USING VIDEO MODELING AND SOCIAL CONSEQUENCES TO INCREASE THE UNSCRIPTED PLAY BEHAVIOR IN CHILDREN WITH AUTISM A Thesis Presented to the faculty of the Department of Psychology California State University, Sacramento Submitted in partial satisfaction of the requirements for the degree of MASTER OF ARTS in Psychology (Applied Behavior Analysis) by Krisann Elizabeth Vistica SPRING 2012 USING VIDEO MODELING AND SOCIAL CONSEQUENCES TO INCREASE THE UNSCRIPTED PLAY BEHAVIOR IN CHILDREN WITH AUTISM A Thesis by Krisann Elizabeth Vistica Approved by: __________________________________, Committee Chair Caio F. Miguel, Ph.D __________________________________, Second Reader Becky Penrod , Ph.D _____________________________________, Third Reader Jill Young, Ph.D ____________________________ Date ii Student: Krisann Elizabeth Vistica I certify that this student has met the requirements for format contained in the University format manual, and that this thesis is suitable for shelving in the Library and credit is to be awarded for the thesis. __________________________, Graduate Coordinator___________________ Jianjian Qin, Ph.D Date Department of Psychology iii Abstract of USING VIDEO MODELING AND SOCIAL CONSEQUENCES TO INCREASE THE UNSCRIPTED PLAY BEHAVIOR IN CHILDREN WITH AUTISM by Krisann Elizabeth Vistica Video modeling has been used to teach a variety of skills to children with autism including self-help, perspective taking, and play skills. Previous research suggests that video modeling is an effective tool to teach appropriate play and social skills to this population. The purpose of this study is to increase the unscripted or “spontaneous” play behavior and social interaction of children with autism. A multiple-probe across playscripts design was used to examine the effects of a video modeling procedure. Participants viewed videos of two play scenarios and were then paired with a play partner to interact with. Results indicate that both children acquired the scripted behavior which increased interactions while their unscripted behavior significantly decreased. This suggests the videos may have suppressed the unscripted behavior of the boys. Implications for increasing unscripted and interactive play behavior are discussed as they relate to the social skill development in children with autism. _______________________, Committee Chair Caio F. Miguel, Ph.D _______________________ Date iv ACKNOWLEDGMENTS The author would like to thank her thesis committee, Dr. Caio Miguel, Dr. Becky Penrod, and Dr. Jill Young for their efforts in the development and fulfillment of this research project. Specifically, the author would like to thank Dr. Caio Miguel for his ongoing support and dedication to the completion of this study. The experience I have gained in working with Dr. Miguel has been invaluable to my academic and professional development. In addition, I would like to express my appreciation and gratitude to all of the members of the Sacramento State Verbal Behavior Research Lab. Their feedback, encouragement, and friendship have made this process a very memorable one. As a final note, I would like to express thanks to my research assistant, Patricia Santos, for her continuous dedication and assistance with data collection and procedural implementation throughout this entire project. v TABLE OF CONTENTS Page Acknowledgments ………………………………………………………………………...v List of Tables……………………………………………………………………………..ix List of Figures …………………………………………………………………………….x Chapter 1. INTRODUCTION ……………………………………………………………….…..1 Common Methods………… ………………………………………………..........2 Video Modeling…………………………………………………………………...6 Video Modeling to Teach Play Skills....…………………………………………11 2. METHOD ……………………………………………………………………..........18 Participants ………………………………………………………….……….......18 Settings and Materials.……………………………………………….………......19 Play Sets………………….………………………………………………………20 Barn Set….……………………………………………………………….21 Garage Set………………………………….…………………………….21 Dependent Variables………………………………..…………………………....22 Scripted Vocalizations…………………………………………………...22 Scripted Play Actions…………………………………………………….23 Unscripted Vocalizations………………………………………………...23 vi Unscripted Play Actions…………………………………………………24 Interactive Play…………………………………………………………..24 Independent Variable……………………………………………………………25 Interobserver Agreement………………………………………………………...25 Social Validity…………………………………………………………………...26 Treatment Integrity………………………………………………………………27 Experimental Design……………………………………………………………..28 Procedures …………………………………………………………....…..……...28 Stimulus Preference Assessment………………………………………...29 Baseline (BL)………….………………..…………………………...…...29 Video Modeling (VM)…………………………...……………...……….30 No Video (NV)…………..…………………………………………...….31 Video + Social Consequences (VS)……………………………………...31 Generalization Probes (G)…….………………………………………….32 3. RESULTS ……………………………………………………………….…….........33 Nick….……………………………………………………….…………….........33 Levi………………………...………………………………………………...…..39 Social Validity…………………………………………………………………...45 4. DISCUSSION ………………………………………………………..…….….…....47 Limitations ……………………………………………………………...……….53 Future Research ……………………………………………………………...….56 vii Implications………………………………………………………………………57 Appendix A. Play Script Data Sheet- Garage Script………………………...………...59 Appendix B. Play Script Data Sheet- Barn Script …..……......…………………….…60 Appendix C. Interval Scoring Data Sheet….……………………………….………….61 Appendix D. Social Validity Scale...……………………......…………………….…...62 Appendix E. Video Modeling Treatment Integrity Checklist..……………………….. 63 Appendix F. Child Preference Questionnaire………………………………………….64 References ………………………..……………………………………………………...65 viii LIST OF TABLES Tables Page 1. Interobserver Agreement..………………………...…………….........................26 2. Treatment Integrity..……………………………………………….....................28 3. Social Validity……………………………………………………..…………….46 ix LIST OF FIGURES Figures Page 1. Results for Nick…………...………………………...…………………………...38 2. Results for Levi……………………………..……………………………………44 x 1 Chapter 1 INTRODUCTION Defining characteristics of children diagnosed with autism spectrum disorders include failure to develop verbal and nonverbal behavior, deficits in social development, and rigid or repetitive stereotypic behavior (Delano, 2007). With regard to social skills, these children typically engage in repetitive and/or isolated play while displaying inappropriate use of objects and/or toys. In addition, they often fail to demonstrate spontaneous, pretend, or symbolic play (American Psychiatric Association, 1994) and the social functioning skills of this population tend to be very limited (Scattone, 2007). Children with autism typically have social skill performance and fluency deficits which refer to not knowing how to perform a particular social skill or failure to discriminate which social behaviors are appropriate in specific situations (Gresham, Sugai, & Horner, 2001). The literature on teaching social and play skills to children with autism has continued to grow in the last decade as researchers and practitioners learn more about the social deficits of the disorder (Bellini & Akullian, 2007) and develop better technology for teaching these skills (Scattone, 2007). It is important that these individuals be taught to engage in appropriate social behavior early on in development to increase their participation in social activities with their peers and adults (Scattone, 2007; Burgess & Turkstra, 2006) as well as give them the tools they will need to function in society. Additionally, early identification of social deficits in children with autism 2 (Wetherby, Watt, Morgan, & Shumway, 2006) and early intervention in social skill training is most beneficial as the developmental gap between children with autism and their typically developing peers becomes smaller (McGee, Feldman, & Morrier, 1997). Common Methods Some common methods used to teach social skills to children with autism have varied from peer-mediation and imitation (Garfinkle & Schwartz, 2002; Haring & Breen, 1992) to theory of mind training (Ozonoff & Miller, 1995). Although literature suggests that students with autism learn equally well from either an adult or a peer model (Ihrig & Wolchik, 1988), using peers to help facilitate social communication and interaction skills has become increasingly popular (Chung, Reavis, Mosconi, Drewry, Matthews, & Tassé, 2007; Maione & Mirenda, 2006; Reagon, Higbee, & Endicott, 2006). MacDonald, Sacramone, Mansfield, Wiltz, & Ahearn (2009) suggested that in order for children with autism to maintain peer social interactions, it is necessary to facilitate peer mediated social reinforcers (i.e., using the peers in the training process to pair with reinforcement) as social consequences are not typically as reinforcing for children with autism. For example, most typically developing children enjoy play, and interacting with other children. Conversely, the play behavior of children with autism is typically reinforcing when playing alone or repetitively with certain toys or materials while their play has no real symbolic or social quality. Chung, et al. (2007) taught social communication skills to children with autism through a peer-mediated social skills training (SST) program. Four boys with autism, 3 ages 6-7 years, participated in a 12-week, training program which targeted six communication skills including contingent response, securing attention, initiating comments, initiating requests, and changing the topic. The typically-developing peers received pre-training and role play practice to orient them to the target skill for that day as well as teach them how to prompt and reinforce the target skill. Each training session followed a similar structure which included explanation, teaching, practice, and video times. A basic comparison design was used for each child to evaluate treatment effectiveness. Results indicated that peer-mediated social skills training combined with behavior management and video feedback quickly improved the social communication skills of children with autism from baseline levels. This study supports previous findings that also suggested peer-mediated social skills training (i.e., initiations, prompting, and reinforcement procedures) produced significant increases in the social behavior of four children with autism (Strain, Kerr, & Ragland, 1979). Pierce and Schreibman (1997) used pivotal response training (PRT) with typically developing peers to increase social behaviors of two boys with autism, ages 7 and 8 years. Multiple peer trainers were used to assess the generalization of treatment effects and a multiple-baseline design was used across different peer trainers. Dependent measures included three social behaviors consisting of maintaining interactions, initiating conversation, and initiating play. Results of this study suggested that the PRT intervention was a naturalistic and effective method for producing higher levels of social 4 behavior. Additionally, the use of multiple peer trainers was effective in increasing generalization in the social behavior of the target children. Shabani, Katz, Wilder, Beauchamp, Taylor, & Fischer (2002) increased social initiations in three boys (ages 6 and 7-years) with autism through the use of a tactile prompt (i.e., vibrating pager). They used an ABAB design to assess the effects of the tactile prompt on verbal initiations and responses to peer initiations. Training sessions consisted of teaching the children to make a vocal initiation related to a toy or play activity when the pager vibrated. Adults immediately prompted (i.e., verbal model) the verbal initiation after activating the timer every minute and reinforced the child for correct initiations after the model. A least to most prompting hierarchy was used to fade the prompts as the children demonstrated independent initiations when the pager was activated. Results suggested that the tactile prompt was an effective and unobtrusive method for teaching verbal initiations in children with autism. Kamps, Leonard, Vernon, Dugan, and Delquadri (1992) used a social skills group training method to increase social interactions between three students with autism (each age 7-years) and their typically developing peers in a first grade class. A multiplebaseline across-subjects design was used to assess the effectiveness of the group social skills training. Training was conducted during the first 10 minutes of 20-minute playgroup sessions. A variety of social skills were targeted including greetings, conversations, giving and accepting compliments, and including others in activities. Feedback was given in the form of reinforcement (i.e., a star placed next to the student’s name) for engaging 5 in trained social interactions at 1-minute checks during the 20 minute play session. Follow up sessions were conducted at the end of the school year where feedback was reduced to two 5-minute checks during a 20-minute play session. Results of this study suggested that the group social skills training increased the frequency, time engaged, and duration of social performance in the children with autism. Additionally, positive changes in the responsiveness of peers to the target children’s social behavior were observed. Although many of these studies have found positive changes in the social behavior of children with autism, there are still many limitations and additional areas of future investigation needed within the body of research exploring the social skills of children with autism. For example, recruitment of participants with similar communication deficits, cognitive levels, and ages was identified as a limitation by Chung, et al. (2007) as it made for a lack of control group. Additionally, the multicomponent training program utilized in their study made it difficult to determine which components of training were absolutely necessary for the increase in skill acquisition. Similarly, Whalen, Schreibman, and Ingersoll (2006) found that their evidence did not explain which behavioral mechanisms were responsible for the changes in the children’s behavior. Thus, a component analysis was necessary. Lastly, Kamps, et al. (1992) suggested the need for examining the quality of social interactions that are being trained. Unclear or ineffective maintenance and generalization training or performance of the acquired skills have been a common limitation of social skill training studies as well 6 (Kroeger, Schults, & Newsom, 2006; Greshm, Sugai, & Horner, 2001; Haring & Breen, 1992). Video Modeling Video modeling is a teaching method that has been found to have potential benefits over the other methodologies previously discussed, including using multiple stimulus and response exemplars, demonstrating the behavior in relevant contexts, consistency in presentation of training (Baker, Lang, O’Reilly, 2009), and facilitating observational learning (Corbett, 2003). In addition, Charlop-Christy, Le, and Freeman (2000) found that video modeling resulted in faster acquisition of tasks than in vivo modeling, and also promoted generalization of skills across people, settings, and stimuli. Past literature has also identified video modeling as being more effective and time efficient than in-vivo modeling (Charlop & Milstein, 1989; Charlop-Christy, Le, & Freeman, 2000).Video modeling has also been more utilized than those previously discussed. Video modeling typically consists of presenting the learner with an individualized video (Delano, 2007) of models (e.g., peers, siblings, adults, or self) engaging in a specific series of behavior, which are targeted for improvement (Bellini and Akullian, 2007; MacDonald et.al, 2009) or acquisition. Kleeberger and Mirenda (2010) discussed the impact of the model in video-based teaching and how generalized imitation is a prerequisite to being able to copy the behaviors modeled in the video. Responding is considered imitative when someone duplicates the behavior or similar behavior modeled by another person (Catania, 2007). 7 Imitation is a fundamental skill that enables learners to acquire more from their environment long-term and also expand their repertoire to new environments (RosalezRuiz & Baer, 1997). Imitating requires attending to the relevant features of the model which may be problematic for children with autism as they often require assistance in attending to the critical features of the environment. In other words, these children are likely to attend to irrelevant stimuli in the environment rather than the important or critical features of the model. As a core deficit identified early in children with autism (Kleeberger & Mirenda, 2010), imitation also entails the ability to attend to who and what to model after (McCoy & Hermansen, 2007). As previously mentioned, it is common for individuals with autism to lack the ability to gather critical information from their environment so to help alleviate for these deficiencies, the video-model serves as a salient multi-component model to control and direct the child’s attention to what actions and vocalizations to perform, including which stimuli the child should be attending to (McCoy & Hermansen, 2007). The video-modeling technology has been used to teach a variety of skills to children with autism including complex imitation skills across novel stimuli (Kleeberger & Mirenda, 2010), appropriate transition behavior (Schreibman, Whalen, & Stahmer, 2000 ), perspective taking (Charlop-Christy and Daneshvar, 2003; LeBlanc, Coates, Daneshvar, Charlop-Christy, Morris, & Lancaster, 2003), daily living/self-help skills (Rosenberg, Schwartz, & Davis, 2010; Charlop-Christy, Le, & Freeman, 2000; Haring, Breen, Weiner, Kennedy, & Bednersh, 1995), social communication skills (Nikopoulos 8 & Keenan, 2004) and play skills (Tetreault & Lerman, 2010; MacDonald et al., 2009; Hine & Wolery, 2006; Reagon, Higbee & Endicott, 2006; MacDonald et al., 2005; Dauphin, Kinney, & Stromer, 2004; D’Ateno, Mangiapanello, & Taylor, 2003; Taylor, Levin, & Jasper, 1999 ). In addition, positive social, language, and academic outcomes have been acquired in children with a variety of disabilities through the use of this video technology (Sturmey, 2003). For example, Kleeberger and Mirenda (2010) taught a 4-year old child with autism generalized imitation skills (e.g., behaviors the child had never viewed before from novel people in novel environments) during song and toy-play activities through the use of a video model. A multiple-baseline across imitation activities design was utilized to evaluate the effectiveness of the video modeling intervention. The procedures included a baseline, a video modeling alone phase (VM), a VM plus highlighting critical components phase (VMH; i.e., the parents made comments while the child watched the video to call attention to the critical components), a VMH plus prompting and reinforcement phase (VMHPR), and a final VM phase. Results of this study suggested that video modeling alone was not sufficient to increase the child’s performance across imitation activities; however, when the VMHPR procedure was implemented, the child demonstrated accurate imitations of the actions for both mastered and novel actions which showed that reinforcement and prompting may have been the vital component for video modeling. 9 In another example, Charlop-Christy and Daneshvar (2003) used video modeling to teach perspective- taking to three boys diagnosed with autism, ages 6 and 9-years. A multiple-baseline across children and within-child across tasks design was used to assess the effects of the video model training on perspective-taking tasks. Additionally, to evaluate generalization of skills to novel stimuli, multiple probes were implemented. The boys viewed a video of adults performing perspective taking tasks while explaining their problem solving strategies. The video was presented twice and the boys were then tested on the perspective taking tasks such as asking the child where he thought the person in the video would look for a cookie). Correct verbal or nonverbal responses were scored when the child correctly answered a perspective taking question (e.g., “Where will Johnny look for the cookie?”). Generalization and maintenance of perspective taking were also taught through multiple exemplar training and assessed across stimuli. Results of this study indicated that perspective-taking skills were learned in each of the three children, as well as the generalization of untrained responses. Video modeling was a fast and effective way to teach new skills to this population. Similarly, LeBlanc et al. (2003) used a multiple-baseline design to assess the effects of video modeling and reinforcement to teach perspective-taking skills to three boys with autism, ages 7 to 15-years. Measures of perspective-taking tasks were administered to the children by presenting the video of an adult completing the task. Once the child finished viewing the video, the video was paused and the child was asked to respond to specific perspective-taking questions. Reinforcement was delivered for 10 correct responses and incorrect responses resulted in repeating the video while prompting the child to attend until imitation of the correct response occurred. Pre-tests, post-tests, and baseline, resulted in no feedback from the experimenters. Stimulus variations were developed for two of the trained tasks to assess for generalization during testing conditions. Results of this study indicated that all three children eventually mastered the perspective-taking tasks even when the novel stimulus variations were introduced as a result of the video modeling and reinforcement procedures used. Another study conducted by Charlop-Christy, Le, and Freeman (2000) used a multiple baseline design across five children and within child to compare video modeling to in-vivo modeling to teach daily- living skills (e.g., tooth brushing and face washing). More specifically, in the video condition for one child’s tooth brushing task, the child was presented with the video two times, and then told, “Let’s do the same, just like what’s on TV”). During the in-vivo condition for face washing, the child watched live models twice, and then was told, “Let’s do the same, just like they did.” These procedures were then repeated until the child reached performance criteria in each condition. Findings of the comparisons indicated that video modeling was a more effective and efficient technique for teaching self-help skills to a child with autism. This technique lead to a faster acquisition of skills than in-vivo modeling and also facilitated generalization of skills, which was not observed after presentations of in-vivo models. Video modeling has also been used to teach social communication skills to children with autism. Nikopoulos and Keenan (2004) implemented video modeling 11 procedures to teach social initiation skills to three children with autism, ages 6 to 7 years. They were interested in each child’s latency to respond to social initiations and their reciprocal play with different toys. The videos consisted of a typically developing peer and the experimenter engaging in simple interactive play with one toy. The child with autism viewed the video clip and then was taken into a room containing the toys used in the video. The child was moved on to the next condition when they demonstrated initiations within the first 25s of entering the room for three consecutive sessions in a row. The second condition followed the same procedures as the first without prior video viewing. The third condition looked at whether the children would make a social initiation for reciprocal play with different toys than the ones used in the video. A multiple- baseline across subjects design was used to evaluate the treatment effects. For all three students, overall social initiation and reciprocal play increased. The latency to engage in social initiations decreased with each of the children and both initiation and reciprocal play were maintained at follow up probes with two different toys. Video Modeling to Teach Play Skills Video modeling has been most commonly used to teach play skills to children with autism. For example, Hine and Wolery (2006) used point-of-view video modeling (i.e., from the child’s perspective) to teach two preschoolers with autism to engage in specific play actions. A multiple-probe across two play behaviors design was used to assess the efficacy of the video modeling. After conducting baseline probes, daily probes, video viewing, and daily practice segments were implemented. During the video 12 modeling segment, two minutes of the child’s favorite cartoon played to encourage attending to the video. After the cartoon clip, the video modeling clip began with the set of materials and a pair of “hands” manipulating them. Finally, another 60s of the cartoon played before the experimenter provided the child with the play materials. After viewing the video, the child was told to play with their toys in the absence of any prompting or reinforcement for engaging in the modeled actions. Generalization probes were conducted to assess performance in a different setting and with different materials. Procedures were the same as in treatment with the only difference being the location or the novel stimuli. Results of this study indicated that point-of-view video modeling was effective in teaching new play skills to the children with autism in the absence of direct reinforcement for modeled actions. Additionally, the generalization occurred in novel settings and with novel stimuli as a result of multiple exemplars. The type of video model may also be a factor in deciding how to best utilize the technology as a teaching tool. Palechka and MacDonald (2010) recently conducted a study which compared commercially-available videos (CAV) to instructor-created video (ICV) models to teach play skills to three children with autism, ages 4 to 5-years. Three children were exposed to one commercially-made video for one play activity and one instructor-made video for another play activity. A multi-element design within participant and across model types was used to compare the rate of acquisition of scripted play using ICV models versus CAV models. A multiple probe design was also implemented across participants. Results of this comparison indicated that both video methods produced 13 complex sequences of play in reasonable amounts of time; however, the instructorcreated videos were found to provide a more exact model of the play behaviors and toys that were to be duplicated in the following play session, which may have contributed to the success of two of the three children’s faster acquisition of play skills. The implications of this research suggest that types of videos may be important variables when teaching play skills to children with autism. In another study, Reagon, Higbee, and Endicott (2006) used video modeling to teach pretend play skills to a 4-year old child with autism and his sibling as a play partner. An AB replicated across four play scenarios design was used to train the pretend play skills. Procedures were similar to Hine and Wolery (2006) in that after an initial baseline phase, the participants were first instructed to watch a video of models depicting the play scenarios and then instructed to play with the materials. Likewise, prompting and reinforcement were not provided during the play sessions. In order to assess maintenance and generalization, follow up sessions were conducted. Throughout the study, data were collected on both scripted and spontaneous play behavior which depicted the rapid acquisition of scripted actions and statements. Findings of this research suggest that when trained, siblings are effective as models and play partners to children with autism. Additionally, the child with autism produced both scripted and unscripted play behavior after viewing the video models in the absence of any extraneous prompts or reinforcement. 14 Similarly, pretend play was also taught using video modeling by MacDonald, Clark, Garrigan, & Vangala (2005). They taught thematic pretend play scripts to two boys with autism (ages 4 and 7-years) using a multiple probe design within child across play sets. Specific scripts were developed to use for each of three different play sets (e.g., a town, a ship, and a house). Adult models were recorded acting out the play scripts using the characters and materials for each play set. These videos were shown to the children two times in a row, and then immediately following, the children were asked to play with the materials for four minutes. The children’s responses were scored correct based on the number of accurate scripted vocalizations and play actions. Unscripted verbalizations and play actions were also scored to evaluate potential novel behavior. Mastery probes and follow-up probes were conducted and suggested that video modeling produced extended sequences of play for the children who did not have a strong play repertoire preceding the study. Rapid acquisition of skills was observed in both children, in addition to some basic perspective taking skills (i.e., acting out actions with the play characters as if the characters were talking). This is a more complex social skill that was learned incidentally through the video modeling procedure. To expand on her previous research, MacDonald et al. (2009), studied the effects of video modeling on teaching two children with autism (ages 5 and 7-years) to engage in reciprocal pretend play with typically developing peers. Two pairs of children participated in this study. Each dyad consisted of a child with autism and a typically 15 developing child. A specific selection criterion was developed and used to identify appropriate peers for the study. Three play sets including an airport, zoo, and grill set were used to teach both pairs of children. Scripted play scenarios were used, which included both vocalizations and play actions with adults as models. Adults were used and matched for gender because they were easy to train and because the researchers believed that it would increase the possibility of imitation. The scripts were created based on videotaped observations and transcriptions of typically developing children playing with the specific play set toys. In this study, the independent variable was the presentation of video models depicting scripted play interactions for each of the play sets. Participants with autism were not exposed to any form of reinforcement or prompting during the study. Dependent measures included the occurrence of five target responses: a) scripted vocalizations, b) scripted play actions, c) unscripted vocalizations, d) unscripted play actions, e) cooperative play, and f) reciprocal verbal interaction chains. Data were collected on the 4-minute play sessions on each child for their particular role in the script. Unscripted verbalizations and actions were scored only during baseline and during mastery probe sessions to assess variations in play as a result of the videos. MacDonald, et al. found that both pairs of children demonstrated extended sequences of reciprocal pretend play across three commercial play sets as a result of the video modeling. Play sequences included increased chains of scripted vocalizations and actions, and although the children 16 primarily acquired scripted vocalizations and actions, a slight emergence of novel behavior was observed in the children with autism. A limitation of MacDonald, et al.’s study was that increased or extended novel play did not occur across participants. Reagon, et al. (2006) suggested that scripted actions and vocalizations generalize to new settings with new materials, which this study failed to control for. Finally, the authors discuss the possibility of the presence of adults acquiring stimulus control over responding as they stood behind the children collecting data. The adult’s presence was a possible confound as they may have exerted some level of control over responding, and it is difficult to tell whether the children performed because of the manipulated video modeling variables or because of the adult’s presence. The purpose of the present study was to extend the MacDonald et al. (2009) research to further assess the effects of video modeling teaching reciprocal play skills to children with autism. Additionally, the present study addressed some of the limitations that were identified in the previously described study conducted by MacDonald et al. While this study used a similar methodology described by MacDonald et al., it was not identical in that the manipulated variables produced extended unscripted play behavior in the participants under study as their “spontaneous” behavior was resulted in social consequences. Boudreau and D’Entremont (2010) suggested that the introduction of video modeling resulted in reduced frequency of both unscripted actions and vocalizations, which may have been a result of the inflexibility acquired while performing only the modeled or scripted behaviors shown in the videos. Keeping these 17 findings in mind, the present study aimed to control for the suppression of spontaneous play behavior by directly consequating any and all occurrences of unscripted actions or vocalizations to increase the frequency of this play behavior. A second goal of the present study was to add to the growing body of research to support these findings and best utilize the technology to improve the reciprocal social and communicative behavior of children with autism (Delano, 2007). 18 Chapter 2 METHOD Participants Two boys diagnosed with autism (Nick and Levi) participated in the study. Nick was a 6.4-year-old male who was enrolled in a typical first grade class full time and received after school behavioral intervention services in his home for 13.5 hours per week. At the start of the study, he had been receiving behavior therapy for 5 years. Nick communicated using complete sentences with a mean length of utterance of approximately 7 words. Per parent report, his main social skill deficits included a lack of interactions with his typically developing peers (e.g., he would ignore his peer when they ask him a question and he rarely initiated his own interactions) as well as engaging in a lot of repetitious or perseverative play. Nick never had any exposure to or experience with video modeling. Levi was a 5.11-year-old male who was enrolled in a typical kindergarten class full time, received behavioral intervention services in his home for 13.5 hours per week, and also received one hour of both Speech and Occupational Therapy services each week. At the start of the study, he had been receiving behavior therapy for a little over 3 years. Levi communicated using complete sentences with a mean length of utterance of 8-10 words. Per parent report, his main social skill deficits included a lack of pretend play skills and the inability to interact with and follow along with his typically developing 19 peers for extended durations (e.g., he would lose interest or engage more in solo play after a few minutes). Only on one occasion did Levi ever have exposure to or experience with video modeling. During the generalization phase of the study, two peer participants were recruited through each family and served as the peer play partners. Nick’s peer was a 6-year-old female who, like Nick, was also in first grade. Levi’s peer was a 4-year-old female who was in preschool. Both peers were enthusiastic to participate, were socially competent, followed adult instructions consistently, and displayed age appropriate play skills. Setting and Materials Baseline and training sessions were conducted in each of the boy’s own homes. The materials necessary for the play activities were placed in open areas on the floor in the living rooms. The videos were presented on a Magnavox© 10” TFT Monitor Portable DVD player. One to four sessions were conducted daily with each participant and the entire session never lasted longer than 30 minutes. The experimenter was responsible for videotaping each session from a distance of 5-10 feet from the play area. The child’s parent(s) and/or sibling would sometimes sit quietly in the room and observe the session and other times would be in another room of the house. Generalization probes were conducted in a different room (i.e., the Play Room for Nick and the Dining Room for Levi), of the child’s house after each play set was mastered. Each session was videotaped for data scoring purposes. 20 An adult research assistant served as the play partner during the sessions to promote reciprocal play interactions and generalization of social responses in the children while two different adults were videotaped acting out the sequence of social interactions. MacDonald et al. (2005) found that adults are easy to train and effective as models to teach play skills to children. Additionally, Ihrig and Wolchik (1988) found that children with autism were capable of learning equally well from both adults and peers as models. Each video was related to a designated play activity and contained 10 scripted vocalizations along with 10 scripted actions associated with the vocalizations. For example, a scripted vocalization included, “Oh no! A flat tire,” while the corresponding action included taking the mechanic out of the truck to look at the man’s tire. The scripts were developed based on previous observations of two typically developing children engaging in similar play actions and social interactions using the same play materials. Play scripts were recorded using a digital video camera and then transferred to a DVD for later convenience of viewing. The Barn script (See Appendix A) DVD was 1:25 minutes in duration and the Garage script (See Appendix B) was 1:13 minutes in duration. Play Sets Two play sets were used to teach reciprocal pretend play to participants. The play sets included a Fisher-Price Little People Animal Sounds Farm© and a Fisher-Price Little People Racin’ Ramps Garage©. Each play set had a base structure (e.g., the barn or the car garage) and multiple characters and objects (i.e., 10 for the Barn and 5 for the Garage) related to the theme of the structure that were used in the video modeling scripts. 21 Each of the play sets had features that made sounds (e.g., a pig sty that oinked when you opened the gate in the barn or a bell that rang as you lifted the elevator up the garage) related to the theme that was also used in the play scripts. The following play sets were chosen based on gender neutrality and age appropriateness. Barn Set The barn base structure was a two-story red barn with an attached feed storage tower. There were three different gate entrances to the bottom story of the barn, and each of the gates made sounds related to the corresponding animal that belonged behind that specific gate. The characters included a farm boy, a farm girl, and six basic farm animals (i.e., a horse, cow, pig, sheep, goat, and chicken). The additional objects included a tractor with a trailer, a basket of apples, and a tall corn stalk. Adults in these videos modeled play interactions by manipulating the characters and animals (e.g., driving the car), speaking for the characters and animals (e.g., “Time to feed the pig”), and manipulating the materials through the characters (e.g., making the farmer get in the tractor to drive) from the Barn Set. Garage Set The garage base structure had a winding ramp for cars to drive down, a car elevator, and, three levels of parking. Features of the garage included a gas pump, a hose for putting air into the tires, and an elevator. The elevator had a bell that rang as it is lifted up to each parking level. The characters included a mechanic, a man, and a police officer in a police car. The additional objects included a tow truck and a blue car. Adults 22 in these videos modeled play interactions by manipulating the characters and vehicles (e.g., driving the car down the ramp), speaking for the characters (e.g., “Time to wash your car”), and manipulating the materials through the characters (e.g., making the character pump gas into the car at the gas pump) in the garage set. Dependent Variables All sessions were videotaped and scored at a later time for the occurrence of the following responses: a) percentage of scripted vocalizations, b) percentage of scripted actions, c) percentage of unscripted vocalizations, d) percentage of unscripted actions, and e) percentage of play interactions. Scripted Vocalizations Scripted vocalizations were defined as vocal statements, phrases, or sentences that matched the statement of the video model while allowing for the omission or substitution of modeled sound effects, conjunctions, articles, prepositions, or pronouns (e.g., substituting “or” for “and” or dropping the “the”). Additionally, the change of verb tense or the rewording of a line while maintaining the same general context of the model were also counted as correct scripted vocalizations. For example, “Time to feed pigs” versus “I’m going to feed the pigs now” was coded as a correct scripted vocalization. A scripted behavior would only be scored as incorrect if the child completely omitted a scripted vocalization or action; otherwise, it would be scored as an unscripted vocalization. 23 Scripted Play Actions Scripted play actions were defined as motor responses that matched the action of the video model and produced the same change in the environment that had been modeled in the video. Data was also collected and scored as correct on actions that were similar but not identical to the modeled responses (e.g., the child is hopping the cow instead of making him run to the barn or driving the tractor around a different way but ending up in the same spot). A scripted action would only be scored as incorrect if the child completely omitted a scripted action; otherwise, it would be scored as an unscripted action. Unscripted Vocalizations Unscripted vocalizations were defined as vocalizations including statements, phrases, or sentences that differed from the script in the video model by more than sound effects, conjunctions, articles, prepositions, or pronouns. They were also considered unscripted if there were changes in the verb tense but were deemed appropriate. In order to score the vocalization as unscripted, the child had to state a word that varied from that of the scripted model (e.g., the child said, “My pig likes to drink water!” instead of saying, “Time to feed the pigs” or “These apples taste so good” versus “Mmm, I like apples!”). These were sentences that did not occur in the script and were therefore coded as unscripted. Repetitions of the same statements were only scored as unscripted the first time they were vocalized within that session. Additionally, vocalizations were not scored as unscripted if the child repeated a line that was modeled from the play partner’s script. 24 For example, a vocalization was not scored as unscripted if the child simply added an article or two such as, “I think I’ll wait upstairs” versus “I’ll wait upstairs.” Unscripted Play Actions Unscripted actions were defined as motor responses that were not modeled in the video scripts but were appropriately related to the context of the play activity. The play action was not scored as unscripted if the action was simply repetitious in nature (e.g., driving a car back and forth in a repetitive motion) or if it was an action that was modeled from the play partner’s script (e.g., if the play partner’s action was to wash the target child’s car and the target child imitated this to engage in the same actions to wash his own car). In order to score the play action as unscripted, it had to vary from that of the play action modeled in the video (e.g., the child gave the apples to the chicken instead of feeding them to the horse and pig). Repetitions of the same actions were only scored as unscripted the first time they were performed within that session. Additionally, actions that were corrective in nature (e.g., picking up a toy that was accidentally dropped) were not scored as a spontaneous response. Interactive Play Interactive play included offering and receiving materials to and from the play partner (e.g., handing the play partner the apples to put inside the barn or accepting a toy that is being offered), as well as talking to each other about the same scripted or unscripted topics (e.g., driving different cars down the ramp and reciprocating information about their cars). Play was not counted as interactive if the duo was simply 25 engaged in parallel play (e.g., sitting and playing next to each other in the absence of interacting) or if it was repetitive or stereotypic in nature (e.g., the child makes the same repetitive initiation to the play partner several times). Independent Variable The independent variable in this study was the presentation of video models depicting scripted play interactions for the two sets of play materials (i.e., Barn and Garage). In addition, two conditions of video viewing were examined that related to program consequences. The first condition included the viewing of videos alone (VM; i.e., no programmed consequences provided for correct scripted or unscripted actions/vocalizations during play). The second condition of viewing included the viewing of videos and the addition of social consequences for unscripted behavior and play interactions. Interobserver Agreement Interobserver agreement (IOA) was assessed on 67% of Nick’s sessions and 67% of Levi’s sessions. IOA was calculated by dividing the number of agreements by the total number of agreements plus disagreements and multiplying by 100. An agreement was scored as any scripted or unscripted vocalization or action in which both observers scored either the occurrence or nonoccurrence of behavior. Disagreements were scored as any scripted or unscripted vocalization or action in which one observer scored an occurrence of a behavior, and the other did not. The primary researcher acted as the primary coder to transcribe and score the occurrences of the dependent measures from all 26 videos across both participants and phases. In addition, a research assistant independently coded IOA for each of the videotapes across all participants and phases. Mean IOA was 94.6% across participants. Table 1 contains individual IOA values for each participant. Table 1 Interobserver Agreement (IOA) – Dependent Measures Nick Levi All # Sessions 43 36 79 # Sessions IOA 29 24 53 % Sessions IOA 67 67 67 % Scripted Responses 100 99 99.5 % Unscripted Responses 92 92 92 % Interaction 99 97 98 Social Validity Social validity for this study was assessed through the use of a post test questionnaire created for the purpose of this study which required parents to rate his/her child performance in relation to the dependent measures. Specifically, parents were asked to provide feedback about the video modeling procedure using a 5-point Likert-scale. The questionnaire consisted of a 10-item scale with 5 meaning “Strongly Agree” to 1 meaning “Strongly Disagree.” Examples of questions include, “Video modeling is a socially appropriate training method for my child” or “I feel video modeling produced little change in my child’s pretend play during the session.” (See Appendix D) 27 Treatment Integrity In order to assess the degree to which the treatment sessions were implemented according to the protocol, treatment integrity checklists were completed (See Appendix E). This checklist included the tasks that were expected to occur in each phase and each condition of the study (e.g., placement of materials, delivery of reinforcement, technical difficulties, etc.). A session was considered correct if the experimenter(s) implemented all procedures as outlined in the protocol for that particular session. A session was considered incorrect if more than one step was implemented differently than outlined in the protocol. The total number of correct sessions divided by the total number of sessions multiplied by 100 will be calculated to determine the overall percentage of treatment integrity of the independent variable. Treatment integrity was assessed for the study during 100% of sessions across each video modeling condition. The only issue that occurred (on multiple instances) throughout the entire study was some minor technical difficulties with the digital timer on five separate occasions. Upon these occurrences, we were able to quickly correct the difficulty with the timer and resume data collection as scheduled. 28 Table 2 Treatment Integrity (T.I.) Nick Levi All Total # Sessions 43 36 79 # Sessions T.I. 43 36 79 % Sessions T.I. 100 100 100 % Correct Sessions 93 94 93.5 Experimental Design A multiple-probe design across play scripts was used to assess the effects of the video modeling intervention on play behaviors. The play sets were counterbalanced across each child to control for order of effects. For example, Nick learned the garage script first and then the barn script, while Levi learned the barn script first and then the garage script. Baseline probes were conducted prior to the introduction of a new play set. A generalization probe was conducted to assess for generalization of the acquired skills in another room of the house at a one-week follow up at the end of the intervention. Procedures Two adults (i.e., the primary researcher and one research assistant) were assigned to consistent roles across participants and throughout the course of the study. The primary researcher was responsible for videotaping the sessions. The research assistant acted as the peer “play partner” during each session with each of the children. Neither the experimenter nor the research assistant acted as the models in the videos to capitalize on 29 the effectiveness of the videos and to keep the actors “neutral” throughout the course of the study. Stimulus Preference Assessment Prior to the start of data collection, parents of the target children were asked to complete a child preference questionnaire (See Appendix F) that was developed for the purposes of this study. This form provided information to develop a selection of potential reinforcers or prizes for the children to earn throughout the course of the study. Nick and Levi each had their own prize box filled with a variety of tangible items and social activity “coupons” (e.g., playing hide and seek or making up a silly story) that had been identified as preferred by their parents. The children had access to the prize box at the end of each session for participating in a session that day. Baseline (BL) Each child was presented with each of two sets of play materials without having viewed the modeling videos. The child was then given the instruction, “Go play” and allowed to play for three minutes in the absence of any programmed consequences from adults. The adult play partner played in a neutral fashion (i.e., not very enthusiastic or energetic) without initiating any play with the child (i.e., the adult engaged in solitary play and only interacted with the target child if he made the initiation). The target child was instructed to sit on the left of what the play set while the adult play partner was placed on the right for playing, as this corresponded with the positions of the characters they would assume during video modeling training. Additionally, the play materials were 30 presented on the floor and positioned in the same way that they would be presented in the videos. Baseline probe procedures continued until responding reached stability (i.e., the data show no evidence of an upward or downward trend, and all of the data points fall within a small range of values; Cooper, Heron, & Heward, 2007). Video Modeling (VM) During the video modeling condition, the materials were set up identical to how they were in baseline. The child was assigned a constant “role” in each video to watch (e.g., “Watch the girl”) which corresponded with the script that they would be performing during the play session. They then viewed the specific target video two consecutive times before being allowed access to the toys. After viewing, the children were immediately directed to the play materials and told, “Go play!” The duo was allowed three minutes to play or until they stopped playing (e.g., sat in the absence of making vocalizations or manipulating any play materials for longer than 15 seconds) before the three minutes elapsed. There were no programmed consequences for this play period, regardless of if the response were scripted or not. If the target child engaged in his scripted response, the adult play partner immediately moved onto their next line. If the target child engaged in an appropriate unscripted response or initiation, the play partner would quickly respond in a neutral way and then immediately move onto their next line of the script. For example, if the child made an unscripted comment such as, “I think the chicken laid an egg!” the play partner would simply respond with, “Really?” and then move onto her next line. In other words, the play partner did not provide a lot of acknowledgement or 31 enthusiasm in her response to the child’s comment. If the target child engaged in an inappropriate or incorrect response, the adult play partner ignored the child for 10 seconds and then moved onto their next line in the script. For example, if the child made his car repetitively crash into the play partner’s character in a rough manner, she would then ignore this for 10 seconds and then move onto her next scripted line. These video viewing sessions continued until the target child demonstrated a steady rate of responding, regardless of whether they engaged in the scripted behavior or not. No Video (NV) During No Video (NV) condition, the videos were removed and everything else was identical to the initial baseline condition as described above. Video + Social Consequence (VS) During the VS condition, the materials were set up identical to how they were in the first two conditions. Any and all occurrences of unscripted behavior and interactions received a social consequence while all scripted responses were ignored and had no programmed consequences. Scripted responses were being ignored as to further increase the amount of unscripted play behavior and interactions that the children engaged in than in previous phases. Social reinforcement was provided briefly by the adult play partner in the form of an energetic and enthusiastic response to the child’s spontaneous vocalization, action, or interaction (e.g., if the child initiated, “Let’s have a race,” the play partner would say, “Oh cool! That sounds like a really fun idea! You go first!” If the child continued to engage in these unscripted responses, the play partner continued to 32 provide the same level of social consequence for each response, creating an extended reciprocal play chain. As mentioned earlier, if the target child engaged in an inappropriate or incorrect response, the adult play partner ignored the child for 10 seconds and then moved onto their next line in the script as to redirect the child back to appropriate play. If the child engaged in one of their scripted responses, the play partner would respond with her next line in the script. The difference between this condition and baseline was that the adult play partner delivered her response to unscripted behavior with a lot more interest, enthusiasm, and energy versus in baseline; she would have simply stated something neutral such as, “nice!” or “cool” while providing minimal eye contact in response to the child’s initiation or spontaneous play. This type of social consequence was a more natural and socially appropriate way of increasing the unscripted play behavior that the children engaged in. Generalization Probes (G) Generalization sessions were conducted one week after mastery of video skills with a similar aged peer play partner (i.e., rather than the adult play partner) in a different room of the house to assess for generalization. The same play sets, characters, and objects were used to observe whether generalization would occur with the same materials but while playing with a different person. 33 Chapter 3 RESULTS Nick Figure 1 depicts the percentage of scripted (closed circles and squares) and unscripted (open circles and squares) play behaviors, as well as interactive play (triangles) performed by Nick across the two different play scripts. During the baseline phase for the Garage play script (upper panel), Nick performed zero scripted actions and one scripted vocalization (M=5%). He also performed moderate levels of appropriate unscripted actions (M=42%) and vocalizations (M=50%). However, some of his spontaneous play was not necessarily related to the context of the play materials (e.g., vocalizing, “We’re going to Disneyland!” while keeping his character in the same position on the roof of the barn for several seconds), in which case was not scored as a correct unscripted response as it did not relate to the theme of the play set. He also displayed low levels of interaction (M=17%; e.g., he would simply make his character say, “Hi” to his partner’s character or he would move an object out of the way for the other character to get by but did not engage in extended sequences of reciprocal interactions). In some cases, Nick would receive a correct score for both unscripted behavior and interaction as he would be manipulating and sharing materials as well as engaging in some spontaneous vocalizations or actions. The introduction of the video modeling phase led to a rapid increase in the scripted actions (M=78%) and vocalizations 34 (M=73%) and drastically suppressed the unscripted actions (M=5%) and vocalizations (M=14%). After only the third session of viewing, Nick was performing the scripted actions and vocalization at 100% correct. Additionally, a rapid increase in play interactions was observed (M=96%) after the introduction of the video models (i.e., he was responding to every single initiation or comment made by his play partner that was built into the script). Once the videos were removed again during the No Video phase, Nick’s unscripted actions (M=57%) and vocalizations (M=45%) returned to near baseline levels whereas his scripted actions (M=3.3%) and vocalizations (M=8%) decreased significantly. Here, he was only performing one to three of the scripted actions or vocalizations and displayed higher levels of unscripted behavior than in baseline. Noteworthy, his unscripted behavior was more thematic than what was observed during baseline. For example, during baseline, Nick made the man with a car fly in the air while vocalizing a random screaming sound such as, “ahhhh” or he bounced the man on top of the car garage while yelling, “We’re going to Disneyland” as mentioned earlier. Nick’s unscripted behavior now included more thematic play such as driving his car to different garage floors or traveling the farm tractor to pick up the corn in the field; two actions that were not modeled in the videos but were appropriately related to the play sets. Additionally, his interactions (M=23%) decreased to near baseline levels once the videos were removed and were variable throughout most of this condition. For example, Nick would show interest and make initiations toward the play partner during some sessions and would engage in more solo play during other sessions. 35 After the introduction of the video plus reinforcement phase, Nick’s unscripted actions (M=61%) and vocalizations (M=52%) increased while near zero levels of scripted actions (M=3%) and vocalizations (M=0%) were maintained from the previous condition. Once his unscripted play behavior started producing more enthusiastic and reinforcing responses from his play partner, he continued engaging in spontaneous play and stopped attempting to perform any of the scripted behavior as it was being put on extinction. Finally, the interaction levels (M=65%) increased and maintained at a higher level than previous phases as a result of the social consequences. All of the scores during the reinforcement phase were found to be significantly higher than the initial baseline phase. One week after the completion of the video modeling procedures, the generalization probe for the Garage Script was conducted. Nick did not perform any of the scripted actions or vocalizations but demonstrated the highest levels of unscripted actions (89%), vocalizations (78%), and interaction levels (78%) with his peer. During the baseline phase for the Barn play script (Figure 1, lower panel), Nick performed zero scripted actions and vocalizations, low levels of unscripted actions (M=33%) and vocalizations (M=41%) and a very low levels of play interaction (M=14%). He engaged in similar actions and vocalizations to those of his Garage Script baseline (e.g., making his character jump and say, “We’re going to Disneyland” or vocalizing several different sound effects as he moved his character about the play set without really paying attention to his play partner’s behavior). Again, the introduction of the video modeling phase led to a rapid increase in the modeled actions (M=49%) and 36 vocalizations (M=49%), but produced variable results with his unscripted actions (M=43%) and vocalizations (M=37%). Initially he performed six to seven of the scripted actions and vocalizations during the session, but consistently stopped responding at the same point in the script each session. After about six correct scripted responses, Nick would begin engaging in unscripted play behavior which was not always appropriate (e.g., silly or repetitive in nature), in accordance with our operational definitions. After the introduction of the video models, a large increase in play interaction was observed (M=57%). However, this level of performance did not maintain stability throughout this condition. Upon removal of the videos, Nick’s scripted actions (M=0%) and vocalizations (M=7%) returned to baseline levels. His unscripted actions (M=68%) and vocalizations (M=46%) slightly increased from the video model phase and his interactions (M=29%) decreased considerably (e.g., he returned to solo play and paid minimal attention to his play partner). Again, the introduction of the video modeling plus reinforcement phase led to a further increase in his unscripted actions (M=69%) and vocalizations (M=50%) and maintained at zero levels of scripted actions and vocalizations. Nick’s play continued to become more thematic as he received social consequences for appropriate comments and actions that were related to the theme of the play set. Finally, the interaction levels (M=67%) were the highest and continued on an upward trend as a result of the social consequences being delivered for all of his interactions. Similar to the first play script, all of the scores during the reinforcement phase were found to be significantly higher than the initial baseline phase suggesting that 37 the direct reinforcement or delivery of social consequences of unscripted play behavior created more opportunities than the video models were able to. During the generalization phase for the Barn Script, Nick performed one of the scripted actions (10%) and one vocalization (10%) but again displayed the highest levels of unscripted actions (72%), vocalizations (78%), and interaction levels (78%) with his peer. He maintained a close proximity and interacted with his peer the majority of the play session. 38 Figure 1. Results for Nick. Percentage of scripted and unscripted actions (circles), scripted and unscripted vocalizations (squares), and interactions (triangles) that occurred during the baseline (BL), Video Modeling (VM), No Video (NV), Video + Social Consequence (VS), and generalization (G) play sessions for Nick. 39 Levi Figure 2 depicts the percentage of scripted (closed circles and squares) and unscripted (open circles and squares) play behavior, as well as interactive play (triangles) performed by Levi across the two different play scripts. During the baseline phase for the Barn play script (upper panel), Levi performed zero scripted actions and very low levels of scripted vocalizations (M=3%) Although not very high, he displayed moderate levels of unscripted actions (M=33%) and vocalizations (M=41%) during baseline probes. His vocalizations and actions were somewhat thematic but mostly consisted of solo play which included many different sound effects and repetitive actions. For example, he made a variety of screeching or crashing car sounds (e.g., “Errrrrrrrrr” or “Ppsshhhhh”) or airplane sounds (e.g., “Neeeer”) as he manipulated the characters in ways that were not related to the play sets (e.g., flying them aimlessly through the air or bouncing them around the play set with no direction). Levi did display some interactive play behavior (M=35%) such as asking the play partner different questions about what she was doing, however, he mostly sat and just observed what his play partner was doing without joining in. After viewing the videos, Levi’s number of scripted actions (M=20%) and vocalizations (M=12%) slightly increased. Slightly differing from Nick’s results, Levi’s unscripted actions (M=80%) and vocalizations (M=68%) did continue to increase as well as his scripted responses during this condition. Likewise, his interactive play behavior increased (M=62%) and he demonstrated much more consistent reciprocal play as a result 40 of the video viewing. Consequently, Levi established a short history of reinforcement simply by playing with his partner as the number of sessions increased which in turn; created more interactions. As his scripted behavior did not significantly increase upon the viewing of the videos, Levi was given a specific “rule” to copy what he saw in the video (e.g., “Watch the girl in the video, and when it’s over, you are going to copy how she played with the toys.”). After receiving this rule, an inverse effect was observed where his number of scripted actions (M=88%) and vocalizations (M=73%) increased dramatically and his unscripted actions (M=25%) and vocalizations (M=11%) greatly decreased. Additionally, his interactive play behavior (M=90%) increased a significant amount as well. Levi was performing nine out of ten scripted actions and vocalizations while staying on cue with his play partner’s scripted lines. During the return to baseline or NV condition, Levi’s performance returned to above baseline levels with his unscripted actions (M=64%) and vocalizations (M=69%). Furthermore, Levi’s scripted actions (M=3%) and vocalizations (M=2%) returned to baseline levels as well. Although the removal of the video produced a decrease in his level of interactions (M=69%) as compared to when video was present, his scores remained higher than in baseline. It should be noted that Levi was constantly observing his play partner and developing ideas for his own unscripted play behavior although he did not always initiate interactions with her (e.g., the play partner would be racing her car down the ramp and Levi would start talking about how fast his car was while not necessarily directing his vocalizations toward her). 41 The introduction of the video modeling plus reinforcement led to further increases in his unscripted actions (M=72%) and vocalizations (M=79%) as compared to the No Video phase. Additionally, Levi did not perform any scripted actions or vocalizations which was maintained from the previous phase. Finally, Levi’s interaction scores (M=90%) exceeded both baseline and video modeling levels as a result of the social consequences delivered. The majority of Levi’s behavior during this condition resulted in social consequences, which in turn created increased levels of interaction with his play partner. All of the scores during the reinforcement phase were found to be significantly higher than the initial baseline phase. At the one week follow up session, the generalization phase for the Barn Script resulted in no observed scripted actions or vocalizations but Levi did demonstrate the highest levels of unscripted actions (83%), vocalizations (78%), and interaction levels (94%) with his peer. Noteworthy, all of his unscripted play behavior during this generalization probe was more appropriate and thematic as compared to the initial baseline probes. During the baseline phase for the Garage play script (Figure 2, lower panel), Levi performed very few scripted actions (M=3%) and zero vocalizations. He performed moderate levels of unscripted actions (M=39%), vocalizations (M=42%), and interactions (M=39%). Levi’s displayed similar play behavior to the Barn script baseline where he engaged in some thematic play and subtle interactions but mostly engaged in solo play as mentioned above. A very significant increase in modeled actions (M=93%), vocalizations (M=88%), and interactive play (M=95%) was observed after the introduction of the video 42 models. Levi did not require a rule this time as his behavior came under the control of the rule from the previous play script. Levi’s unscripted actions (M=8%) and vocalizations (M=5%) significantly decreased as a result of the video models suggesting that the video extinguished any spontaneous responding. Levi’s scripted actions (M=7%) and vocalizations (M=0%) returned to near baseline levels after the removal of the video models and an increase in unscripted play behavior was observed again. His unscripted actions (M=63%) and vocalizations (M=63%) increased from the video alone phase and his interactions (M=72%) decreased slightly yet remained higher than baseline. The introduction of the video plus reinforcement phase led to further increases in his unscripted actions (M=74%) and vocalizations (M=92%) and maintained at zero levels of scripted actions and vocalizations. Levi demonstrated appropriate thematic play both in his vocalizations and actions throughout this condition. Finally, the interaction level (M=96%) was found to be the highest as a result of the social consequences delivered for engaging in spontaneous behavior and interactions. Similar to the first play script, all of the scores during the reinforcement phase were found to be significantly higher than the initial baseline phase. After the one week follow up for generalization of the Garage Script, Levi performed zero scripted actions or vocalizations but again displayed high levels of unscripted actions (89%), vocalizations (72%), and interaction levels (78%) with his peer. Levi’s overall percentages decreased slightly while playing with the peer which may have been due to the history of reinforcement that was established with the 43 adult play partner, thus finding that the interactions with the peer may not have been as reinforcing to him. 44 Figure 2. Results for Levi. Percentage of scripted and unscripted actions (circles), scripted and unscripted vocalizations (squares), and interactions (triangles) that occurred during the baseline (BL), Video Modeling (VM), Video Model + Rule (VMR), No Video (NV), Video + Social Consequence (VS), and Generalization (G) play sessions for Levi. 45 Social Validity The overall social validity of the study indicated that both families were satisfied with the video modeling procedures. They strongly agreed with items such as, “Video modeling increased my child’s interactive play behavior during the sessions” and “I think video modeling is a socially appropriate training method for children with autism.” Both of the boy’s families disagreed with items such as, “Video modeling produced strong negative emotions in my child.” The results of the social validity questionnaires indicate that both sets of parents considered this type of procedure to be a socially valid training method for children with Autism spectrum disorders. 46 Table 3 Social Validity- Likert-Type Scale Ratings for Social Validity Nick Levi 1. I liked the video modeling 5 5 procedure 2. My child has demonstrated more 1 3 thematic play since the video modeling procedure. 3. I am satisfied with my child’s 3 4 Response to the video modeling procedure. 4. Video modeling increased my 5 5 child’s interactive play behavior during the sessions. 5. I thought the video modeling 5 5 Procedure decreased my child’s spontaneous play behavior. 6. Video modeling is a socially 4 4 appropriate training method for children with autism. 7. I think video modeling procedures 5 5 promoted my child’s unscripted play behavior during the session. 8. My child will benefit in the long run 3 3 from the procedures implemented in the study. 9. Video modeling produced negative 1 1 Emotional reactions in my child. 10. I feel video modeling produced little Change in my child’s pretend play 4 2 During session. ________________________________________________________________________ 47 Chapter 4 DISCUSSION The current study supports previous research related to using video modeling to teach play and interactive skills to children with autism (Hine and Wolery, 2006; MacDonald, et al., 2005; Palechka & MacDonald, 2010; Reagon, Higbee, & Endicott, 2006). Specifically, the current study replicated and extended the work of MacDonald, et al (2009) by showing the acquisition of scripted and thematic play behavior as a result of video modeling. Additionally, our results support the work by Boudreau & D’Entremont (2010) that shows that video modeling can have a suppressive effect on unscripted or “spontaneous” play behavior. Similar to previous findings, Nick and Levi demonstrated rapid acquisition of the scripted vocalizations and actions after viewing the video models as compared to initial baseline levels. Prior to the video modeling, both boys engaged in play that was somewhat “random,” lacking any real theme related to the play sets (e.g., they engaged in repetitive actions and the majority of their vocalizations were sound effects related to the vehicles in the play sets). Although they did have some appropriate pretend play skills in their repertoire prior to the beginning of the study, it was evident that the video modeling contributed to the increase in their appropriate and related vocalizations and actions. Upon the introduction of the video models, Nick immediately performed the scripted vocalizations and actions while the majority of his spontaneous play behavior 48 had completely suppressed. Although the video modeling condition showed us that Nick was successful in learning his portion of the play scripts and it did increase his thematic play, this may have been disadvantageous to him as he stopped engaging in the little unscripted behavior that he had initially displayed in baseline. Additionally, Nick’s parent did report that he had a long history of reinforcement for imitation which may have influenced how quickly he memorized the script. Once the videos were removed, Nick’s unscripted play scores returned to baseline levels and continued to progress in an upward trend as the intervention continued. Again, his unscripted actions and vocalizations were now more thematic and related to the play materials suggesting that the video models did have a permanent effect over Nick’s play behavior. For example, Nick would pretend to drive through a car wash on a different level of the garage or made his farm character ask the other characters what their names were and if they wanted to go for a ride with him in the tractor. Unfortunately, his interaction scores decreased upon the removal of the videos. Once social consequences were delivered for every initiation, spontaneous action, or vocalization he produced, his scores increased even further than in the video modeling condition. This suggests that social consequences were effective in increasing his overall performance scores. Differently than Nick, little change was observed in Levi’s scripted responses upon the introduction of the video models, whereas his spontaneous play behavior actually showed a dramatic increase. For example, his play became far more thematic and he was observed to make more reciprocal comments and actions to that of his play 49 partner. After four sessions of little to no change in his scripted responses, we gave him a “rule” to copy what the girl was doing and saying in the video. The rule was given to assess whether or not Levi was able to imitate the scripted responses modeled in the videos. Once the videos were removed again, both Levi’s unscripted play and interaction scores were higher than in the initial baseline. Upon receiving the social consequences for every initiation, spontaneous action, or vocalization he produced, Levi’s scores slightly increased again. This suggests that, like Nick, the social consequences were effective in increasing his overall performance scores and he was able to achieve and maintain high levels of interaction with his play partner. Both Nick and Levi displayed reasonably low levels (M= 13% for Nick and M= 37% for Levi) of interaction prior to the start of the study. During baseline, the children primarily engaged in parallel play with the adult play partner (i.e., they sat next to the adult but rarely made eye contact, did not respond to, or direct any initiations toward them. Following the introduction of video modeling, the children both produced significant increases in their interaction levels throughout the remainder of the study (M= 13% for Nick and M= 37% for Levi). These results support previous findings (MacDonald, 2009; Reagon, Higbee, & Endicott, 2006; Taylor, Levin, & Jasper, 1999) suggesting that video modeling can be a method of facilitating sequences of interactions with a play partner. It should be noted that the video modeling helped more with the participant’s overall interaction levels rather than the actual play language and actions they were engaging in. Naturally, engaging in appropriate vocalizations and/or actions is 50 a very important play skill, but the interaction a child has with his or her peers while using those play skills is the more valuable skill for them to learn. Interestingly, both Nick and Levi’s results for their unscripted actions and vocalizations indicated that the video models did, in fact have a suppressive effect on their spontaneous play behavior. This supports the findings of Boudreau & D’Entremont (2010) suggesting that repeated exposures to the same videos may suppress the production of unscripted or spontaneous play behavior in children who have some in their repertoire to start. The videos controlled the play behavior of both boys although no explicit instructions were given to copy what they observed in the video (with the exception of the rule given to Levi for the Barn script). It was interesting that they rarely added any additional actions or vocalizations to the scripted lines or actions even though they were both capable of doing so according to their baseline scores. Once they completed their portion of the scripted responses, they stopped playing with the toys all together. On only a few instances, Nick continued to play on his own after he completed the script; however, this additional play was not always appropriate. Levi never continued on to add his own actions or vocalizations into the play after he completed the scripted responses (e.g., he would stop what he was doing and look at the video camera to indicate that he had finished his script and was done playing). As part of the social validity survey, one of the questions was, “I thought the video modeling decreased my child’s spontaneous play behavior,” and interestingly enough, the parents agreed with this statement. 51 Another interesting finding of this study was when video was removed with both boys, the nature of the spontaneous vocalizations and play changed from its original form. Both boys demonstrated play that was more thematic and appropriate (i.e., related to the context of the materials and not stereotypic or repetitive in nature), suggesting that video modeling is a teaching tool that may affect future probability of behavior. Although both boys demonstrated some appropriate play and language skills, the scripted videos worked as a tool to facilitate spontaneous play ideas in each of the boys. In addition to the increases noted above, the generalization phase resulted in maintenance of high scores while playing with a similar age, typically developing peer. However, it is important to note that the social consequences provided by the peer were not as strong or enthusiastic compared to those delivered by the adult play partner. This has interesting implications in that our method of reinforcement may not have even been necessary to evoke higher levels of unscripted play and interactions as both boys continued to perform at high levels simply with the basic social consequences that their peers provided. Here again, we can support past findings that video modeling can be a powerful teaching tool in that the videos can foster thematic play that children with autism can learn to use when playing with same-aged peers (D’Ateno, Mangiapanello, & Taylor, 2003; MacDonald, 2009; Sancho, et al., 2010). In this study, we chose to use a less obvious and more socially appropriate method of reinforcement (i.e., compared to other types of reinforcement such as edible; Sancho, Sidener, Reeve, Sidener, 2010, verbal praise, or token delivery; Boudreau & 52 D’Entremont, 2010) to increase the amount of unscripted play behavior and interactions in both boys during the video plus reinforcement phase. We wanted to challenge this notion as we saw play increase as a result of very subtle reinforcement which was embedded into the play sessions. Reagon, Higbee, & Endicott (2006) suggest that video modeling alone was effective in producing scripted and unscripted behavior as well as interactions without the aid of outside reinforcement. Additionally, Boudreau & D’Entremont (2010) suggested that future studies should, “lessen or eliminate reinforcement all together” as these consequences did not necessarily produce dramatic changes in their play. It is likely that these children would encounter reinforcing social consequences through contingencies in the natural environment (e.g., a peer commenting, smiling, or responding to an initiation that the target child makes) which is what we attempted to simulate through our adult play partner’s interactions with the children during the social consequence phase. Koegel, O’Dell, & Koegel (1987) discuss the importance of employing naturalistic reinforcers to broaden treatment gains in children with autism. They also discuss the efficiency of video modeling in the absence of programmed consequences such as prompting and praise. Both boys ended up requiring a similar number of sessions to complete the entire training process (M= 20) with each boy participating in fewer sessions during their second play script. This may suggest that the video modeling process may have some history effects related to it and as the boys became more familiar with the training 53 procedures, they were able to acquire the modeled responses in fewer sessions for their second scripts. In summary, the results of the current study indicate that video modeling is an effective method for teaching thematic play and social skills to children with autism. Both of the participants met the performance criteria and ended up with significantly higher play and interactive scores than their initial baselines. Additionally, the overall quality of their play became more thematic and relevant to the play materials they were using. Limitations While the current study demonstrates strong experimental control and replication of the results across participants, there are some limitations that should be considered. Firstly, we were not able to produce the most naturalistic results possible as we used an adult as the play partner for convenience of data collection purposes. Although we did one generalization probe with a typically developing peer, the results would be much more meaningful had we conducted the entire study using a same-aged peer. If familiar peer participants were utilized throughout the entire study (versus for generalization only), we may have seen similar results, with the exception of potential extended maintenance of skills and generalization. In other words, the training with a familiar same-aged peer may have fostered generalized play skills across activities, play materials, and environments beyond the realm of the study. Chandler and Lubeck (1992) discuss the need for increased use of baseline and normative comparisons to better 54 control for potential treatment inflation which is something that may have occurred in the present study due to the utilization of an adult as a peer play partner. The scores may have been inflated with an adult as the play partner possibly due to the fact that children receiving behavior intervention services typically have a history of reinforcement with adults, therefore; the scores may have been inflated because of this. We were not able to include information about normative scores and how the target children performed in comparison to their typically developing peers in this study; however, future research should look into expanding on this idea and continue to utilize peers as much as possible during social skill training. Another limitation that should be noted in relation to using an adult as the play partner was the lack of control we had for the adult’s responding during the Baseline and No Video phases. Because the adult play partner was only instructed to respond in a neutral fashion (i.e., not very enthusiastic or energetic) without initiating any play with the child, her responding may have solicited more behavior at times and less at other times. In other words, no explicit guidelines were created around what specifically to say or not say in response to the child’s initiations. Future studies should look at developing a system to control for this issue and ensure that the play partner is consistently responding during each occurrence of an initiation. Each of the play scripts was developed based on observations of typically developing children interacting with the exact play materials used in this study; however, midway through data collection, it was noted that the garage script seemed easier for the 55 boys to learn than the barn script. The Garage Script required that the target child use only one character and one object throughout the entire script and the same for the play partner. Conversely, the Barn Script required each person to change characters, objects, and roles throughout the script, which may have been more difficult for the boys to learn. Interestingly, the Barn Script is where both boys demonstrated quite a bit of varied responding during the video modeling phase, which ended up taking them a little longer to learn that particular script. Future researchers should ensure that the play scripts are more consistent in terms of effort required to perform them in order to control for response requirement problems. A fourth limitation may have been that the participants may have experienced satiation from the repeated exposure to the videos and became less motivated to interact in ways that promoted spontaneity in their play. The repeated exposure to the videos may have extinguished the spontaneous play. As Boudreau & D’Entremont (2010) describe, Nick and Levi may have had displayed some “inflexibility” in their performance of only engaging in what was modeled in the videos while missing out on opportunities to add spontaneous vocalizations or actions to their play. It seems that performing what was observed in the videos and engaging in additional spontaneous play were competing behaviors making it difficult for the children to engage in both. Further examination of how to possibly fade the play scripts or create multiple exemplars to promote more generalized responding may be of great use in the elimination of the route responding that often accompanies the video modeling (D’Ateno, Mangiapanello, & Taylor, 2003). 56 A final factor to consider includes the profile of child that we used in this study. Both Nick and Levi had fairly advanced play skills to begin with, so we may have seen different results had we recruited participants with fewer play and language skills. These types of play scripts and videos would be more appropriate for children who have minimal appropriate vocalizations, actions, and interactions as the scripts would have given them some basic and appropriate play behaviors to engage in with a play partner. Again, the idea is for the videos to serve as a guide or building block of how to play and what to say, while also influencing increased interactive behavior. It is possible that the unscripted or “spontaneous” play behavior of the boys in this study could have simply increased with basic reinforcement or social consequences. Future Research Future research should seek to replicate and extend the current study to continue to examine the effects of video modeling and how to best utilize it as a teaching tool for children with developmental disabilities. In addition, the results found in this study may be strengthened and further supported if additional participants are used; especially those who have fewer play skills in their repertoire to begin with. This is a population who would most likely benefit from this type of intervention. Another area to examine in future research would be to measure the overall duration of interactions pre and post video modeling procedure as well as the quality of interactions taking place. While the children in this study may have interacted on several different occasions, we had no measure for their quality (i.e., meaningful or contextually 57 related interactions) or of any significant length. We recommend that both of these variables are investigated in future studies as they would provide valuable information into further developing the best methods for teaching social and interaction skills. It is important that we are not only teaching the children to interact but to engage in quality interactions with their peers. The play skills in this procedure were very basic and were designed for young preschool or kindergarten aged children with more severe deficits in their play and social skills than these two children displayed. It would be interesting to use these video modeling procedures with children who have less advanced play skills and see if the scripted videos promote their responding to the next level in the teaching procedure. The current study did not include criteria for probing of independence of either teaching procedure. Implications Of primary importance is the finding that in addition to the positive changes in their social behavior, for both of the boys, the video modeling actually suppressed their unscripted or spontaneous play. This suggests that video modeling may not have been the most appropriate method to increase the spontaneity in their pretend play skills. Nonetheless, it helped to foster their vocalizations into more thematic play when it was removed in which case, the videos could be used in a different way. Furthermore, the use of video modeling served as a method of decreasing problem or unrelated behavior (e.g., unrelated, stereotypic, or inappropriate sound effects or play) 58 as appropriate behavior increases through the implementation of the procedures. The more appropriate play and social skills we can teach, the more incompatible they become with the problematic or inappropriate play behavior. The use of social consequences (i.e., more energetic and enthusiastic responses) on the unscripted actions and vocalizations was a key component in the increase of unscripted behavior as well as interactions in both children. Video modeling could be an important tool to use in future programming for unscripted play behavior. This type of video modeling procedure may be more appropriate for children who have fewer verbal skills or who have less advanced play skills. As these children displayed moderate language and play skills to begin with, the videos did not serve as a discriminative stimulus for developing new play ideas in addition to the scripted responses like we had hoped. However, we do feel that the results of this study suggest that video modeling is an effective and efficient method of teaching a wide spectrum of skills to children with autism. 59 APPENDIX A Play Script Data Sheet- Garage Script Garage Play Script Date : _____________ Session Time: _____________ Mark data as (+) = Correct Scripted Actions + Vocalizations Observer: ____________ (-) = Incorrect Vocal S core Action S core Child: _____________ (NR) = No Response Unscripted Actions + Vocalizations (P) = Peer (TC) = Target Child P: "Good Afternoon!" (walking mechanic over to man in blue car). 1) TC: "My car is dirty!" (drives car up to mechanic) P: "No problem! Follow Me!" (drives tow truck around to entrance on side of garage) 2) TC: "I need gas too!" (follows tow truck to entrance on side of garage) P: "Scrub, scrub, swish, swish" (makes mechanic wash blue car) 3) TC: "I'll wait upstairs!" (puts man in elevatormoves green lever up to top floor- takes man out) P: "Stop! Security!" (drives police officer over to other man) 4) TC: "Hello Officer!" (makes man wave to police officer) P: "Okay! Goodbye!" (drives police car down the ramp) 5) TC: "Gotta get my car!" (gets back into elevatormoves green lever down to bottom floor- takes man out) P: "All clean! Don’t forget your gas!" (pushes blue car to man outside) 6) TC: "Thanks Mister! Glug, glug, glug" (drives man in blue car and fills up with gas) P: "Want to race?" (drives mechanic in tow truck to talk to man in blue car) 7) TC: "Sure- you go first!" (puts cars on top of garage) P: "Here I go- Weeeeeee" (rolling tow truck down the ramp) 8) TC: "My turn! Zoooom! " (rolling blue car down the ramp) P: "Oh no! A flat tire!" takes mechanic out of tow truck to look at man's tire) 9) TC: "Thanks for fixing it!" (gives mechanic a high five) P: All done working!" (puts mechanic back into tow truck) 10) TC: "See you soon!" (puts man in blue car and drives away) Score: # of Correct Responses/ Total Opportunities for Child to Respond Vocal Score- TC:_______/10 P:________/10 Action Score- TC: _______/10 P: ________/10 Additional Notes/Observations/Bx Issues: 60 APPENDIX B Play Script Data Sheet- Barn Script Barn Play Script Date : _____________ Se ssion Time : _____________ Mark data as (+) = Correct Scripted Actions + Vocalizations Obse rve r: ____________ (-) = Incorrect Vocal S core Action S core Child: _____________ (NR) = No Response Unscripted Actions + Vocalizations (P) = Peer (TC) = Target Child P: "Time to feed the animals!" (walks farm girl to the front of the barn) 1) TC: "Beep Beep… Got the apples!" (drives farm boy in tractor to pick up apples in barn shed- lifts door to get apples & puts in tractor). P: "Lets feed the cow" (walks cow out of barn over to the apples + makes cow eat apples) 2) TC: "I like apples! Oink Oink" (walk pig out of gate over to eat apples) P: "You animals need some exercise!" (makes farm girl talk to the animals and the farm boy) 3) TC: "Let's have a race." (walks horse out of barn, over near the barn shed) P: "On your mark, get set, go! Neighhh!" (runs horse toward tractor) 4) TC: "Yay! I win! (runs cow fast towards tractor) P: "Let's get a drink" (walks horse to get a drink on side of barn) 5) TC: "I'm thirsty! Gulp Gulp Gulp" (follows horse to side of barn, pretends to drink) P: "I want to visit chicken" (puts horse up top in chicken coup- slides chicken from left to right to make chicken sound) 6) TC: "I'm going to visit Mr. Goat" (walks into the barn where the goat is) P: "There's hay everywhere!" (makes farm girl pretend to sweep) 7) TC: "Im going for a ride" (drives tractor back to barn shed- makes farm boy put apples inside) P: "Okay Mr. Pig. Go to your home!" (opens pig gate, puts pig back in barn) 8) TC: "Time for lunch! Hop in!" (driving tractor to pick up farm girl) P: "Goodbye Animals!" (puts farm girl in back of tractor) 9) TC: "Honk honk honk!" (drives tractor around barn) P: "Thanks for the ride friend." (makes farm girl wave to farm boy) 10) TC: "See you tomorrow!" (parks the tractor near the barn shed) Score: # of Correct Responses/ Total Opportunities for Child to Respond Vocal Score- TC:_______/10 P:________/10 Action Score- TC: _______/10 P: ________/10 Additional Notes/Observations/Bx Issues: 61 APPENDIX C Interval Scoring Data Sheet Interval Scoring Data Sheet Child: Script: __________________ IOA Initials: ____________ Session: _______________ Date: ____________ Spontaneous Play (Partial Interval) Opperational Definitions: (V= Vocalization, A=Action) Unscripted vocalizations. Defined as vocalizations that differ from the script in the video model by more than sound effects, conjunctions, articles, prepositions, or pronouns, or changes in the verb tense but are deemed appropriate. In order to score the vocalization as unscripted, the child must state a word that varies from that of the scripted model (e.g., the child said, “M y pig likes to drink water!” instead of saying, “Time to feed the pigs” or “These apples taste so good” versus “M mm, I like apples!” “M y pig likes to drink water” did not occur in the script and was therefore coded as unscripted. Repetitions of the same statements will only be scored as unscripted the first time they are vocalized within that session. Additionally, vocalizations will not be scored as unscripted if the child repeats a line that was modeled from the play partner’s script Unscripted play actions. Defined as motor responses that are not modeled in the video scripts but are appropriately related to the context of the play activity. The play action will not be scored as unscripted if the action is simply repetitious in nature (e.g., driving a car back and forth repetitively in the same motion) or if it is an action that was modeled from the play partner’s script. In order to score the play action as unscripted, it must vary from that of the play action modeled in the video (e.g., the child was making the cow moo to the chicken instead of making him run to the barn). Repetitions of the same actions will only be scored as unscripted the first time they are performed within that session. Additionally, actions that are corrective in nature (e.g., picking up a toy that was accidentally dropped) are not scored as a spontaneous response. 10 sec 10 sec 10 sec 10 sec 10 sec 10 sec 1 Minute V A V A V A V A V A V A 2 Minutes V A V A V A V A V A V A 3 Minutes V A V A V A V A V A V A Vocalizations (V) # Intervals circled ________/ 18 Total = __________ % Actions (A) # Intervals circled ________/ 18 Total = __________ % IOA # Agreements________/ Total #Agreements+Disagreements = __________ % Cooperative Play (Whole Interval) Opperational Definition: (P= Proximity, O=Orientation, I=Interaction) Interactive Play. Interactive play includes offering and receiving materials to and from the play partner (e.g., handing the play partner the apples to put inside the barn or accepting a toy that is being offered) as well as talking about the same scripted or unscripted topics (e.g., driving different cars down the ramp and talking about how fast they were going). Play is not counted as interactive if the duo is simply engaging in parallel play (e.g., sitting and playing next to each other in the absence of interacting) or if it is repetitive or stereotypic in nature (e.g., the child makes the same repetitive initiation to the play partner several times). 10 sec 10 sec 10 sec 10 sec 10 sec 10 sec 1 Minute P O I P O I P O I P O I P O I P O I 2 Minutes P O I P O I P O I P O I P O I P O I 3 Minutes P O I P O I P O I P O I P O I P O I Proximity (P) # Intervals circled ________/ 18 Total = __________ % Orientation (O) # Intervals circled ________/ 18 Total = __________ % Interaction (I) # Intervals circled ________/ 18 Total = __________ % IOA # Agreements________/ Total #Agreements+Disagreements = __________ % 62 APPENDIX D Social Validity Scale After observing the video sessions, please answer all questions honestly and to the best of your ability. This will help to interpret the overall social validity of the study for future reference. Thank you for your participation in this survey. 1 = Strongly Disagree, 2 = Disagree, 3 = Neutral, 4 = Agree, 5 = Strongly Agree 1. I liked the Video Modeling procedure. 1 2 3 4 5 2. My child has demonstrated more thematic play since the video modeling procedure. 1 2 3 4 5 3. I am satisfied with my child’s response to the video modeling procedure. 1 2 3 4 5 4. Video modeling increased my child’s interactive play behavior during the sessions. 1 2 3 4 5 5. I thought the video modeling procedure decreased my child’s spontaneous play behavior. 1 2 3 4 5 6. Video modeling is a socially appropriate training method for children with autism. 1 2 3 4 5 7. I think video modeling procedures promoted my child’s unscripted play behavior during the session. 1 2 3 4 5 8. My child will benefit in the long run from the procedures implemented in this study. 1 2 3 4 5 9. Video modeling produced negative emotional reactions in my child. 1 2 3 4 5 10. I feel video modeling produced little change in my child’s pretend play during session. 1 2 3 Please provide any additional comments/suggestions about the procedure: 4 5 63 APPENDIX E Video Modeling Treatment Integrity Checklist Observer: __________________________________ Date: _________________ Child: _____________________________________ Session:_______________ Play Script: _________________________________ Materials/Equipment 1. Play Set (Barn or Garage) 2. Play Characters and Objects 3. DVD Player 4. Play Script DVD 5. Video Camera 6. Digital Timer 7. Adult play script cheat sheet 8. Prize Box Procedure 9. During video viewing, child and adult play partner were positioned to correspond with roles they assumed in the video models 10. Child viewed video two times 11. Timer was set for 3 minutes 12. Experimenter told child to, “Go Play” after video viewing 13. During play, child and adult play partner were positioned to correspond with roles they assumed in the video models 14. Child stays in the designated play area for duration of play 15. Correct scripted responses were reciprocated with the next line of the script 16. Appropriate unscripted responses received no social consequence 17. Appropriate unscripted responses were reinforced 18. Incorrect/Inappropriate responses were ignored for 10 seconds before moving onto next line in script 19. Session lasted no longer than 30 minutes 20. Child was given the opportunity to select an item from the prize box Yes No N/A Comments Yes No N/A Comments Total # of Steps Completed: _________ Total # of Correct Steps: ____________ # of Correct steps /Total # of steps completed : _____________ Total % of Tx Integrity for the session: _________________ 64 APPENDIX F Child Preference Questionnaire The purpose of this survey is to obtain information about the toys that you believe would be useful as rewards for your child. If you have signed the consent form, please answer the following questions regarding your child’s preferences: 1. Some children really enjoy toys (e.g., puzzles, books, blocks, figurines, musical toys, playdoh, etc.) and special treats (e.g., candy, fruit snacks, cookies, cereal, crackers, sodas or juices, etc.) that they do not have access to on a regular basis. What are the specific toys/edible items your child really likes? __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ 2. Please go back to the list in question number one, and place a number in each box to rank these toys from most favorite (1) to least favorite (10). 3. Are there any toys/edible items that you would prefer your child not to play with during our study? ________________________________________________________________________ 65 REFERENCES American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, DC: Author. Baker, S.D., Lang, R., & O’Reilly, M. (2009). Review of video modeling with students with emotional and behavioral disorders. Education and Treatment of Children, 32(3), 403-420. Bellini, S. & Akullian, J. (2007). A meta-analysis of video modeling and video selfmodeling interventions for children and adolescents with autism spectrum disorder. Exceptional Children, 73(3), 264-287. Boudreau, E. & D’Entremont, B. (2010). Improving pretend play skills of preschoolers with autism spectrum disorders: The effects of video modeling. Journal of Developmental Disabilities, 22, 415-431. Burgess, S. & Turkstra, L.S. (2006). Social skills intervention for adolescents with autism spectrum disorders: A review of the experimental evidence. EBP Briefs, 1(4), 121. Catania, A.C. (2007). Learning (Interim 4th ed.). Cornwall-on-Hudson, NY: Sloan Publishing, LLC. Chandler, L.K., Lubeck, R.C., & Fowler, S.A. (1992). Generalization and maintenance of preschool children’s social skills: A critical review and analysis. Journal of Applied Behavior Analysis, 25, 415-428. 66 Charlop-Christy, M.H. & Daneshvar, S. (2003). Using video modeling to teach perspective taking to children with autism. Journal of Positive Behavior Interventions, 5(1), 12-21. Charlop-Christy, M.H. Le, L., & Freeman, K.A. (2000). A comparison of video modeling with in vivo modeling for teaching children with autism. Journal of Autism and Developmental Disabilities, 30(6), 537-552. Charlop, M.H. & Milstein, J.P. (1989). Teaching autistic children conversational speech using video modeling. Journal of Applied Behavior Analysis, 22, 275-285. Chung, K., Reavis, S., Mosconi, M., Josiah D., Matthews, T. Tasse’, M.J. (2007). Peer mediated social skills training program for young children with high functioning autism. Research in Developmental Disabilities, 28, 423-436. Corbett, B.A. (2003). Video modeling: A window into the world of autism. The Behavior Analyst Today, 4(3), 367-377. D’Ateno, P., Mangiapanello, K., & Taylor, B.A. (2003). Using video modeling to teach complex play sequences to a preschooler with autism. Journal of Positive Behavior Interventions, 5(1), 5-11. Dauphin, M., Kinney, E.M., & Stromer, R. (2004). Using video-enhanced activity schedules and matrix training to teach sociodramatic play to a child with autism. Journal of Positive Behavior Interventions, 6(4) 238-250. Delano, M.E. (2007). Video modeling interventions for individuals with autism. Remedial and Special Education, 28(1), 33-42. 67 Garfinkle, A.N., & Schwartz, I.S. (2002). Peer imitation: Increasing social interactions in children with autism and other developmental disabilities in inclusive preschool classrooms. Topics in Early Childhood Special Education, 22(1), 26-38. Gresham, F.M., Sugai, G., & Horner, R.H. (2001). Interpreting outcomes of social skills training for students with high-incidence disabilities. Exceptional Children, 67(3), 331-344. Haring, T.G., Breen,C.G., Weiner, J., Kennedy, C.G., Bednersh, F. (1995). Using videotape modeling to facilitate generalized purchasing skills. Journal of Behavioral Education, 5(1), 29-53. Haring, T.G., & Breen, C.G. (1992). A peer-mediated social network intervention to enhance the social integration of persons with moderate and severe disabilities. Journal of Applied Behavior Analysis, 25, 319-333. Hine, J.F. & Wolery, M. (2006). Using point-of-view video modeling to teach play to preschoolers with autism. Topics in Early Childhood Special Education, 26(2), 83-93. Ihrig, K. & Wolchik, S.A. (1988). Peer versus adult models and autistic children’s learning: Acquisition, generalization, and maintenance. Journal of Autism and Developmental Disorders, 31(3), 343-349. Kamps, D.M., Leonard, B.R., Vernon, S., Dugan, E.P., Gershon, B., Wade, L. & Folk, L. (1992). Teaching social skills to students with autism to increase peer interactions 68 in an integrated first-grade classroom. Journal of Applied Behavior Analysis, 25, 281-288. Kleeberger, V. & Mirenda, P. (2010). Teaching generalized imitation skills to a preschooler with autism using video modeling. Journal of Positive Behavior Interventions, 12(2), 116-127. Kroeger, K.A., Schultz, J.R., & Newsom, C. (2006). A comparison of two groupdelivered social skills programs for young children with autism. Journal of Autism and Developmental Disorders, 37, 808-817. LeBlanc, L.A., Coates, A.M., Daneshvar, S., Charlop-Christy, M.H., Morris, C., & Lancaster, B.M. (2003). Using video modeling and reinforcement to teach perspective taking skills to children with autism. Journal of Applied Behavior Analysis, 36, 253-257. MacDonald, R., Sacramone, S., Mansfield, R., Wiltz, K., & Ahearn, W.H. (2009). Using video modeling to teach reciprocal pretend play to children with autism. Journal of Applied Behavior Analysis, 42, 43-55. MacDonald, R., Clark, M., Garrigan, E. & Vangala, M. (2005). Using video modeling to teach pretend play to children with autism. Behavioral Interventions, 20, 225-238. Maione, L. & Mirenda, P. (2006). Effects of video modeling and video feedback on peerdirected social language skills of a child with autism. Journal of Positive Behavior Interventions,8(2), 106-118. 69 McCoy, K. & Hermansen, E. (2007). Video modeling for individuals with autism: A review of model types and effects. Education and Treatment of Children, 30(4), 183-213. McGee, G.G., Feldman, R.S., Morrier, M.J. (1997). Benchmarks of social treatment for children with autism. Journal of Autism and Developmental Disorders, 27(4), 353-364. Nikopoulous, C.K. & Keenan, M. (2007). Using video modeling to teach complex social sequences to children with autism. Journal of Autism and Developmental Disorders, 37, 678-693. Nikopoulous, C.K. & Keenan, M. (2004). Effects of video modeling on social initiations by children with autism. Journal of Applied Behavior Analysis, 37, 93-96. Nikopoulous, C.K. & Keenan, M. (2003). Promoting social initiation in children with autism using video modeling. Behavior Interventions, 18, 87-108. Ozonoff, S. & Miller, J.N. (1995). Teaching theory of mind: A new approach to social skills training for individuals with autism. Journal of Autism and Developmental Disorders, 25(4), 415-433. Palechka, G. and MacDonald, R. (2010). A comparison of the acquisition of play skills using instructor-created video models and commercially available videos. Education and Treatment of Children, 33(3), 457-474. 70 Pierce, K. and Schreibman, L. (1997). Multiple peer use of pivotal response training to increase social behaviors of classmates with autism: Results from trained and untrained peers. Journal of Applied Behavior Analysis,1, 157-160. Reagon, K.A., Higbee, T.S., & Endicott, K. (2006). Teaching pretend play skills to a student with autism using video modeling with a sibling as model and play partner. Education and Treatment of Children, 29, 517-528. Rosalez-Ruiz, J. & Baer, D.M. (1997). Behavioral cusps: A developmental and pragmatic concept for behavior analysis. Journal of Applied Behavior Analysis, 30, 533-544. Rosenberg, N.E., Schwartz, I.S., & Davis, C.A. (2010). Evaluating the utility of commercial videotapes for teaching hand washing to children with autism. Education and Treatment of Children, 33(3), 443-455. Sancho, K., Sidener, T.M., Reeve, S.A., Sidener, D.W. (2010). Two variations of video modeling interventions for teaching play skills to children with autism. Education and Treatment of Children, 33(3), 421-442. Scattone, D. (2007). Social skills interventions for children with autism. Psychology in the Schools, 44(7), 717-726. Schreibman, L., Whalen, C., & 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. 71 Shabani, D.B., Katz, R.C., Wilder, D.A., Beauchamp, K., Taylor, C.R., & Fischer, K.J. (2002). Increasing social initiations in children with autism: Effects of a tactile prompt. Journal of Applied Behavior Analysis, 35, 79-83. Shipley-Benamou, R., Lutzker, J.R., & Taubman, M. (2002). Teaching daily living skills to children with autism through instructional video modeling. Journal of Positive Behavior Interventions, 4(3), 165-175, 188. 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(1), 41-54. Sturmey, P. (2003). Video technology and persons with autism and other developmental Disabilities: An emerging technology for pbs. Journal of Positive Behavior Interventions, 5(1), 3-4. Taylor, B.A., Levin, L., & Jasper, S. (1999). Increasing play-related statements in children with autism toward their siblings: Effects of video modeling. Journal of Developmental and Physical Disabilities, 11(3), 253-264. Tetreault, A.S. & Lerman, D.C. (2010). Teaching social skills to children with autism using point-of-view modeling. Education and Treatment of Children, 33(3), 395419. Wetherby, A.M., Watt, N. Morgan, L., & Shumway, S. (2006). Social communication profiles of children with autism spectrum disorders late in the second year of life. Journal of Autism and Developmental Disorders,37, 960-975. 72 Whalen, C., Schreibman, L., & Ingersoll, B. (2006). The collateral effects of joint attention training on social initiations, positive affect, imitation, and spontaneous speech for young children with autism. Journal of Autism and Developmental Disorders, 36, 655-664.