Healthcare Training Institute - Quality Education since 1979
Psychologist, Social Worker, Counselor, & MFT!!

Section 17
Peer-Mediated Interventions to Increase the Familial Social Interaction of Children with Autism

Question 17 found at the bottom of this page
Answer Booklet | Table of Contents
Get PRINTABLE format of this page

Intervention Approaches
The review will be organized by the approach used to change peer expectancies. First, interventions that arrange the situation or contingencies to promote optimal peer effort will be discussed. These include integrated play groups, peer buddy and peer tutoring approaches, and group-oriented contingencies. Second, interventions that promote peer effort by teaching peers skills for initiating with and reinforcing children with autism will be discussed. These include peer networks, pivotal response training, and peer initiation training. Third, interventions that change peer expectancies and promote interaction by teaching social skills to the children with autism will be discussed. These include target child initiation training and initiation training of both the target child and peers. Finally, comparative studies will be reviewed.

Manipulation of the Situation or Contingencies to Promote Interaction
The following interventions involve arranging the situation or contingencies to promote peer interaction with children with autism. As noted earlier, increased peer effort to interact enhances the likelihood that children with autism will attend to peer models--and thereby enhances learning. Specific methods include integrated play groups, the assignment of peer buddies, and group-oriented contingencies.

Integrated Play Groups.During integrated play groups, as used by Wolfberg and Schuler (1993,1999), an adult provides a structured environment and guides participation between children with autism and socially competent peers. A key facet of this approach involves providing a supportive environment to optimize interaction rather than using adult direction. Other important components of this method are a natural integrated setting, well-designed play spaces that take into account accessibility and size, and play materials that promote interaction. Integrated play groups also typically establish a consistent schedule and routine, use a small number of familiar peers, and match play activities to the child's developmental level (Wolfberg & Schuler, 1993). An adult monitors the play situation for evidence of developing play skills, interprets for and coaches the peers, and encourages the children to engage in activities slightly more advanced than their current abilities (Wolfberg & Schuler, 1999). In addition, the adult encourages the target child to engage in and maintain interaction by using prepared cues, such as posters, when the child seems uncertain. Such prompts are faded as the child begins to incorporate the strategies on his or her own.

Wolfberg and Schuler (1993) used a multiple-baseline design for three target participants in three different integrated play groups. All children were 7-year-old boys diagnosed with autism who had very little appropriate play, participated in a high degree of repetitive play, and had little to no language. All participants nearly doubled the amount of interaction with peers involving attention to a common activity in the final treatment condition. All participants also engaged in less repetitive play and more functional play, and all but one child engaged in more pretend play. Parents and teachers reported similar improvements outside the experimental setting, indicating some social validity and generalizability. However, there is no evidence that parents were kept blind to study hypotheses. In addition, initial behavior gains were not maintained when treatment was withdrawn. Although this is evidence that the behavior gains were due to the treatment, it also demonstrates that the treatment effect was dependent on adult support.

Roeyers (1996) also examined the possible impact of integrated play groups. However, in his study, the typically developing peers were informed about autism, and each was assigned to a target child; in addition, adults were less involved. Roeyers randomly assigned 85 children diagnosed with autistic disorder or pervasive developmental disorder not otherwise specified to an experimental or control group. All children were between 5 and 13 years of age and lived in the Dutch-speaking part of Belgium, but information on their level of impairment was not provided. The experimental group significantly increased the amount of time spent in interaction, increased the length of sustained interaction, increased their degree of responsiveness to the partner's initiations, increased the number of social initiations made, and decreased the amount of time spent in self-stimulatory behavior compared with the control group. Most increases represented a change of 20% or more over behavior prior to the intervention. However, despite these positive results, the interactions of the target children remained inconsistent and idiosyncratic.

Peer Buddy and Peer Tutor Approaches. Peer buddy and peer tutor approaches focus on dyads with one typically developing peer and one child with autism, rather than a group of children. Peer buddy approaches involve assigning each child with autism to a buddy, who is told to stay with, play with, and talk to the child with autism. Laushey and Heflin (2000) investigated this approach with two 5-year-old children diagnosed with an autism spectrum disorder. Both children had some language and could read at the kindergarten level but experienced social difficulties. Using a reversal design, the results indicated that the children with autism increased their social interaction 36% and 38% during the treatment phase, as compared with the baseline phase, in which children were integrated but not assigned a buddy.

Peer tutoring approaches consist of tutor-learner pairs and promote the incidental learning of social behaviors through natural interactions. Peer tutoring approaches have generally been studied with high-functioning school-aged children with autism (Kamps et al., 1994; Kamps, Dugan, Potucek, & Collins, 1999). Kamps et al. (1994) examined the peer tutoring approach using a multiple-baseline-across-participants-with-reversal design. Participants included three 8- and 9-year-old boys with autism, who were high functioning in terms of language and intellectual abilities but lacked social competence, and all other children in a third-grade classroom. Each week, students were assigned a different tutoring partner. Tutoring produced increases in interaction from 80 to 120 seconds per 5-minute sample for the three children with autism. In addition, the mean interaction time of peers increased, and the children with autism displayed improved academic achievement.

Kamps et al. (1999) studied a slight variation of the peer tutoring approach by having moderate-to high-functioning school-age children with autism tutor typically developing first-grade students who were experiencing academic difficulties. An ABAB withdrawal design with replication was used to examine the impact of three 9-year-old children with autism and five fourth-grade girls tutoring six first-grade students. All three children had higher mean free time social interaction following the peer tutoring. In addition, the first-grade children who were tutored improved academically compared to those who were not tutored. A second part of this study used the same tutoring program with four 10- to 12-year-old children with autism as tutors. The results were similar but more variable and less dramatic.

Group-Oriented Contingency. Group-oriented contingencies require that all children in a class engage in a specified behavior in order to receive a reinforcer. Both studies that used this approach also provided social skills training prior to using the group-oriented contingency (Kohler et al., 1995; Lefebvre & Strain, 1989). Group-oriented contingencies promote the emergence of corollary, or untrained, supportive behaviors among the children to influence one another's performance. An advantage of this approach is that teachers are able to more efficiently manage large groups.

The group contingency method has been found to increase social interactions of 4- to 6-year-old children with autism (Kohler et al., 1995; Lefebvre & Strain, 1989). Kohler et al. used group-oriented contingencies with three 4-year-old children with autism and six typical peers ranging in age from 3 to 4 years. Prior to the study, none of the children with autism engaged in more than occasional interactions with peers, and only one child used appropriate play skills. A withdrawal-of-treatment design was used with alternating baseline, social skills training, and group contingency conditions. The classwide social skills training package was developed by Odom, Kohler, and Strain (1987) and included play organizer suggestions, share offers and requests, and assistance offers and requests. The class-wide supportive skills training included reminding one another to use these skills. The amount of time that children with autism and their peers engaged in social interaction increased from 28% to 65% during group-oriented contingency conditions. However, rates of interaction remained variable. Peer prompts ranged from 2.6 to 7.6 times per session during group-oriented contingencies but returned to zero during baseline phases. In addition, social interactions in which the peers used supportive prompts were longer and more reciprocal. These results occurred independent of teacher and adult praise.

Lefebvre and Strain (1989) examined the use of group-oriented contingency in a similar withdrawal-of-treatment design with three children with autism ranging in age from 4 to 6 years. The social skills training targeted specific behaviors, including: say your friend's name, face him or her, keep trying, ask for a toy and hold out your hand, listen and help, give a toy to your friend by placing it in his or her hand, and remember to give the requested toy. Group-oriented contingencies following the social skills training produced a higher rate of interaction than that found at baseline. However, there was considerable fluctuation in the amount of interaction that the three target children engaged in.

Peer Instruction in Social Interaction Strategies to Promote Interaction
The following interventions involve teaching peers specific social skill strategies to facilitate interaction with children with autism. Such strategies make it easier and more rewarding to interact with children with autism. Increased peer effort to interact enhances the likelihood that the children with autism will attend to the peer models and thereby enhances learning. Specific methods include peer networks, pivotal response training, and peer initiation training.

Peer Networks.Peer networks are based on the premise that an enhancement of peer understanding of, and interest in, children with disabilities will promote increased interactions. Peer network interventions thus develop a social support network by soliciting an intact group of peers to provide support for individuals with disabilities. Helping peers better understand and support children with autism is important because, as McEvoy and Odom (1987) noted, children with disabilities who have received training on how to interact with their peers will be successful only if there are receptive peers with whom to interact.

Two studies have used this approach with school-aged children (Garrison-Harrell, Kamps, & Kravitz, 1997; Kamps, Potucek, Lopez, Kravitz, & Kemmerer, 1997). Garrison-Harrell et al. used a multiple baseline design to investigate this method across three 6- to 7-year-old students who were diagnosed with autism. All target children were nonverbal or had minimal communication ability. Fifteen typical first-grade students were included in three peer networks of five peers per target child. Peers were taught how to use the target child's augmentative communication system; in addition, they were taught social skills, including initiating conversation, responding to conversation, giving compliments, sharing, providing instructions, and maintaining conversations. The target children then spent 20 minutes with their peer networks in three different settings, which were individualized to match the target child's interests. Following the intervention, peers reported higher acceptance of the students, and the target students increased the frequency and duration of their interactions across settings. However, these researchers did not test generalization to other settings. Kamps et al., using a similar strategy and sample, reported that the intervention improved interaction time for target students, and that the results generalized to nonintervention settings for two of the three children.

Haring and Breen (1992) used the peer network approach with two 13-year-old boys, one with autism and one with moderate mental retardation and severe language delay. Similar to the above studies, peers were taught how to initiate interactions with, reinforce, and prompt responses from target students. However, in this study adults also taught appropriate responses to target students, and one target child was taught to use a self-monitoring system. The results indicated an increased frequency of appropriate social interactions in nonstructured contexts. In addition, the peer network members reported improved attitudes and ratings of friendship toward the students with disabilities.

Pivotal Response Training.Pivotal response training, as described by Pierce and Schreibman (1995,1997a, 1997b), involves using role-play techniques to teach peers how to provide target children with social reinforcement, including paying attention, letting the child choose, varying toys, modeling appropriate social behavior, reinforcing attempts, encouraging conversation, extending conversation, taking turns, providing narration for play activities, and teaching responsivity to multiple cues. The approach is expected to increase social behaviors by providing multiple models who incorporate the target child's preferences in natural or loosely controlled contexts (Pierce & Schreibman, 1995).

Pierce and Schreibman (1995,1997a, 1997b) tested this model in three studies using multiple-baseline designs. Participants in the first study were two 10-year-old children with autism who were socially nonresponsive and who had expressive verbal abilities similar to a typical 3-year-old's (Pierce & Schreibman, 1995). After several weeks of intervention, both children began to initiate play and social conversation with the trained peer, and these gains were maintained during a follow-up period. There was evidence of some response generalization, but only one child generalized to untrained peers. Two other studies involving 7- and 8-year-old children with autism (Pierce & Schreibman, 1997a, 1997b) yielded similar results, with less repetitive play and increased social conversation. Interactions with untrained peers reached levels near 100% after treatment, compared with near-zero levels at baseline, a change that is clinically as well as statistically significant (Pierce & Schreibman, 1997b).

Peer Initiation Training. The goal of teaching peers techniques for initiating interactions is that the children with autism will then be involved in more interactions in which they can receive reinforcement for appropriate responses. One line of research teaches typical peers to initiate "play organizers," which includes such things as share offers and requests, assistance offers and requests, and strategies to gain the target child's attention. In addition, peers are taught how to appropriately use affection and complimentary statements with children with autism.

This approach has been evaluated for use with preschool-aged children (e.g., Goldstein, Kaczmarek, Pennington, & Shafer, 1992; Kohler, Strain, Maretesky, & DeCesare, 1990; Odom & Strain, 1986; Odom & Watts, 1991; Sainato, Goldstein, & Strain, 1992). Odom and Watts used a multiple-baseline design to investigate the utility of the peer initiation training approach with three children with autism between the ages of 3 and 5 years. All children engaged in infrequent social interactions and had communication abilities ranging from the 9- to 35-month levels. Four preschool-aged typically developing children received the peer-initiation intervention as described above. Though there was considerable variability across participants, the children with autism substantially increased their interactions during the intervention phase when teachers prompted the peers to use the initiation strategies. Odom and Watts also examined the impact of adding correspondence training/ visual feedback in which the teacher provided reinforcement to the peers when they used the initiation strategies by giving a visual cue during the play session and providing a tangible reward following the play session. This feedback intervention, combined with the peer-initiation intervention, produced increased engagement from the children with autism in a setting where adults gave verbal prompts to the peers and a setting in which they only gave feedback. However, the peer initiation intervention alone without verbal prompts from teachers regarding initiation did not lead to increases in social interactions. Sainato et al. similarly found that teaching the peers initiation strategies was not enough to ensure that they would use them.

Though peer-initiation strategies have had success in increasing the social interactions of preschool-aged children with autism, those interactions have consisted primarily of responses, rather than social initiations (Odom & Strain, 1986; Odom & Watts, 1991; Sainato et al., 1992). Using a modification of the peer-initiation intervention developed by Odom and Strain, Mundschenk and Sasso (1995) investigated use of this strategy with 7- to 10-year-old children with autism. As in previous studies, peer initiations were found to increase the responses of the children with autism. In addition, responding generalized to non-trained peers when at least three trained peers were present. Peer-initiation training was also found to increase the rate of initiations by the children with autism from 2% to 7%. However, because the children with autism were not specifically taught social initiation and response skills, their interactions remained idiosyncratic.
- DiSalvo, Carla & Donald Oswald; Peer-Mediated interventions to increase the social interaction of children with autism: consideration of peer expectancies; Focus on Autism & Other Developmental Disabilities, Winter 2002, Vol. 17, Issue 4.

Personal Reflection Exercise #10
The preceding section contained information about peer-mediated interventions to increase the social interaction of children with autism.  Write three case study examples regarding how you might use the content of this section in your practice.

QUESTION 17
According to DiSalvo, what is pivotal response training? Record the letter of the correct answer the Answer Booklet.