation sessions may have served as a prompt for participant and peer to engage in social behaviors—or more likely, given the results, as a deterrent. Although there was much attention given to desensitizing the students to the researchers’ presence, this cannot be eliminated as a possible confounding factor.

The density of children on the playground may have affected results. At various times, there could be from one to three classes on the playground. This meant that between 12 and 36 children could be present. The opportunities for interactions would increase as the population grew and likewise the opportunity to avoid interactions decreased or, at least, the opportunity to avoid proximity to others decreased. However, there was no evidence that this was the case.

Although older children have shown changes in behavior after only one or two viewings, it is possible that four or five viewings by 2- to 3-year-old children are insufficient. Bellini, Akullian, et al. (2007) had their children, ages 4 and 5 years, watch videos on 17 occasions over 4 weeks. On the other hand, Buggey et al. (2011) used only five viewings over a 1-week span to obtain positive results with 4-year-old children. Again, we are in need of further research to determine whether more viewings would provide better results.

Conclusion

The findings of the present study suggest that using VSM with peers to stimulate social initiations with children with autism is problematic when the children are less than 3 years of age. There are insufficient studies using VSM with this age group to make strong claims about efficacy; however, this is the third such study targeting social behaviors with preschoolers where no improvements were noticed. Children with autism (Buggey, 2012) and oppositional-defiant disorder (Clark et al., 1993) participated in previous studies. Adding four typically developing children to this list indicates that there may be developmental factors that limit VSM’s efficacy. As it stands, children who are above 4 years of age seem to benefit from VSM intervention (e.g., Bellini, Peters, et al., 2007; Buggey, 2005; Wert & Neisworth, 2003), whereas children below 4 years of age have not (Buggey, 2012; Clark et al., 1993). More research is needed to confirm or refute these findings. In addition, researchers using VSM with this age group need to address other behaviors, such as language, life skills (e.g., toileting, dressing), and physical skills. Firmly establishing lower age limits of VSM efficacy across behaviors will be of great value to professionals and parents.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.

Bandura

(2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52, 1–26.

Bauman

M. L.

(1999). Autism: Clinical features and neurobiological observations. In Tager-Flusberg

(Ed.), Neurodevelopmental disorders (pp. 383–399). Cambridge, MA: MIT Press.

Bellini

Akullian

(2007). A meta-analysis of video modeling and video self-modeling interventions for children and adolescents with autism spectrum disorders. Exceptional Children, 73, 261–284.

Bellini

Akullian

Hopf

(2007). Increasing social engagement in young children with autism spectrum disorders using video self-modeling. School Psychology Review, 36, 80–90.

Bellini

Peters

J. K.

Benner

Hopf

(2007). A meta-analysis of school-based social skills interventions for children with autism spectrum disorders. Remedial and Special Education, 28, 153–162.

Buggey

(2005). Applications of video self-modeling with children with autism in a small private school. Focus on Autism and Other Developmental Disabilities, 20, 180–204.

Buggey

(2009). Seeing is believing: Self-modeling applications with children with autism and other developmental disabilities. Baltimore, MD: Woodbine House.

Buggey

(2012). Effectiveness of video self-modeling to promote social initiations by 3-year-olds with autism spectrum disorders. Focus on Autism and Other Developmental Disabilities, 27, 102–110.

10.

Buggey

Hoomes

Sherberger

M. E.

Williams

(2011). Facilitating social initiations of preschoolers with autism spectrum disorders using video self-modeling. Focus on Autism and Other Developmental Disabilities, 26, 25–36.

11.

Clark

Beck

Sloane

Goldsmith

Jenson

Bowen

Kehle

(1993). Self-modeling with preschoolers: Is it different? School Psychology International, 14, 83–89.

12.

DiSalvo

C. A.

Oswald

D. P.

(2002). Peer-mediated interventions to increase the social interaction of children with autism: Consideration of peer expectancies. Focus on Autism and Other Developmental Disabilities, 17, 198–207.

13.

Dowrick

P. W.

Raeburn

J. M.

(1995). Self-modeling: Rapid skill training for children with physical disabilities. Journal of Developmental and Physical Disabilities, 7, 25–36.

14.

Ganz

J. B.

Earles-Vollrath

T. L.

Cook

K. E.

(2011). Video modeling: A visually based intervention for children with autism spectrum disorder. Teaching Exceptional Children, 43, 8–19.

15.

Gelbar

N. W.

Anderson

McCarthy

Buggey

(2012). Video self-modeling as an intervention strategy for individuals with autism spectrum disorders. Psychology in the Schools, 49, 15–22.

16.

Hine

J. F.

Wolery

(2006). Using point-of-view video modelling to teach play to preschoolers with autism. Topics in Early Childhood Special Education, 26, 83–93.

17.

Kennedy

C. H.

(2005). Single-case designs for educational research. Boston, MA: Allyn & Bacon.

18.

Lewis

Ramsay

(2004). Development of self-recognition, personal pronoun use, and pretend play during the second year. Child Development, 75, 1821–1831.

19.

Litras

Moore

D. W.

Anderson

(2010). Using video self-modeled Social Stories to teach social skills to a young child with autism. Autism Research and Treatment. Advance online publication. doi:10.1155/2010/834979

20.

Lovaas

O. I.

(1987). Behavioral treatment and normal intellectual and educational functioning in autistic children. Journal of Consulting and Clinical Psychology, 55, 3–9.

21.

Nikopoulos

C. K.

Keenan

(2003). Promoting social initiation in children with autism. Behavioral Interventions, 18, 87–108.

22.

Rosenwasser

Axelrod

(2001). The contributions of applied behavior analysis to the education of people with autism. Behavior Modification, 25, 671–677.

23.

Schreibman

Whalen

Stahmer

A. C.

(2000). The use of video priming to reduce disruptive transition behavior in children with autism. Journal of Positive Behavior Interventions, 2, 3–11.

24.

Strain

P. S.

Kerr

M. M.

Ragland

E. U.

(1979). Effects of peer-mediated social initiations and prompting/reinforcement procedures on the social behavior of autistic children. Journal of Autism and Developmental Disorders, 9, 41–54.

25.

Strain

P. S.

Kohler

F. W.

(1998). Peer-mediated social intervention for young children with autism. Seminars in Speech and Language, 19, 391–405.

26.

Strain

P. S.

Kohler

F. W.

Goldstein

(1996). Learning experiences an alternative program: Peer mediated interventions for young children with autism. In Hibbs

Jensen

(Eds.), Psychosocial treatments for child and adolescent disorders: Empirically based strategies for clinical practice (pp. 573–586). Washington, DC: American Psychological Association.

27.

Wert

B. W.

Neisworth

J. T.

(2003). Effects of video self-modeling on spontaneous requesting in children with autism. Journal of Positive Behavior Interventions, 5, 300–305.

28.

Woods

J. J.

Wetherby

A. M.

(2003). Early identification of and intervention for infants and toddlers who are at risk for autism spectrum disorder. Language, Speech, and Hearing Services in Schools, 34, 180–193.